[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/diy/ - Do It Yourself


View post   

File: 270 KB, 2550x1754, TB303 block diagram.jpg [View same] [iqdb] [saucenao] [google]
2043890 No.2043890 [Reply] [Original]

Thread failed due to solder joint fatigue:>>2037814

>I'm new to electronics. Where to get started?
It is an art/science of applying principles to requirements.
Find problem, learn principles, design and verify solution, build, test, post results, repeat.

>Project ideas:
http://adafruit.com
http://instructables.com/tag/type-id/category-technology/
http://makezine.com/category/electronics/

>Don't ask, roll:
https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png

>Archive of Popular Electronics magazines (1954-2003):
https://worldradiohistory.com/Popular-Electronics-Guide.htm
>Some guy’s list of electronics resources:
https://github.com/kitspace/awesome-electronics
>Microchip Tips and Tricks PDF:
https://ww1.microchip.com/downloads/en/devicedoc/01146b.pdf
>Li+/LiPo batteries required reading:
http://www.elteconline.com/download/pdf/SAFT-RIC-LI-ION-Safety-Recommendations.pdf

>Principles (by increasing skill level):
Mims III, Getting Started in Electronics
Geier, How to Diagnose & Fix Everything Electronic
Kybett & Boysen, All New Electronics Self-Teaching Guide
Scherz & Monk, Practical Electronics for Inventors
Horowitz and Hill, The Art of Electronics

>Design/verification tools:
LTSpice
MicroCap
falstad.com/circuit/circuitjs.html
NI Multisim
CircuitLab
iCircuit for Macs
KiCAD (PCB layout software, v5+ recommended)
Logisim Evolution

>Components/equipment:
Mouser, Digi-Key, Arrow, Newark, LCSC (global)
RS Components (Europe)
eBay/AliExpress sellers, for component assortments/sample kits (caveat emptor)
Local independent electronics distributors
ladyada.net/library/procure/hobbyist.html

>Related YouTube channels:
mjlorton
w2aew
jkgamm041
eevblog
EcProjects
greatscottlab
Photonvids
sdgelectronics
BigClive
Ben Eater

>I have junk, what do?
Shitcan it

>> No.2043896
File: 20 KB, 580x314, RFPD3890.png [View same] [iqdb] [saucenao] [google]
2043896

>>2043890
Thread brought to you by the RFPD3890 GaAs / GaN Power Doubler Hybrid Amplifier
http://pdf.datasheetcatalog.com/datasheet/qorvo/rfpd3890.pdf

Also the MN3890S CCD delay-line is kinda interesting too, but I thought I'd go for something analog this time around.

>> No.2043899
File: 190 KB, 1062x1375, 1612344421727.png [View same] [iqdb] [saucenao] [google]
2043899

>>2043896
was about to do picrel
the HSMP-3890 RF PIN switch diode would have been fun too

>> No.2043905

>>2043899
I skipped past all the transistors and diodes because ICs are more unique, but I guess there were some interesting ones in there.

>> No.2043924

>>2043896
>The part employs GaAs MESFET, GaAs pHEMT and GaN HEMT die
wait they use those for transistors? what the hell does all of that mean?

>> No.2043956

Is this the right thread to ask for help? I'm new to electronics, and I have a PSU that I'd like to save. It fails to power on the PC but after retrying multiple times it eventually boots. I've changed some capacitors that had low capacitance, but that didn't solve the problem. To explain the problem in more detail, after pressing the power button on the PC, the CPU fan spins for a second and then the PC dies. It's strange however, that using a multimeter to check the voltages on the ATX cable everything's there. Every cable that should have voltage has it - 12v, 5v and 3.3v are all there.
What could cause a PSU to give the initial power but not maintain it? The PSU is a Grimtec SFX500, built by the chinese R-Senda, and I can't find any info online.

>> No.2043957

>>2043956
The power switch itself could be bad (on the front of the PC case).

>> No.2043958

>>2043957
Nah. I've tried shorting the two pins that turn on the PC and it's the same problem. I've also tried other PSUs and the PC boots on first try.

>> No.2043959

>>2043958
Then it's a guessing game, really. Maybe it's a failing component(s), cold solder joints or a short. Have you tried testing outside of the case with a load attached to 5V and 12V at the same time?

>> No.2043974

>>2043959
>a load attached to 5V and 12V at the same time?
How would I go about doing that?

>> No.2043982

I've been thinking about building my own RGB video cable for the GameCube. I've found all the parts I'd need on the internet except one, the 75ohm shielded cables for the RGBS + audio signals. I've been looking for the small ones, like in composite cables so I could bundle a bunch of them together through a sleeve. Anyone knows how to search for them? Because I can't seem to find them...

>> No.2044017

>>2043982
https://www.ebay.com/itm/6ft-White-VIDEO-Component-Audio-Video-Patch-Cables-5-wires-HD-TV-RCA-COMPONENT-/180996545393
>>2043974
Plug devices into the Molex connectors.

>> No.2044075 [DELETED] 

>>>2043889
no

>> No.2044077

>>>2043889
>Like how name-brand soldering stations use transformer PSUs instead of SMPSs.

what a dumb thing to say. an SMPS is 50x more likely to die in the first ten years than a simple iron transformer. so if you want a product that lasts 20 years, iron is the way to go.

>> No.2044115
File: 647 KB, 1024x582, royal rumble 17.png [View same] [iqdb] [saucenao] [google]
2044115

>>2043899

>> No.2044186

>>2043896
>>2043899
Is there a standard way of going about these? Just read the general description and see how to implement them?

>> No.2044188

I want to remake a voltmeter for my variac since the one it came with had measurements that related nothing to the actual value of the voltage. I have been having a hard time getting any answers since
>OOOOO AC SCARY
Yes it can be but it is a fact of life that it must be learned and handled. So maybe ohm can help me. I am looking for some kind of general circuit that will measure 0 to about 150V rms. It will have to be powered from the primary side of the transformer and measure the secondary. I do not know any microcontroller that can handle these voltages so I also expect some kind of rectification and regulation but how to perform this semi safely on a board is a bit beyond me. Anyone have any possible circuit designs or block diagrams of a circuit that can roughly do what I am asking for?

>> No.2044193

>>2044077
Considering all server PSUs are switchers, I'd feel confident dismissing your argument. The only other place I see transformer PSUs is in audio equipment, for low-noise purposes, possibly for metrology or whatever too. There's nothing inherently unreliable about switchers at all, in fact they deal with varying voltages on the primary side much better. In both cases, it all depends on how you design them.

>>2044186
Going about what? Putting a new "thread brought to you by", or using a random IC in a project? For the former, I search up for datasheets with the last 4 digits of the thread's OP, and see what looks the most unique. For the latter, I'd scour through the datasheet and hope it has an example circuit. If not, I'd look for an appnote.

>>2044188
Digital multimeters can handle those voltages just fine, presumably with voltage dividers. Have a look at the circuit diagram of one of those, especially an autoranging one.
But if you're a lazy shitter like me, just buy an eBay AC panel meter with a sufficiently low minimum voltage. They do current too.

>> No.2044196

>>2044193
Unfortunately I am constrained to the dimensions of the face plate of the vairac. So whatever I make has to fit into something like a D91-20. https://www.aliexpress.com/item/381416369.html

I bought one of these but for some reason they were taking the source and measurement from the same line so it was not accurate at all and even worse at the extreme of the vairac. I gutted this pulled the lcd off and learning how to drive it. I also plan on designing a new board so whatever circuit design (has to be simple) I use will fit on it.

>> No.2044198

>>2044193
>For the former, I search up for datasheets with the last 4 digits of the thread's OP, and see what looks the most unique. For the latter, I'd scour through the datasheet and hope it has an example circuit. If not, I'd look for an appnote.
Ahhh, sorry. I'm decently new so I didn't know what was the goal of these "brought to you by X" posts. Now that you answered, I think it's one of those 'opening your eyes' things, thank you anon.

>> No.2044218

>>2044193
>Considering all server PSUs are switchers, I'd feel confident dismissing your argument.

well i spent some 15 years at a bench doing production tests and repairs, and you cant dismiss that.
and today, people still bring me gadgets to repair, so i know that, when a gadget has an SMPS, that supply is the gonna be the cause of the problem about 50% of the time. linear supplies cause maybe 1-2% of faults.

>> No.2044222

>>2044218
And as a guy who has worked on things for about a year or two this guy is correct. Often see failures on SMPS devices have failures on the supply side. Linear powers supplies I often see failures from cooked traces or another parts of the device that tried to short that supply to ground.

>> No.2044225

>>2044115
it's fierce
PI are absolute faggots that demand corporate registration for datasheets. foul them for concealed weapons

>>2044196
the problem is that you still need to isolate power from the primary, or isolate signal from the secondary. the first fork of the problem is probably easier than the second

>>2044193
>server
you have a very narrow perspective on power supplies, anon. not that I've seen a computer PSU last for 10 years
>The only other place I see transformer PSUs is in audio equipment
large and small electrical appliances often use low-frequency iron where isolation and robustness are required, space and efficiency are of little relevance, the expense of certification and listing for a new design bespoke off-line flyback SMPS isn't worth it, and the low-voltage loads aren't more than a relay or two. it's a fairly specific set of conditions but it's more likely than you think
>nothing inherently unreliable about switchers at all
aside from component count, and you should have a look at a switcher's input network to understand how much it costs to make them robust against line surges, which are fairly common at the end of a long power supply wire when switching motors or heaters on and off
iron sucks in a lot of ways, but it's tough as nails and almost impervious to power supply surges and spikes, and still has something of value to offer

>> No.2044236

>>2044218
>you cant dismiss that
Well there will be selection bias there, be it the kind of appliance that people bring to you, or just that cheaper less reliable devices tend to be built with SMPSs anyhow. I've no idea what kind of overall quality the appliances with broken SMPSs that you see have. But it certainly gives your argument more weight than the unsubstantiated "50x more likely to die in the first ten years".
I think I'd like to see a study with all kinds of biases accounted for.

>>2044225
>a switcher's input network
I thought those were pretty standard for all appliances. Common mode choke, X caps, Y caps, normal chokes, MOV, etc. Not exactly required for an iron transformer I guess.
>almost impervious to power supply surges and spikes
Wouldn't those be reflected in the secondary voltage? Or are they fast enough that the inductance of the transformer kinda "ignores" them? Maybe I should run all my SMPS appliances off an isolation transformer.
>a long power supply wire when switching motors or heaters on and off
I'd have thought that's something you want to prevent at the motor/heater end, with soft-start circuits or some sort of line filter. Be nice to your grid and it will be nice to you.

>> No.2044244

>>2044225
>the problem is that you still need to isolate power from the primary, or isolate signal from the secondary. the first fork of the problem is probably easier than the second

Why would this be the case? Both sides of the transformer are referenced to earth mains. I mean I could always wireup a coil transformer but would that really be necessary instead of a voltage divider and regulator?

>> No.2044253

>>2044236
>I thought those were pretty standard for all appliances. Common mode choke, X caps, Y caps, normal chokes, MOV, etc.
but those components need to be dimensioned, it depends on the operating frequency and power draw of the switcher, and may initiate a few rounds of EMC testing and fixing. point being the dI/dt inside a switcher is a source of problems that not everyone considers worth dealing with. but energy conservation regulations may yet push switchers into these applications
>Wouldn't those be reflected in the secondary voltage?
only to the extent they can induce changes in current in the primary, which is still inductive. usually they're too short to matter much
>isolation tranny
it would be the cleanest power you ever had without a proper power conditioner
>I'd have thought that's something you want to prevent at the motor/heater end
1. not really, the filtering gets more expensive with the apparent power consumed by the load. it's a rough neighborhood, and sensitive devices are on their own. stay out of the bad part of town and put your electronics on a separate branch circuit if you can
2. tfw you are the controller inside the dishwasher or coffee maker or other controlled apparatus, ie on the motor/heater end, of a 15m branch circuit, and the motor load is <1m away from you. awkward

>>2044244
oh, I thought it was an isolating variac for some reason, nm

>> No.2044257

>>2044253
Unfortunately it is not. I did however purchase an isolation transformer. Had to clip the ground on the isolation side. Now sure why they call it an isolation transformer but then connect earth mains to all the outlets. Sorta defeats the purpose.

>> No.2044268

>>2044244
Power the meter with a battery.

>> No.2044269

>>2044268
I could do this yes but then need to have a switch circuit to turn off and on. It was meant to only come on when the transformer is plugged in but this is all besides the point. I do not have a base circuit to start off with, with some semblance of safety in mind. I mean I am not looking to make a UL listed device but still want to do the job semiproperly.

>> No.2044285
File: 47 KB, 483x446, voltmeter least-effort solution.jpg [View same] [iqdb] [saucenao] [google]
2044285

>>2044269

someone asked the same question a few days ago. this remains the simplest solution.

you can actually use any old galvanometer you have lying around, or can rip out of old equipment; just add diode and resistance in series. and print out a replacement scale.

>> No.2044289

>>2044285
Yes I could but that would defeat the whole purpose of this exercise. Might as well just buy a new variac while I am at it save me the trouble.

>> No.2044291

>>2044269
https://www.digikey.com/en/product-highlight/c/cui/ultra-compact-open-frame-ac-dc-power-supplies
This will isolate it.

>> No.2044295

>>2044291
thank you very much for this. This seems very useful. What might I consider for the secondary side for measurement?

>> No.2044305

>>2044295
A DMM? How is the meter you have wired now? Should it not normally measure the secondary side?

>> No.2044309

>>2044305
he's apparently trying to make a permanent installation into an enclosure, not simply get numbers

>> No.2044310

>>2044305
I do not have a meter right now. The one I had was not designed properly so I dismantled it and salvaged the lcd/backlight. Planned on doing something like a microcontroller to drive the lcd and ADC to measure the voltage. But since I am working with mains from 0 to 150rms there are some considerations to take here and wanted advice from you guys.

>> No.2044312

>>2044309
This guy is also correct. I am tryinng to get the new meter to fit in the place of the old one. So I have some dimensions I have to stick to. I have a similar variac to https://www.walmart.com/ip/VEVOR-Variac-Variable-Transformer-Powerstat-AC-Voltage-Regulator-500VA-5Amp-0-130-Volt/874825594
I am replacing that in the front.

>> No.2044319

>>2044312
Something like
https://www.ebay.com/itm/100A-AC-Digital-LCD-Power-Panel-Meter-Monitor-Power-Energy-Voltmeter-Ammeter-NEW/122198389743
but since the dimensions are unknown to me, search for "AC panel meter" or ammeter.

>> No.2044321

>>2044319
I am unsure of the dimensions myself but D91-20 keeps coming up for both the digital and analog meters that I have seen.

>> No.2044359
File: 4 KB, 421x365, opamp.png [View same] [iqdb] [saucenao] [google]
2044359

Whats up my ohmies.
What's the point of the diodes in the feedback line here?
There's three circuits like this on a board I'm looking at and each takes in the input from a current sensor (4V maximum).
The outputs of all 3 circuits are summed together into a fourth op amp.
I'm trying to figure out what the point of the circuits are because the current sensor outputs are actually measured by ADCs so it must be something besides just measurement

>> No.2044414

>>2044321
according to alibaba the bezel window should be 35*20mm

>>2044359
different gain, depending on the direction of the change
don't be coy, give all details. the 10k and 15k4 in parallel looks weird. anything could be happening off the right side of the circuit and we would never be the wiser

>> No.2044420
File: 22 KB, 698x563, current.png [View same] [iqdb] [saucenao] [google]
2044420

>>2044414
>don't be coy, give all details
I'm still putting the schematic together but this is what I have so far. It's actually only 2 current sensors, not 3.

There's three quad op amps on the board. This one feeds into one of the circuits in the 2nd one and two of the other circuits in the 2nd one are used for PTC sensors. Third one has 3 unused circuits and haven't checked out what the fourth circuit is for.

All of them are powered on +15 / -15V

>> No.2044426
File: 27 KB, 1176x600, Full-Wave-Precision-Rectifier-Circuit-Diagram.png [View same] [iqdb] [saucenao] [google]
2044426

>>2044420
I had a look at a few op amp topologies and it actually looks like this full wave rectifier.
I know the current sensors definitely give out a -4 to +4V DC signal so why would it need to be rectified?

>> No.2044431
File: 149 KB, 996x714, 1603585208219.png [View same] [iqdb] [saucenao] [google]
2044431

>>2044426
half right

>> No.2044435

>>2044431
Diodes are pointing the other way

>> No.2044436

>>2044435
they chose to rectify the other half of the wave, then

>> No.2044438

>>2044435
that makes it a half wave rectumfrier

>> No.2044450

>>2043890
>>Don't ask, roll:
>https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png
fuck it, lets go

>> No.2044456

>>2044450
>rain alarm
Hygrometry sensors?

>> No.2044457

>>2044438
>rectumfrier
I hate you, but you're one of us so I also have to love you too

>> No.2044459
File: 1.42 MB, 2760x1600, GarageNovelance10710834.jpg [View same] [iqdb] [saucenao] [google]
2044459

>>2044450
Fuck it. I'll join you, what's the worst that can happen?
>>2043890
>https://github.com/Rocheez/4chan-electronics-challenges/blob/master/list-of-challenges.png

>> No.2044460

>>2044459
Wait, can't I get arrested for this? I live in a mega blue state northeast. Retrying...

>> No.2044465

>>2044460
>what's the worst that can happen
>OH SHIT PEKKA ITS THE COPS
relax, it's not /k/ or it wouldn't be on the list

>>2044444

>> No.2044467
File: 985 KB, 1070x796, 1595264849632.png [View same] [iqdb] [saucenao] [google]
2044467

>>2044457
You too, Anon. lmao

>> No.2044527

I was here quite a few threads ago. I'm the DIY plasma CNC anon with some more questions concerning HF EMI.

I finally got around to welding together a steel cabinet to house all the electronics. Those include an ATX mobo, a couple hard drives, PSU, and the stepper controller board that will interface via USB.

By necessity this steel cabinet will have to have holes in it. Should I have to worry about shielding the holes for wires/vents etc?

I also bought a 100ft length of shielded CAT5 cable to wire up the servos. If I do have further issues with EMI, do those stupid little clip on ferrite chokes even do anything or should I not even waste my money?

The more videos I watch on DIY plasma tables the more I continue to worry my shit's gonna break in short order. These people seemingly have non-stop issues and constant fights with HF EMI, some worse enough that they can't even film the things up close while they're running because their cameras are affected. Most of the time it's issues with their control boards resetting once the arc starts, some have issues with the computers exhibiting some strange shit as well.

>> No.2044549

>>2044467
I wonder if there’s a market for stock images of soldering where they’re not obviously burning their hands?

>>2044527
Holes should be fine. A microwave had holes in it and no microwaves leak out. So long as it’s mostly covered, a few small holes won’t impact anything. Don’t forget proper grounding, and putting LC filters on power inputs to the computer and controller should help a lot.

>shielded CAT5
Why not CAT6 or 6A? Clip-on ferrites should have an effect.

>> No.2044552

>>2044549
>Why not CAT6 or 6A?
high-frequency content doesn't seem very high on that anon's priority list, spec whore

>> No.2044553

>>2044549
>Why not CAT6 or 6A?
because I got a deal on shielded cat5 and I figured shielded wire is shielded wire.

Is there some magic new property of CAT6 that's better for shielding I'm not aware of? I'm not using it for ethernet anyway, all unused strands will be terminated into one ground connection.

>> No.2044561

>>2043890
Say I had a small metal tube ~2mm in diameter that I wanted to heat. Is there a heater that I could use for this? Or should I just wrap it with some nichrome wire?

>> No.2044569

>>2044553
The best ones have shielding both on each twisted pair, and on the whole shebang. See the Wikipedia article for more. I suspect more shielding is a good idea for this purpose, but if you’ve already got that CAT5 I’d try that first. If you’re not using it for Ethernet, then I’d still want to drive the twisted pairs as differential pairs, with somewhat low resistances across them at the receiving end to ensure robustness. It might also be worth pursuing galvanic/optical isolation in your signal wires, to prevent ground loops, but I’m not sure. Or even just go full-retard and repurpose Toslink terminals and cables for your data, which could be fun.

>> No.2044570

>>2044561
any reason not to just wrap it with nichrome plus an insulating sheet e.g. mica? how hot?

>> No.2044575

>>2044569
>The best ones have shielding both on each twisted pair, and on the whole shebang
all of which is going to do exactly zero against low-frequency EMI
>then I’d still want to drive the twisted pairs as differential pairs
the wires aren't the aggressors. grounding one half of the pair should be enough, unless the machine is developing an uneven ground with all the plasma power
>pursuing galvanic/optical isolation
this is probably going to make it work

>> No.2044576

>>2044561
Nichrome wire isn’t insulating, you’ll want something to stop your windings from shorting against the tube, and possibly against themselves too. If you’re going low temperature, kapton tape isn’t a bad idea. If you need something hot, you may need some sort of ceramic cement adhesive.

Instead of nichrome, you may be able to find some sort of flat flexible heating element/resistor, but they’d probably be no more heat resistant than kapton.

There may be heaters designed for a similar purpose, like a 3D printer extruder, or hot glue gun, or whatever.

>>2044570
Mica is a decent suggestion, but I’m not sure if it will wrap around a 2mm tube.

>> No.2044578

>>2044575
>low-frequency EMI
how low is low? This plasma cutter has a pilot arc feature, which drives a high frequency arc to the work piece in order to establish a stable plasma arc through dirt, rust or paint.

I know these HF machines can be made to work with CNC, as plasma cutters that require you to touch off on the work piece to start the arc aren't exactly popular with professionally built CNC plasma tables.

>> No.2044581

>>2044575
>low-frequency EMI
How low frequency is it? I’d have thought harmonics in the 10’s of MHz at the least. That’s borderline transmission-line distance for the kinds of runs a signal line is going to do. Can’t hurt, but again he should still try with CAT5 first.

I’d like to see optical be more of an option for hobbyists, even if it’s cheap plastic optical. But even without fibre, I think I’ve seen some such circuits use optocouplers on the output of twisted pair. I think transformers are somewhat common on Ethernet, maybe?

>> No.2044592

>>2044570
Between 180C and 200C, maybe even lower like 165C.

>>2044576
I'll have to see how a hot glue gun works. I checked out 3D printer extruder and those use a cartridge heater. I might be able to make something like that work but I'd have to use a heater block too.

>> No.2044600
File: 118 KB, 1280x720, soldering iron heaterr.jpg [View same] [iqdb] [saucenao] [google]
2044600

>>2044592

the element from a soldering iron should do the job, as it's insulated inside and out. use light dimmer to adjust temp.

>> No.2044605

>>2044600
>use light dimmer to adjust temp.
Or buy one with an integrated thermistor and make yourself a nice little analog proportional control circuit. Analog pid would be interesting too, but almost certainly not necessary.

>> No.2044610

>>2044581
I'd try UTP through a metal conduit with soldered fittings at the termination points on either end.

>> No.2044628

I want to be an electronics and electrical engineer, is there anything I should do on the side to be competitive/active in the profession other than doing personal projects?

>> No.2044631
File: 1.72 MB, 1357x1281, kode with karly.png [View same] [iqdb] [saucenao] [google]
2044631

>>2044628
learn to code

>> No.2044639

Does everyone here regularly visit /mcg/? I only ever come here but I feel like I can see my bros typing style there... Maybe I'm mentally ill.

>> No.2044643

>>2044639
Never. Micros are covered here pretty regularly.

>> No.2044659
File: 798 KB, 2592x1944, WP_001123.jpg [View same] [iqdb] [saucenao] [google]
2044659

Being pick and place machine sucks

>> No.2044662

Do you think I'll ever need more than 3A for a power supply? Trying to decide whether or not I should go for a 0-5A or a 0-3A (more expensive but more utility options), for the foreseeable future I plan to do just electronics but I also don't want to limit myself. Maybe I can just buy both, a cheap 0-5A one and a more expensive 0-3A one.

>> No.2044669

>>2044600
Great idea, I'll have to see if I got an old one laying around somewhere I can take apart.

>>2044605
My "nice" soldering iron has a knob on the base, do you mean something like that?

>> No.2044683
File: 1.16 MB, 1385x1847, 20210301_202907.jpg [View same] [iqdb] [saucenao] [google]
2044683

>>2044414
Thanks though I think I rather just make a new one all together than just buy something especially one that might just be another pos. At least I have the chance of making a pos do what I want. First step. Finally done and got this going on a test rig. Next is sigh..... Learning to code.

>> No.2044705

>>2044631
This, but in assembly language.

>>2044659
>he tins the pads beforehand
>he doesn't have solder paste

>>2044669
yes i guess so

>>2044683
>Learning to code
If it's just an arduino nano, then you really don't need to do anything other than copy-paste code from the examples. Understanding how to drive the display will be the most difficult part, but even then it will just be delays and digitalwrites and such, depending on what's driving it. Could put in the extra effort to calculate the true rms voltage, I guess. I'd focus more on the hardware design of the voltage divider and driver. You could always use something like a digital multimeter chip (integrated display driver and adc and maybe dividers too), there's probably one or two popular ones used by hobbyists.

Not having a way of measuring and displaying current is the major shortcoming of sticking to that panel-meter form-factor, surely you can just widen the hole for a larger one?

>> No.2044710

>>2044705
>he tins the pads beforehand
Yes. Why not?
>he doesn't have solder paste
I don't even have a hot air or oven.
And yes, I melt my own solder from tin and lead, because I stole like 10 kg ingots of this shit.

Also, my laser printer died. Fucking piece of shit

>> No.2044713

>>2044710
>I melt my own solder from tin and lead, because I stole like 10 kg ingots of this shit
Well so long as you have flux I don't really see an issue.
>inb4 you stole a brick of colophony from a violin shop

>my laser printer died
Don't turn it on, take it apaaat!

>> No.2044718

>>2044713
>Well so long as you have flux I don't really see an issue.
Flux is also diy, because, fuck off, im not paying this amounts for rosin in alcohol.
>inb4 you stole a brick of colophony from a violin shop
Kek. Close enough. I bought 1 kg from idk, some beauty and idk what they use it for.
>Don't turn it on, take it apaaat!
I hate printers. I found this shit on the side of the road, and it worked 3 years... until some sensor prob got covered with dirt. Not to mention drum on that shit is dead. Can't care enough to repair. At least it is not economical.

>> No.2044719

>>2044705
Don't really need to know the current I have another device for that as is. I just want to be able to dial in the voltage since I have nothing on it to allow me to do so. Do you know of any chips people often use for the application you mention before?

>> No.2044720

>>2044718
>beauty
beauty shop

>> No.2044739

>>2044719
not that anon, but an old and trusty integrated meter ic is ICL7006. Includes ADC and LCD driving output

>> No.2044741

>>2044718
>beauty rosin
Should be mostly fine, but remember that soldering rosin is mildly activated with additives to make it work better. 6 years ago I bought a tub of soldering rosin paste for $15 or so, and I've only used 2/3 of it. I'd say it's well worth the investment. If you want to dilute it, that's still best done yourself with IPA.

>found this shit on the side of the road, and it worked 3 years
Well at least salvage the feed servos/steppers from it. Put them towards a shitty diy CNC. The laser too I guess. Might have some nice bearings or gears or whatever in there, but it might not be worth salvaging.

>>2044719
>Do you know of any chips people often use for the application you mention before
No but I've seen one or two schematics that look like diy DMMs, the sort of thing people would make 20-30 years ago. After 10 minutes of searching I came up with the 7106:
http://pdf.datasheetcatalog.com/datasheets2/23/239259_1.pdf
Which I found from this kit manual:
http://shahed.ac.ir/stabaii/Files/DVM%20digital%20voltage%20meter%20m-2666k.pdf
Needs some signal conditioning for AC (rectify, filter) but should be fine.

The MAX1491, TC7129, ICL7107, and probably others too can be found by searching "ADC display driver". The pinouts will differ depending on the type of display they drive, and how they drive it. I think all the ones I selected are designed for LCD driving, but I think the 7106 is meant to drive the segments individually (hence the high pin-count), while at least one of the others multiplexes the display. Choose whichever one fits the pinout of your display. At least one of them should be available on alibay, and you don't need anything higher than 3 digits.

>>2044739
>old and trusty integrated meter ic
I knew someone here would know one.

>> No.2044758

>lpl uses a pocket emp generator to open a digital lock
The thing sounds like some shitty spark-gap generator, but also looks really damn useful. Some cursory searches reveal it's some sort of HV generator like one of those ebay "taser" modules, connected to a coil via a spark-gap.

>> No.2044759

>>2044739
>>2044741
Gotta admit I am a bit disappointed but this would most likely fit my purpose so I might purchase a couple to dick around with. Probably an easier route anyways. Though it would be interesting to learn chip to chip communications or something. Thanks for the recommendation guys. One other thing what pin would power the ac symbol on the lcd? I see most of the pins are for the segments even the decimal and the negative bit what about to light up something like volts.

>> No.2044783

>>2044759
>bit disappointed
By what, the simplicity and lack of challenge? If you haven't dabbled much in electronics before I'd say it's a fine place to start. Even a small project like this has its nuances.
>chip to chip communications
If you're an arduino-babby it's just install a library and meme it.
>the ac symbol on the lcd
Oh it has those too? Just leave it permanently on with the V symbol, fed from Vcc. Possibly through a resistor, not really sure if LCDs need them. A pinout of the LCD with inferred labels would be good too. I'd start building up a circuit diagram in KiCAD. Making a custom LCD symbol will help you keep track of the pinout too. Paper works too I guess.

>> No.2044795

>>2044628
get an engineering degree

>> No.2044806
File: 1.38 MB, 2286x2286, 20210302_150654.jpg [View same] [iqdb] [saucenao] [google]
2044806

I opened my TV remote (again) and apparently the nrf51822 does everything except sending out the IR pulses, which they dedicate an entire stm32f0 to.
I still haven't figured out why it drains a pair of AAA in a week.

>> No.2044872

>>2044741
>but remember that soldering rosin is mildly activated with additives to make it work better.
If normal rosin doesn't cut it, it is time for zinc chloride flux for plumbing. Or maybe sand paper
>I bought a tub of soldering rosin paste for $15 or so
There is a thing tho. All fluxes I used, except real expensive ones, were shit compared to simple rosin. They had more solvent than rosin, which means it just boils off and you need fuck ton of it.
>If you want to dilute it, that's still best done yourself with IPA.
I used ethanol.
>Well at least salvage the feed servos/steppers from it.
This is low end model, I'm pretty sure in those they have only one motor that does it all...
Btw, do they have IR lasers or visible light lasers?

>> No.2044912

>>2044628
if you want to work in embedded
>learn to code in C and how to do it without an IDE
if you want to work in low level design
>pirate sim software and try to simulate circuits you're seeing during your curriculum
>master the theory of analog/digital circuits
>apply for PhD if you still have energy
if you want to work in high level design
>have common sense

>> No.2044988

>>2044581
>I’d have thought harmonics in the 10’s of MHz at the least
yeah that's fair, and a problem
>optical be more of an option for hobbyists
it would make a lot of things easier in this situation for sure. for now, there's the time-tested MIDI way of sticking an opto at the receiving end, assuming everything else is fast enough
>transformers
nah, too much fucking around, too little gittin' R done, but (like optos) the cables are balanced and run in current mode. if you don't care so much about the voltage at the other end, the wire C kind of drops out, in a manner of speaking. if you create a proper transmission line with impedance matching, most of the L kind of drops out
proper termination would also help keep anything untoward from coupling into the lines, if ground bounce could be ruled out and EMC susceptibility were more likely the culprit

>>2044576
if not mica, then a thin layer of some sort of cement that can withstand what is a relatively low heat in the greater perspective

>>2044783
>not really sure if LCDs need them
small MCUs have LCD drivers built into them. LCDs don't play well with dc. you should read some app notes about the waveforms that make these things work

>>2044806
>two 32-bit ARM Cortex-M in a TV remote
>tfw your TV remote is smarter than 90% of what's on TV

>>2044872
if you don't care whether the connection is there by this time next year, sure
>dilution makes a flux bad
only because you don't understand the science
>lasers
usually IR. I know little about lasers but you may be able to find a frequency multiplier to get green or blue or something

>> No.2044998

>>2044795
Okay
>>2044912
>master the theory of analog/digital circuits
So I should read the textbook? Isn't digital circuits just logic gates?
>if you want to work in high-level design
>have common sense
why common sense? I don't have it but wonder why.

>> No.2045005
File: 18 KB, 800x414, SJ-A801-handzender-Afstandsbediening-2.jpg [View same] [iqdb] [saucenao] [google]
2045005

I have a noob question. I'm trying to add a 2-channel remote to an automatic gate opening system. The system supports multiple remotes (up to 8) and lets you program what button activates what function. Rather than buying an original keyfob for 50 bucks, I cheaped out and got one for 6 bucks. Both the original keyfobs as my cheap one work on the same frequency, 433 MHz, but I doesn't seem like the system detects the cheap one when I try to program it into the system. I already tried erasing all keyfobs and adding them again one by one, but the system only recognizes the originals.
Am I shit out of luck and is the system more complex than I thought, or is there something else I could try?

>> No.2045008

>>2045005
I think you have to copy the code from the original remote.

>> No.2045025

>>2044783
Lcds are not DC devices. I can't just tie it I Vcc sure that'll work for awhile but it'll damage the lcd. This is a lcd for an ac volt meter but I assume it can be for any volt meter they just tied it always on so it states what you're measuring. I would have thought that there would be a pin for this.

>> No.2045027

>>2044998
>Isn't digital circuits just logic gates?
at a certain level, yes, but digital has its own degrees of structure, in common buses and their concepts and tricks (I2C, 9-bit async serial, AMBA, Ethernet, etc.) for which one should develop an intuitive sense

>>2045005
>believing that cheap keyfobs are interchangeable
yes, the system is more complex than you thought. did you even know if the same chip is used when you placed the order

>> No.2045036

>>2045008
>>2045027
I found a different reseller of the original remotes that mentions that it uses a proprietary code from the system's manufacturer. I did not check if the chips are the same 2bh.
I could probably add on the receiver for the cheap keyfob (another 6 euro) to the original system by putting the relay on the wired pushbutton input, but I'm not sure if that's worth the bother.

>> No.2045050

>>2045027
>their concepts
Any idea where I learn these concepts? I have a textbook for digital logic but it only goes up to designing with flip flops.
>(I2C, 9-bit async serial, AMBA, Ethernet, etc.) for which one should develop an intuitive sense
I'll try my best anon.

>> No.2045085

>>2045025
it's just like any other segment. you'll need to generate waveforms for it either way, unless it's the only segment on its common and you can leave it flapping in the breeze

>>2045050
that's almost enough. digital logic hasn't changed that much at the conceptual level over the past 30-40 years. we still use FFs and gates all over the place. we mostly just built HDLs on top of it all and hardly worry about assembling discrete gates anymore
a good text covers the concepts: combinatorial vs. sequential logic, finite state automata (and state diagrams), cause-and-effect, glitches (effects, causes and mitigations), various kinds of adders, counters, multipliers, etc. does yours? if so, you should have all the rudiments and some idea of how to combine them
I learned my basic digital from a text (Roth, Fundamentals of Logic Design, 1976) that ended with an applications exercise of a museum exhibit lighting controller, which sounds lame and why wouldn't you just a PIR sensor today, but in fact is the tricky part of any FIFO controller, and also gently suggests important concepts like asynchronous clock domains. continuing education means looking at other people's designs and understanding the decisions that went into the design as the designer was working on it

>> No.2045093

>>2045085
Well I want it on that is the thing and the chips I looked at don't seem to have a particular pin for it. Maybe I did not see it so where should I tie it to? Or should I also find a way to squeeze a not gate onto the board to accomplish this?

>> No.2045095
File: 194 KB, 646x748, textbook.png [View same] [iqdb] [saucenao] [google]
2045095

>>2045085
I learned everything before 'one hot' and 'verilog', since we just build HDL's on top of them, I should get learning how to write in HDL then, is this where I need to learn how to solder? I've always thought that the digital logic was, well, only for computer stuff and has to be done on the computer only (we used Logisim) and didn't know that you can go physical with it with the HDL thing.
>glitches (effects, causes and mitigations)
I don't think I learned this one, unless you mean that one obscure chapter going over error detector/corrector which I realized I skipped over.
>continuing education means looking at other people's designs and understanding the decisions that went into the design as the designer was working on it
This sounds interesting, I should pick yours off a library, thank you anon.

>> No.2045118

>>2045093
Well since at least 1 of ~3 segments will always be on, you could just run diodes from those three segments to your AC and V symbols. Assuming current doesn't need to flow the other way.

>> No.2045158

>>2045093
oh, in that case it's just another segment. your controller can probably already handle it one way or another. it depends on the LCD's matrix layout, mostly

>>2045095
ah, neat, that's a lot further than I understood at first
designs written in HDLs or captured as schematics can be used to produce bitstreams for programmable logic device control. you just need to tell the toolchain which Verilog wires/regs connect to which pins, load the resulting bitstream into the device (manufacturer specific), and you can then connect your design to outside world stimuli and responses, usually using pin headers from a dev board rather than soldered connections
but Verilog also supports facilities useful for simulation. you can simply plan out a program of stimuli, write a test harness in Verilog to generate that program on wires, apply them to your design in simulation, and probe the responses at the outputs (also any wire/reg inside the design). workflows vary by toolchain tho, see manufacturer
>glitches
these are important consequences of the varying timing of logic elements. they're more prevalent in physical, discrete-gate implementations, as the PLD/gate array toolchain and normal design practice will avoid them. check the index to make sure you didn't learn about them in an earlier chapter. there are design strategies to avoid them, eg instead of placing logic gates before FF clocks to enable them, modify the control input(s) to the FF, or use the FF's native enable input. also generally put FFs at the output end of a stage of synchronous logic

>>2045118
>Assuming current doesn't need to flow the other way.
>current
lurk moar until you read this http://ww1.microchip.com/downloads/en/Appnotes/AN658-LCD-Fundamentals-and-the-LCD-Driver-Module-of-8-Bit-PIC-MCUs-00000658C.pdf

>> No.2045164
File: 346 KB, 1276x702, drive waveforms.png [View same] [iqdb] [saucenao] [google]
2045164

>>2045118
>>2045158
Just reading 658 now, I see what you mean.
>the net AC voltage across a dark pixel should be maximised
>the net AC voltage across a light pixel should be minimised
>the net DC voltage across any pixel should be zero
I assumed they'd not be this complicated to drive, since Applied Science had to do basically the same thing to drive a multiplexed electroluminescent dot-matrix. Guy coulda just hooked an LCD dot-matrix driver to some stepup transistor circuits.

Anyhow, if his common is constantly going up-down-up-down, can't he just connect the segment to a 50/50 voltage divider or to a cap to ground? Or leave open-circuit/with a really high-impedance voltage divider/capacitor, in case the segments are the other way around.

>> No.2045171
File: 28 KB, 416x328, 1587687165178.png [View same] [iqdb] [saucenao] [google]
2045171

>>2045164
>can't he just connect the segment to a 50/50 voltage divider or to a cap to ground?
a CMOS inverter would work with the ICL7x06 on a non-muxed display if the goal is to keep a single segment active continuously. if it needs to be switched, an XOR gate works, pic

>> No.2045185

>>2045171
>a CMOS inverter
Oh yeah that would double the AC voltage compared to a passive solution.
>XOR gate
Yes that's a natural development.

Though with any luck, there will be a decimal point that he can select that isn't being used for anything, and just use that for both. Not like he'll ever use it in a range outside of "200V".

>> No.2045189

Is there any potential 'interference' that can be caused if I put my desktop monitor/kb+m setup on the same desk as my DMM/PSU/Oscilloscope?
It'll be really convenient if I could do that because then I don't have to move over to the other side of the room and crossreference something on the PC.

>> No.2045191

>>2045189
It's what I do, the computer itself isn't an issue but my computer's PSU can be. My T12 soldering station's PSU is even worse though.

>> No.2045192

>>2045191
>computer's PSU can be. My T12 soldering station's PSU is even worse though.
Ohh, could you elaborate? I've always been overly conscious about messing up my computer ever since that one time I fucked up on it.

>> No.2045194

>>2045192
Just noise that I pick up with my oscilloscope. Was making a high-ish gain preamp, I heard the humming through my headphones while testing.

>> No.2045197

>>2045194
>Just noise that I pick up with my oscilloscope.
Ah fuck, right. How can I forget about this. I even live next to a train rail so I'd occasionally have really loud train noises. Train (mid-high noise) + obnoxious heater (loud) + PC fan (mildly) is going to be a fun combination heh. Thank you anon, can't believe I forgot about this desu.

>> No.2045200
File: 76 KB, 500x417, 1588496221968.jpg [View same] [iqdb] [saucenao] [google]
2045200

>>2045185
>there will be a decimal point that he can select
where in the 7x06 datasheet does it say that the chip even cares about decimal points?

>> No.2045203

>>2045185
If that decimal point is always on I can probably do it that way.

>>2045164
They are two different technologies and sure you can drive it on dc for they will die after awhile so often they use this method of swapping logic levels on the common plate and the segments to achieve an DC signal of 0 for an AC signal needed to drive it. Gives the LCD the a 10Vpp from 0-5DC. It is very clever.

>> No.2045269

>>2045203
>you can drive it on dc for they will die after awhile

i did this on a project, where i took a 1Hz signal and applied it to an LCD segment (in series with a small cap 1n - 100n) without any synchronization with the backplane. result: it worked fine for years. after about 3 years, the segment started to get a lil bit dark, but it's still visibly flashing after some 5 years. so, death is assured, but it can very slow.

>> No.2045312
File: 31 KB, 1075x554, circuit.png [View same] [iqdb] [saucenao] [google]
2045312

>>2044420
Right, this is the entire circuit.
The current sensors give out 0 to 4V if current is flowing in the forward direction through them and 0 to -4V if backwards.
The Vref connection is the reference voltage for the microcontroller's ADC and Analog 5 is one of the analog inputs for the microcontroller.

Am I right in thinking the first pair of op amps will invert the signal from the current sensors if it's positive and keep it at 0 if it's negative?

>> No.2045315

>>2045312
Output of the last op amp is driving an optocoupler by the way

>> No.2045320

>>2045312
have you tried putting the circuit into a simulator yet?

>> No.2045323

>>2045320
>simulator
nah, I'll have access to a bipolar bench supply tomorrow so I'm gonna simulate the current sensors with resistor dividers with pots and see what happens

>> No.2045324

>>2045320
Is there a decent simulator that lets me do shit like adjust pots while the simulation is running?

>> No.2045339
File: 46 KB, 1349x602, 1587568560187.png [View same] [iqdb] [saucenao] [google]
2045339

>>2045324
if you want a breadboard simulator, falstad in the OP will do that and more

>>2045323
well, I did, and you are pretty much correct

>> No.2045345

>>2045339
Yeah, I just built it in falstad there and it looks like I guessed.
Cheers, I had a pain in my face trying to figure out how to use a pot to control PWM in spice a few weeks back

>> No.2045351

>>2045197
>I even live next to a train rail
Are they electric trains? If so you’ll be seeing some awful commutation EMI, not suitable for listening to AM radio at all. Not sure how much it will infect signals in circuits that aren’t meant to be EMF receivers, but I imagine it will to some degree.

>>2045200
Well if he has a multiplexed display he may want to use a different IC. I’d recommend getting multiplexed drivers anyhow, since it’s easier to convert a non-multiplexed LCD into a multiplexed one than it is to get multiplexed driving signals out of an IC like the 7106.

>> No.2045354

>>2045351
>Are they electric trains?
From NYC so maybe in 50000 years we'll transition to electric, I think most of the subway is operating on 1900s-50s technology. Thank you though, I'll be on the lookout for my own environment when looking at potential sources of error.

>> No.2045364
File: 33 KB, 460x365, file.png [View same] [iqdb] [saucenao] [google]
2045364

Don't laugh at me, ok...?
If the graph was x axis = I, y axis = V, then the slope would've just been R rather than 1/R right?

>> No.2045365

>>2044988
>if you don't care whether the connection is there by this time next year, sure
Nothing ever happened after using zinc chloride. Only problem is that it is conductive and will corrode copper if you leave it on. I mean, we all have seen green stuff on copper pipes near solder joint, because plumber didn't care enough to remove flux.
Shit aint in airplane after all.
>only because you don't understand the science
You don't need to understand science when Zheng is selling you vaseline or zinc chloride, and proper stuff like, idk, FluxPlus is expensive as fuck.
So...
>zheng vaseline
It boils off, leaving no flux and you need to use fuckton of it.
>zheng zinc chloride
A rip off, since zinc chloride is cheaper in hardware store.
>FluxPlus and other proper fluxes
They obviously work. But, c'mon, a bit of rosin and activators can't cost about of money they want.
>Rosin in alcohol
Just works.

>> No.2045368

>>2045364
x is the horizontal axis, y is vertical
You can tell because it's y(x)

>> No.2045380

I bought a BLE enabled thermometer for $2
How the fuck did they make it $2

>> No.2045405

>>2045269
Sure. But I could just do it properly since that is the standard and most chips will do it.

>> No.2045410
File: 753 KB, 3206x1750, IMG_20210302_www203648.jpg [View same] [iqdb] [saucenao] [google]
2045410

Dubai at home

>> No.2045416
File: 30 KB, 365x460, 1613478371873.png [View same] [iqdb] [saucenao] [google]
2045416

>>2045345
yeah, I wouldn't know how to do a pot off the top of my head in SPICE, but piecewise linear is straightforward enough for voltages

>>2045354
if he does the 7106 won't be of any use to him ofc

>>2045364
like 0.009 hours in MS paint related? sounds right

>>2045365
>and will corrode copper if you leave it on
that's fair, thorough post-cleaning is essential, even for the stronger electronics-grade fluxes. even no-cleans used in hand soldering are better cleaned off a board, since you won't know for sure whether any particular bit of flux on the board was necessarily activated
here's the rub: the pipes we use are quite a bit smaller than plumbers use, and stranded insulated wire wicks so it simply cannot be thoroughly cleaned. if you use activated flux on stranded wire you will come a gutser
if you need that much flux activity, perhaps it's better to tin the parts of the work separately using whatever fuckin' HCl dip, then join them under something milder
>proper stuff
$1 per cc is about right for flux at around an ounce sized order. you don't have to use that much of it. I have a bunch of 2cc tubes of no-clean, each of which gets me through about 400cm^2 of board surface with a lot left over for random breakout boards and wires
>boils off
>need to use a fuckton
I suspect a technique problem
>rosin in alcohol
it's a weak flux, kind of a mess to clean if you should need to, but benign in service and cheap. as expedient electronics fluxes go, this is my fave
>zinc chloride
citric acid can be found in the ethnic foods aisle, but it is cheaper than ZnCl in bulk and makes a decent water-washable flux that doesn't attack metallic Cu or Fe as hard. also great de-ruster for hand tools

>>2045380
it would be easier to guess with board photos. prob just cutting every possible corner
>reference/demo code from the chip mfr
>no FCC testing
>cheapest possible antenna
>shanzhai IP policy
>etc.

>>2045410
>casually moving your minisplit around
Dubai extravagance confirmed

>> No.2045419

>>2045380
>How the fuck did they make it $2

communism, dude.
IT is a like a train.
YOU are like cardboard obstacle.

>> No.2045424

>>2043896
I recently saw some pretty cool GaN half bridges with the gate driver integrated with the FET.

>> No.2045430

>>2045419
Communism only works if you allow capitalism to flourish inside of it.

Living in a completely capitalistic world where everyone buys your shit because lolslavewages also helps.

>> No.2045453

>>2045416
>citric acid can be found in the ethnic foods aisle
What fucking country is citric acid from? Shouldn’t it just be in the generic baking ingredients isle? Alongside cornflour and sugar and any substitutes that don’t fit in the hippy isle like stevia.

>>2045424
>gate driver integrated
Oh that’s neat! I wonder what the advantage of using semiconductors like GaN are? I’ve only ever heard of SiC transistors before, outside of Ge and Si, of course.

>> No.2045457

>>2045416
>casually moving your minisplit around
This is other room kek, but they are identical size and paint wise.
>if you need that much flux activity, perhaps it's better to tin the parts of the work separately using whatever fuckin' HCl dip, then join them under something milder
True, which is why rosin is more than enough.
>I suspect a technique problem
Nah. Just shit chink flux I bought twice in different places. And since I don't want to waste money on name brand fluxes, I use rosin. Because it is unironically just better than zhengflux
>citric acid can be found in the ethnic foods aisle, but it is cheaper than ZnCl in bulk and makes a decent water-washable flux that doesn't attack metallic Cu or Fe as hard. also great de-ruster for hand tools
For some reason they dont sell here citric acid in powder. Only lemon juice...

>> No.2045464

>>2045457
>For some reason they dont sell here citric acid in powder.
Can you get Vitamin C tablets in a pharmacy?

>> No.2045474

>>2045464
I can, but pharmacy is always a rip off. I better search for stuff somewhere else.

>> No.2045521

>>2045453
I saw it on the Mediterranean foods shelf in passing the other day
I just buy mine by the kilo from a biodiesel shop on ebay. did I mention it's great for descaling boilers, kettles, and coffee machines?
>I wonder what the advantage of using semiconductors like GaN are
imagine the sort of tens-of-milliohm Rds(on) and hundreds-of-pF Cgs you would expect from a fast 20V switching MOSFET, but at 650V. compare to the specs of a typical Si NMOS at 600V and come back after you lift your jaw off the floor

>>2045457
>True, which is why rosin is more than enough
not always, aged component leads build up full oxide coatings, which an active flux may be needed to cut through. when I was a soldering scrub and didn't know about extra flux, I had a 20-year-old roll of 100k carbon film resistors that were just not very solderable, which I eventually shitcanned
>they dont sell here citric acid in powder
that sucks. check ebay, also look at dishwasher rinse aids and cookware descaling products. other organic acids also have flux action and are used commercially. read a few SDS for commercial fluxes, it's fun and educational
if you're really in a pinch, brush some off of sour candies

>>2045464
>ascorbic acid
looks like it might work, haven't tried, don't know anyone who has

>> No.2045539
File: 636 KB, 942x908, file.png [View same] [iqdb] [saucenao] [google]
2045539

Going to try making a digital clock, I think I have everything I need. Do I need an AWG for it?

>>2045474
>>2045457
>I can, but pharmacy is always a rip off. I better search for stuff somewhere else.
I have no idea what you two are talking about but consider Amazon, I buy tablets off there occasionally, primarily melatonin ones because I can get strawberry flavor along with some magnesium supplements. Y-you're in the states r-right? (They're much cheaper than pharmacies for over the counter stuff).

>> No.2045544

>>2045539
>Do I need an AWG for it?
I don't see how 24AWG would not suffice for everything in a desk clock that's not connected directly to the mains

>> No.2045545

>>2045521
>a fast 20V switching MOSFET, but at 650V
Oh damn. I find it really weird how for diodes and BJTs and stuff like that we want to minimise the band-gap to minimise energy losses, but with FET technology it's actually not bad at all to use high bandgap semiconductors.

>> No.2045547

>>2045544
Sorry, my brain malfunctioned while typing the question. I should've realized that I didn't. I should get some sleep. Goodnight.

>> No.2045552
File: 765 KB, 1279x1280, Untitled.png [View same] [iqdb] [saucenao] [google]
2045552

Is there some trick to putting knobs w/ screws on pots? I'm trying to put on these knobs with screws to 1/4" smooth shaft pots, and if I tighten them all the way I can't turn anything, and if I loosen them they come off or spin 360 degrees

>> No.2045558

>>2045552
oh, just figured out what i was doing wrong. i was pressing down on the knobs when tightening the screw, whereas it goes a lot better if i just let it rest in place when tightening

>> No.2045561

>>2045539
oh, a waveform generator, lol no, just use whatever crystals you have, try to buy 10ppm crystals. it may not lose a whole second every day

>>2045547
fug dat, the most entertaining designs come from the least sleep

>>2045552
maybe use a spacer between the knob and panel when you install them

>> No.2045568

>>2045558

an auspicious day this is, as dude has mastered the art of screwing.

>> No.2045573

>>2045561
yeah, i do have a spacer.

>>2045568
im just not very experienced tb.h

>> No.2045574

>>2045416
Even with cut corners $2 is still very cheap for a device with a BLE uC and a mems temp/RH sensor.

>> No.2045593
File: 161 KB, 717x722, file.png [View same] [iqdb] [saucenao] [google]
2045593

Trying to wrap my head around Zener diode's notation/parameter names.
Is Vr where it is able to start conducting in the reverse bias region? Or is it Vz?
Is Ir the current at which the diode can conduct in reverse bias or is it Izt? It just so happens that my book uses Vz, Vz_0, Vzk, Iz, Izt, Izk, which makes me lose my shit.
https://www.mouser.com/datasheet/2/308/1N5333B_D-1801430.pdf

>> No.2045605

>>2045593
The difference between breakdown voltage and zener voltage will just be by what current they're specified at. Breakdown voltage will be sub-1mA kind of thing, maybe even 100 or 10µA. While the zener voltage is the voltage you care about when using it as a zener diode, i.e. as a shunt regulator that takes care of a few hundred mA. The breakdown voltage isn't terribly useful to know unless you're using it in a low-current/high-impedance system. I guess it gives some idea of how much the reverse voltage will vary as you change the current. If you use a zener with an emitter follower, you might only have 1mA through it, for example.

I_ZT, I_ZK, I_R are all reverse currents, just for different measurements in that table. I doubt you'd get 100% continuity between different manufacturer's datasheets on the names for these, they're kinda arbitrary. With any luck, your textbook should explain what its own different measurement names refer to.

>> No.2045643

>>2045521
>if you're really in a pinch, brush some off of sour candies
Sugar is PITA to remove from PCBs.
>not always, aged component leads build up full oxide coatings, which an active flux may be needed to cut through
Scrubbing is best way to fix this problem.
> read a few SDS for commercial
MSDSes usually say this:
>Modified Rosins (Rosin) : ~45%
>Terpineol : ~5%
>Maleic Acid: ~5%
Rest is probably solvent/vaseline/idk
or, just
>rosin
>ethanol
>2-butanol
So basically, it is either pure rosin in solvent, or has some spices. And those are from reputable brands (obviously, because chinks dont give you MSDS)
Also, I checked old datasheet on soviet flux, they recommend washing it off with Freon-113... I wonder, will R-141b dissolve shit? Because ngl, alcohol evaporates really really slowly. Or maybe butane? Because alcohol is slow. Or maybe try brake clean?
>>2045539
I can get citric acid from chemical supply store, but every time I go there, they think im terrorist of sort.

>> No.2045651
File: 30 KB, 430x630, 5-master_0.jpg [View same] [iqdb] [saucenao] [google]
2045651

>>2043890
Radios piss me off but I keep making them. Thanks Armstrong, you make me feel like a big dummy.

>> No.2045704

>>2045643
>alcohol evaporates really really slowly
Use isopropyl alcohol, not ethanol. Evaporates nice and quick, dissolves shit slightly better, and still isn't expensive. Actually I find it sometimes evaporates too quickly, and leaves the dissolved rosin back on my board, so I have to go multiple times.

>>2045651
>I keep making them
What kind of circuits are you making? Radio circuits have many different sorts to them, I'd be interested in trying out one or two if you have recommendations. I've got this bag of 4 tuning caps I got for a couple dollars, plus a CD4046, if that helps.

>> No.2045764
File: 96 KB, 794x1118, RDA5807.png [View same] [iqdb] [saucenao] [google]
2045764

More datasheets need to have internal pin configurations. It's really handy for optimising designs, and for using parts in ways not intended. Full transistor diagrams are even better, but not really practical for high-transistor-count highly-integrated ICs.

>> No.2045767
File: 38 KB, 568x340, confidential.png [View same] [iqdb] [saucenao] [google]
2045767

>>2045764
Also I just realised that this circuit somehow drives the 32.768kHz crystal on pin 6 single-endedly. Didn't know that was a thing. It's the middle row in >>2045764, and I assume the upper line is for external crystal, while the lower is for driving it with a clock cycle. I guess there's feedback going to the crystal via that MOSFET's source and the 20pF capacitor?

>> No.2045793

just learned USB has 0.4v or so on the datalines, not 5v as i thought
how is this actually measured? are there mosfets with vth that low? BJT's need ~0.7v and the lowest mosfet i've seen was ~0.6v to switch

>> No.2045794

i dont mean how its measured but what component is at the end of the cable that detects these low voltages?
it could be done with analog but isnt the whole point that digital shit is faster

>> No.2045815
File: 11 KB, 545x476, AM-univ.gif [View same] [iqdb] [saucenao] [google]
2045815

>>2045704
I've made a TRF set tuned for 6-20MHz for fun but am in the worst place for shortwave, that and I have 2 transmitters within a 1km radius outputting 50kw so....waste of time. Made a ton regen sets but those use vacuum tubes which you may or may not want to play with. Pic related was another I built, worked pretty well. Coming to the point where I want to make my own superhet though I can't get IF trannys

>> No.2045888

>>2045704
In my shithole isopropanol is twice as expensive as ethanol.
What about methanol?

>> No.2045907
File: 16 KB, 362x333, Capture123.png [View same] [iqdb] [saucenao] [google]
2045907

How do circuits where the opamp is the sole power source work? It seems like it shouldn't do anything, since the input voltage is zero. How do you analyze these?

>> No.2045913

>>2045907
equalize the inverting and non-inverting inputs and do KCL, the purpose is to derive some formula of the output voltage using the component values

>> No.2045916

>>2045574
apparently on these shanzhai sites there are whole manufacturing kits for quite a lot of things. it's like open source in reverse: you have a license to manufacture this thing using this kit, but you wer expected to contribute back to the community *and* change it up in some significant way, like a novel form factor or a functionality upgrade, sort of like the elite BBS of old

>>2045643
try some no-clean and water-soluble datasheets

>>2045764
that's confidential. stop giving away the secrets of our trade. if all the muggles learn that we read datasheets, our job market is ogre and every arduino pajeet is a competitor

>>2045767
neat, that's a reasonable mechanism

>>2045793
>0.4v or so on the datalines
during bus suspend, perhaps
read the USB 1.1 electrical specs, it's complex enough that you won't guess the correct answer and later versions only get more so

>>2045794
>digital shit is faster
analog shit is still cheaper, and there's no moral philosopher sitting behind you saying mixed signals aren't kosher. if there is, ignore him
USB uses analog and digital signaling for various tasks like hot-(un)plug detection, charger detection, speed selection, etc. that don't need or can't use the overhead of a full bus peer with descriptors

>>2045815
loop antennas are pretty directional

>>2045888
IPA is the standard cleaning solvent in electronics. most materials are selected to withstand (or dissolve in) it, without much thought to what other solvents will do. e.g. methanol will damage certain plastics

>> No.2045927

>>2045916
Not the case here, it's a xiaomi device and the only things special about it are the price and mijia compatibility.
They sure as hell aren't "contributing" back or whatever, it's proprietary.
The BLE uC they use also isn't all that integrated, it has quite a bit of passives around it, it needs an external LCD driver and supply is externally regulated I think.

>> No.2045929

>>2045927
>passives
have you priced them out on LCSC
enough bullshit gotcha speculation, it's an imageboard, post images

>> No.2045934

>>2045907
Parasitically induced noise will feed back until it oscillates. Your pic is an oscillator.

Ideal components and perfect shielding don't exist, therefore there will always be noise. But if you simulate it in LTSPICE, you will have to either wait a long time for it to start, or induce a transient voltage using the .ic command.

>> No.2045985

Retard here.

I tried replacing switches in my mechanical keyboard with new ones because I spilled some shit in it. Should be pretty simple, right? It's just a fancy switch.

However, I did one row. The right half of the row works perfectly fine. The other half, however, does not. Some buttons don't appear to send out a signal at all. Others send out (several) wrong keys. For example pressing space sends "5 + return/newline".

How can I fix this apart from not being retarded?

>> No.2046018

>>2045985

you screwed up the soldering, causing open circuits, or more likely, short circuits. get a light and a magnifying glass, and make like you're Sherlock Holmes.

>> No.2046038

>>2045985
>spilled some shit
did you clean the shit out thoroughly? some could have dripped down onto the pc board to very helpfully short out rows/columns for you. same goes for flux byproducts created during assembly, be sure to clean with isopropanol afterward

>> No.2046046

>>2046018
The offending keys don't seem to connect to any other keys in their rows and columns. Not sure why though. There's nothing visibly wrong even in full Sherlock Holmes mode. I could try posting a pic but not sure if it would be helpful

>>2046038
As I said, retard here. I put the pcb into water/soap and let it dry for several weeks. This kind the mechanism of the switches, hence why I'm switching them. All keys seemed to work in my os before I tried replacing the switches, though (electrically at least; not me mechanically).

>> No.2046069

>>2045916
>every arduino pajeet
don't worry, they don't read datasheets

>> No.2046073 [DELETED] 
File: 18 KB, 550x357, file.png [View same] [iqdb] [saucenao] [google]
2046073

Doing a practice test for some diode analysis using CVD. I have some calculation errors that I'm not sure how I got them. The answers are given below, my Va doesn't match theirs by a little so I'm not sure if it's just my paranoia or I genuinely I got it wrong (they said Vx but I think its because my professor forgot it was Va so Vx = Va).

Here's how I did it:

1) For diode analysis, we're told to make some assumptions on diodes' state first. My assumption for this analysis is that D1 = on, D2 = off.
2) I changed D1 to a voltage source with 0.7V (this is what I'm using for turn-on voltage), and my D2 to open circuit.
3) Because the current doesn't travel through R3 and R4, I just ignore those 2 resistors and know that the Va is probe is equal to the voltage across that parallel line with D1 and R2.
4) I did mesh analysis, drawing a clockwise loop:

-5v (given voltage source) + (1500)(I) [voltage across resistor 1] + 0.7v (diode) + (1500)(I) [v across resistor 2] = 0

Solving for the current, I get 0.00143A. Since the current going through R1 and R2 is the same, they're both 1.43 mA which matches with the answer given below.
-------Here is probably where I did something wrong.---
5) I assume that I need to know Vb because at the end of the branch/line(?), it is ground, so if I know Vb then that'll equal to Va.
I_r1 = (5v - Vb)/(1500)
Vb = 5v - 0.00143A * 1500
Vb = 2.855V
Va=Vb
Va (answer key at bottom) = 2.850V
Vb (mine) = 2.855V

Was my comprehension/method of doing the question wrong or is it just some rounding that led to my paranoia for missing that 0.005V?
And to verify my work, since Vb = 2.855V is greater than 0.7V, the diode should turn on which should agree to my assumption.

>> No.2046075
File: 46 KB, 803x590, file.png [View same] [iqdb] [saucenao] [google]
2046075

(Had to repost because incomplete crop, sorry)
Doing a practice test for some diode analysis using CVD. I have some calculation errors(?). The answers are given below, my Va doesn't match theirs by a little so I'm not sure if it's just my paranoia or I genuinely I got it wrong (they said Vx but I think its because my professor forgot it was Va so Vx = Va).

Here's how I did it:

1) For diode analysis, we're told to make some assumptions on diodes' state first. My assumption for this analysis is that D1 = on, D2 = off.
2) I changed D1 to a voltage source with 0.7V (this is what I'm using for turn-on voltage), and my D2 to open circuit.
3) Because the current doesn't travel through R3 and R4, I just ignore those 2 resistors and know that the Va is probe is equal to the voltage across that parallel line with D1 and R2.
4) I did mesh analysis, drawing a clockwise loop:

-5v (given voltage source) + (1500)(I) [voltage across resistor 1] + 0.7v (diode) + (1500)(I) [v across resistor 2] = 0

Solving for the current, I get 0.00143A. Since the current going through R1 and R2 is the same, they're both 1.43 mA which matches with the answer given below.
-------Here is probably where I did something wrong.---
5) I assume that I need to know Vb because at the end of the branch/line(?), it is ground, so if I know Vb then that'll equal to Va.
I_r1 = (5v - Vb)/(1500)
Vb = 5v - 0.00143A * 1500
Vb = 2.855V
Va=Vb
Va (answer key at bottom) = 2.850V
Vb (mine) = 2.855V

Was my comprehension/method of doing the question wrong or is it just some rounding that led to my paranoia for missing that 0.005V?
And to verify my work, since Vb = 2.855V [anode] is greater than 0.7V, the diode should turn on which should agree to my assumption.

>> No.2046079

>>2046075
>matches with the answer given below
if you'd added one more significant digit to your Ir calculations, your results would have been perfectly equal to theirs to 3 decimal places
I'm not 100% sure the prof is rounding correctly but it's been a few years since my scientific math module in high school

>> No.2046082

>>2045916
I'm aware but also I live in a shoebox so I can't be fucked to do anything with it. It's like the 1951 Zenith H500 I have...need a fucking 1L6 but there's no motivation since I'm in a poor spot for reception.

>> No.2046084

>>2046082
Considered making an ELF setup?

>> No.2046095
File: 25 KB, 966x187, file.png [View same] [iqdb] [saucenao] [google]
2046095

>>2046079
wq0ihjnwqkjepklmkekkqwelqw
God damn it wolfram, When I put in units for my calculation, they started rounding but when I put in regular numbers, they give exact value.
Thank you so much anon, I genuinely thought that my method was wrong because my friend used Thevenin instead (I'm absolutely atrocious at Thevenin) and got the exact answer while I was off...

>> No.2046099
File: 165 KB, 821x1321, 12AT7-3.jpg [View same] [iqdb] [saucenao] [google]
2046099

>>2046084
>ELF
Not really, I was trying to catch weirdo shortwave stations, I could probably do it with a regen but they end up putting out a lot of garbage at the antenna which might get me v&, something like this but without the shitty 12AT7

>> No.2046121

>>2046099
I'm no expert, but if you're not aiming at SSB signals you could go for direct demodulation/homodyne, with a PLL of some kind. I think if you use a ring diode mixer and follow it with a few passive filters, you won't have any issues with sudden noise spikes giving you clipping distortion.

>> No.2046142
File: 38 KB, 571x528, 1598400778358.png [View same] [iqdb] [saucenao] [google]
2046142

>>2046099
you can build a decent square receiving loop with a couple of 1m lengths of PVC pipe and some wire, or smaller if you like. stick a two-tranny amp on it and it's great. I built this without the relay because I'm not transmitting
also consider a folded dipole antenna hanging from the ceiling

>> No.2046168
File: 122 KB, 1446x438, Shitty Circuit.jpg [View same] [iqdb] [saucenao] [google]
2046168

So I've made a shitty camp light out of a 12v strip of led's, a buck boost converter, and a micro usb port. The port connects to my powerbank which outputs 5v 2.1A. I'm having a problem where the chip on the buck boost board (circled) is getting really hot really quickly. How can I stop this? Reduce current into the board with a resistor? Or do I just have to add a heatsink to the chip?

>> No.2046184

>>2046075
You don't even need to know the current. The there is 5v dropped across 2 identical resistors and 1 diode., So subtract the .7 diode voltage, it's 4.3v over 2 resistors or 2.15v each. Va=V1- Vr1, or Vd1+Vr2

>> No.2046194

>>2046168
what is the converter rated for? how much current do the LEDs draw? in boost mode you only have about 2.1*(5/12) amps available at the output, excluding conversion losses

>>2046184
let the scrub develop the system and technique and worry about intuition later

>> No.2046199
File: 503 KB, 1200x900, 4360-01.jpg [View same] [iqdb] [saucenao] [google]
2046199

I very much enjoyed the 'designs by the masters' chapters of the art of electronics.
Does anyone of you have recommendations for books that handle similar analysis/design of advanced circuit design?
One that jumps to mind is 'audio power amplifier design', which builds up from basic knowledge to fairly advanced designs, but is of course very specific to audio power amps.

>> No.2046204

What is the recommended cheapo clamp meter with DC capability. Fluke is out of my budget, ~300USD is way too expensive and I don't need extreme precision, even 3%+10 would be okay as long as the meter actually achieves this spec.
I heard that UNI-T can be good depending on the model.

>> No.2046208

>>2046204
Fluke released some multimeters only in China, which are pretty affordable yet decent in performance. Look up the fluke 17b. Should be about 100is USD

>> No.2046215
File: 710 KB, 1080x2340, Screenshot_20210303-231925_Banggood.jpg [View same] [iqdb] [saucenao] [google]
[ERROR]

>>2046194
This is the specs of the converter

>> No.2046224

>>2046199
I can’t even begin to understand that schematic, looks like I’ve still got a long way to go. Should really get around to reading AoE.

>>2046204
Consider getting a standard DMM that fits your budget and specs (see the EEVblog spreadsheet), then buy (or make) an addon current clamp. Do you need AC or DC?
I barely do any current measurements myself, most of what I do is on a PCB with little room for any kind of current measurement.

>> No.2046225

>>2046075

D1 voltage = 0.7
R2 voltage is 1500 x 1.433mA = 2.1495
so add 'em up and you get 2.8495
so, book is right, you are wrong.

>> No.2046242

>>2046224
>Should really get around to reading AoE.
I do really recommend it, and spend the time of doing the optional exercises as well. It really helps drill the knowledge and drive home the core understanding.

>> No.2046243 [DELETED] 
File: 38 KB, 672x346, file.png [View same] [iqdb] [saucenao] [google]
[ERROR]

Retard here. How do I do question 3a?
I get questions 3b and 3c, however, for 3a I'm stupidly lost. Supposedly my actual test will have more derivations which is why I want to get good at it.

Vz_0 = where the slope of 1/Rz hits the voltage axis of the I-V characteristic map
Rz = zener diode resistor
Vs = input
Vout = output

Deriving an equation is my biggest weakness so I'm just trying to understand how it's even done. I'm thinking that I'll have to convert that reverse bias zener diode into it's uhh, other form, with the Vz source and the Rz source, and then... KVL??? I think I can convert that drawing into a better looking one and then possibly KVL it to get V_out.

>> No.2046245

>>2046208
17b isn't a clamp meter. Why don't you actually read his post.

>> No.2046248

>>2046243
Wait maybe after I convert the zener into it's reverse bias form with the Vz and Rz, I can then detatch the load and then figure out the formula from there, and then include the load and KCL it somehow...?

>> No.2046252

>>2045916
>try some no-clean and water-soluble datasheets
No-clean means you don't need to wash it way. And rosin is technically no-clean.
Actually, no-clean datasheets mention this: 10-30% rosin, rest is IPA with spices.
Water-soluble fluxes are zinc chloride, acids and such.
>IPA is the standard cleaning solvent in electronics. most materials are selected to withstand (or dissolve in) it, without much thought to what other solvents will do. e.g. methanol will damage certain plastics
Idk. What about freons? R141b cost just as much as IPA... Boils at 32C.

>> No.2046253

>>2046245
I read over it, no need to be rude.
If you want to go that way, why not do your own research instead of wanting to be spoonfed on an internet forum.

>> No.2046258

>>2046248
Nvm the non-standardized notation names fucked me up. I think I got it.

>> No.2046263

>>2046253
I'm not the original poster and I'm not in the market for a clamp meter or any multimeter really. I have four Fluke bench meters I'm quite happy with and some shitty extech handheld that I threw in a drawer somewhere.

>> No.2046277

>>2045929
Nah it uses torx screws for some reason, don't feel like opening it up.
Images are all over the internet anyway.

>> No.2046285

Do any of you guys have a career in electronics? If so, what were the qualification required by your employer?

>> No.2046292

>>2046285
>Do any of you guys have a career in electronics?
Yes, currently working as a researcher in a university
Although I suppose it is less directly related to electronics more towards conceptual design/research and control systems
> what were the qualification required by your employer?
I was offered a contract by my supervisor after my master's thesis. Generally I think everyone working as a researcher at uni has at least a master's degree, a lot have phds.

>> No.2046293 [DELETED] 

>>2046292
>urrently working as a researcher in a university
Are you a chink? How many of your colleagues are chinks/spies?

>> No.2046296

>>2046293
I am not a chink, dutch actually
There's actually not many chinks in my department, most finish just the bachelor, some do master and then vanish. Not many actually go into research after that. I think I started my Bsc with 120 people, 100 actually turned up of which probably 10 or so being chinese. At the end of the bachelor I think there were maybe 40 people left in total, of which 4 chinese, all great people. End of master I think only 20 people left of which 2 chinese (the other 2 didnt attempt).
I think other departments do have more chinks working there, yes.

>> No.2046304

>>2046199
BigClive does that
if you can make an audio amp from scratch, and you're paying attention, you probably understand well enough that you can probably make anything

>>2046215
again, how much load are you putting on the converter? a whole 5m of LED strip isn't happening. you have barely 10W to expect at the output, which is not even a meter or two, depending on the LED density

>>2046225
you skipped one 1k5 resistor

>>2046252
>And rosin is technically no-clean.
no, it's practically no-clean. rosin accumulates dust and moisture, which you may not want in sensitive circuits. RMA really should be cleaned in any case because of the activator. technically, even no-clean should be cleaned off for maximum assembly life
>10-30% rosin
interesting. but those spices may well be the actual flux. they just have to hook up with oxides in the melt and pull them out, there's not that much to do. you won't need much bulk to do it. hence the liquid flux pen which is mostly concentrated solids
>freons
they use some of those too, in industry. expensive but effective, not the best first-line choice

>> No.2046308

>>2046296
Interesting. If you had to guess, which departments are they most represented? I would guess biotech, aerospace and physics.

>> No.2046310

>>2046304
>you probably understand well enough that you can probably make anything
If you think that, you have not seen enough circuits. Some of the 'designs by masters' are a step above. For example:
- Front end of the HP 34420 7.5 digit multimeter
- Keysights multi-slope AD converter
- SR560 preamp
- Nano amp precision programmable current source
etc
Really bleeding edge stuff that surpasses the dynamic range/accuracy/precision requirements for audio.

>> No.2046311

>>2046308
>which departments are they most represented? I would guess biotech, aerospace and physics.
There's no aerospace at my uni, but I think bio things do have a fair amount of them, yes. But i have no idea about exact ratios or that my observations have confirmation bias.

>> No.2046320

>>2046285
I work for a company making some instrumentation used in industrial settings. I'm an electrical tech. I don't do product development, but rather keep production lines running and am involved in doing product failure analysis. I only have an associates in an EE tech program from a local CC but I have a good amount of hands on experience from various independent projects I've done myself. The job is decent. My resume was pretty meh so I'm surprised I got it but it was a good fit. I'll probably finish my BS and have the company fund it.

>> No.2046321

>>2046242
>and spend the time of doing the optional exercises as well
Ok you know how /a/ has watch-along livestreams? How about we have a read-along AoE, where each day we look at a new page, do any exercises, and compare our answers? Or maybe one chapter or design per thread, or whatever.

Actually anything for which we all design something similar and compare our results would be very educational.
The /3dpg/ guys had the idea of having a CAD-off, designing competing mechanical solutions for simple problems, until they remembered that such a thread would probably be forcibly moved to the 3D modelling board.

>> No.2046322

>>2046310
Everyone knows high precision instrumentation and RF and microwave design can only be understood by the greybeards.

>> No.2046323

>>2046322
>high precision instrumentation and RF and microwave design can only be understood by the greybeards.
Thats the problem, I understand the latter, but not so much the former. The texts in AoE about these magical precision circuits are wonderful for learning more about it though.

>> No.2046325
File: 1.84 MB, 2048x1535, 9817.jpg [View same] [iqdb] [saucenao] [google]
[ERROR]

>>2046184
What the HECC, is this some black magic?!
>>2046194
>let the scrub develop the system and technique and worry about intuition later
One day................................. Right now I just have to be able to understand the difference between Vz and Vz_0 heh.....
>>2046296
>I am not a chink, dutch actually
I think he was making a reference to the meme that Chinese international students would work in research centers out of their country and then leak data when their government asks for it. Still, out of the left field. (I'm a Chinese American so sometimes my friends would meme me on what the pure Chinese students do.) I hope that my ethnic background won't hinder me from getting a job in the DoD though.
Have any advice on how to get a research position as an undergrad? No, I promise I won't leak it.

>> No.2046329
File: 2.52 MB, 332x334, 1603630535703.gif [View same] [iqdb] [saucenao] [google]
[ERROR]

>>2046325
>I'm a Chinese American
>Have any advice on how to get a research position as an undergrad?
I'll set you up with Hunter Biden for a nominal fee.

>> No.2046331

>>2046323
AoE doesn't teach you anything about how to design the circuits that go in $10k DMMs and 20GHz+ spectrum analyzers and VNAs. That's a class of design knowledge way way beyond AoE.
AoE is good, it's an intermediate reference that provides some good design theory and principles for a wide variety of analog circuits. It IS a good book, but it has limits which I think is what the other guy was trying to point out.

>> No.2046332

>>2046329
Still being a sore loser eh?

>> No.2046336
File: 135 KB, 361x287, qiqi.png [View same] [iqdb] [saucenao] [google]
[ERROR]

>>2046329
I'm gon beat you up anon :(, the country invested more into my childhood/development than my own parents. I don't even know how to speak beyond conversational mandarin heh, kind of sad desu. Can't even communicate with my own parents that well.

>> No.2046338
File: 49 KB, 600x701, 1609426843523.jpg [View same] [iqdb] [saucenao] [google]
[ERROR]

>>2046332
Just put you on the no-fly list.
>>2046336
I know kung flu.

>> No.2046342

>>2046338
Soon you'll hang and so will trump.

>> No.2046345

>>2046331
>it's an intermediate reference
When do you think one is ready for that book? When you've spent a year or two messing around with electronics? Genuinely curious, been meaning to pick it up but a lot of people said that you should know a decent amount of content first.

>> No.2046347
File: 773 KB, 333x358, 1602030529893.gif [View same] [iqdb] [saucenao] [google]
[ERROR]

>>2046342
dilate, hackaday contributor

>> No.2046365

sneed

>> No.2046371

>>2045907
With an ideal op-amp, the amplitude will either decay exponentially (closed-loop gain <1), increase exponentially (closed-loop gain >1), or oscillate at a fixed but unspecified amplitude (closed-loop gain =1). For the exponential-growth case, oscillation will be "bootstrapped" by power-on transients or noise.

For a practical circuit, you'd set it up so that the nominal closed-loop gain is slightly greater than 1 and rely upon the fact that the actual gain will drop as the output saturates. For lower distortion, you'd add some form of automatic gain control (AGC) to maintain a specific amplitude. Early oscillators often used a filament lamp as an amplitude-controlled resistor (increasing amplitude increases the filament temperature which increases resistance). Modern AGC circuits generally use FETs.

>> No.2046388

>>2046321
how about just choose a good ic for the ‘thread brought to you’ thing, and set a project based off that? easier for people to get involved, and encourages designing circuits around parts outside of our fields

otoh, if its some 32 pin digital chip with an application that barely stretches beyond its example schematic, it would just be tedious. probably better if we just select or vote on a project like ‘discrete 7400 series uart transceiver’ or ‘0-12V 0-2A linear variable cv and cc power supply that goes all the way to 0’.

either way would be cool to see anons who feel particularly inspired actually building these designs

>> No.2046396

>>2046321
>The /3dpg/ guys had the idea of having a CAD-off, designing competing mechanical solutions for simple problems, until they remembered that such a thread would probably be forcibly moved to the 3D modelling board.
Isn't 3dpg... mostly just 3d modeling? Never been there much but now that you mentioned it, they probably belong to both boards.

>> No.2046405

>>2046396
There’s a lot of trial and error trying to get printers printing smoothly, talk about hardware upgrades, tribalism between enders and prusas, and some talk about the properties of different filaments. I’d say it’s at least 60% diy, probably closer to 75%. Though maybe it’s more like 20% 3D modelling 40% technology, 40% diy.

>> No.2046409

>>2046405
>talk about hardware upgrades
You can upgrade 3d printers? I never knew that, unless you mean upgrading other equipment, which is something I didn't consider as well.
>tribalism between enders and prusas
Wat, is it one of those us vs them type of tribalism?
>talk about the properties of different filaments
Ohhh interesting, I have little to no CAD skills, so I don't know if I should pick one up, even though modeling only cost of 20-30%.

>> No.2046410

>>2046409
>Wat, is it one of those us vs them type of tribalism?
no the only real option is prusas, the rest are just shitposters

>> No.2046415

>>2046409
There’s a lot of hardware upgrading to be done. After about 1 or 2 years with a bottom-end printer, many people have replaced more than half the parts.

>>2046410
Typical prusashitter

>> No.2046419

>>2046320
>>2046292
Seems like you need at least an associate's degree to get a job. I was looking to get an electrical/electronics fundamentals technical certificate, guess I'll have to take the Associate's degree path :/

>> No.2046423

>>2046419
drunkposting here but most engineering jobs nowadays require a bachelors. i work with guys 40+ that don't have degrees but they're grandfathered in. it could be better outside the defense sector but you shouldn't expect much luck. even just getting a 2 year AA and then another 2 years for an EE from a shit state school should be enough.

>> No.2046433

>>2046304
>they use some of those too, in industry. expensive but effective, not the best first-line choice
Well, I guess I will use too, because they jacked up IPA prices 2-3 times due to corona... To the point ozone depleting phased out liquid is cheaper.

>> No.2046434

>>2046419
As someone scouring the job market at the moment, there are a fair few jobs asking for people with certificate equivalents (NZQA) but they're more for repair or assembly. Not that I wouldn't kill for a repair or assembly job, it would still give somewhat of an opportunity to display critical thinking and potentially move up to a more cognitive task.
The less technical repair jobs seem to be field-work, driving to some housewife's place while the hubby is working and shoving a new fuse up her coffee machine. Shame I don't have a car.
t. physics bachelor

>>2046433
>they jacked up IPA prices 2-3 times due to corona
Ooh, that's why you were complaining about the price. I'd object, but it's probably better than inhaling methanol. I wonder if there are any common organic solvents that dissolve hair/fingernails?

>> No.2046445

>>2046434
Associates EE tech again. I don't do assembly, there's dedicated people for that but I do work with them frequently when troubleshooting equipment. It's really not a critical thinking job. It's follow the work instructions and call the techs if you have issues. It's very repetitive and I imagine is pretty mind numbing as well and it can be very stressful if you're busy and have a lot of large or hot orders. It's almost certainly better than like food service or something but you won't really engage your brain much.

For me, I fix stuff broken equipment, occasionally calibrate, analyze returned units for the failure mode and implement corrective actions if required, run tests on sample units when required, design small PCBs or fixtures as required for the main production floor (orCAD/Solidworks), figure out how to increase efficiency and or reduce scrap, and various other odd jobs here and there. That's all pretty reasonable stuff you can do on an AS in EE/EET. Can you read schematics/drawings, use basic instruments like DMM/oscilloscope, have some familiarity with common industry (CAD/SPICE software) <---(pro-tip, pirate a copy of orCAD and Solidworks for your home PC) and you're fine.

Get in with a halfway decent company with tuition reimbursement and you're set to come out of school debt free. I generally make okay money, not six figures or anything but it's fine for one guy that lives modestly and there's insurance, 401k, etc. I'd assume that's pretty standard though. Non-union, salaried but still get OT pay.

>> No.2046467

>>2046445
I have and often use a scope and function generator, read datasheets and schematics daily, and consider myself proficient in KiCAD and LTSpice. KiCAD I know is used to a small degree in industry, not sure about LTSpice. Feel kinda like a dime-a-dozen electronics hobbyist with no relevant degree.

>> No.2046484
File: 10 KB, 777x474, beep.png [View same] [iqdb] [saucenao] [google]
[ERROR]

Hello all
I'm putting together a custom horn for my car, pic related is the general layout, im following instructables com/Car-Horn-Custom-Sound-Effects/
This is my first DIY attempt, but I think i could integrate in a PA speaker.
If I do that, would it be easier/more reasonable to install a CB as well? I'm not sure how i would do this

>> No.2046491
File: 181 KB, 1510x992, test.png [View same] [iqdb] [saucenao] [google]
[ERROR]

ATtiny programmer, design now even smaller than ever. Will do a few via tests before etching.

>> No.2046576

If I'm powering a 3.3V part from a li-ion should I use a buck or a buck-boost? Would the buck-boost allow operation when the battery drops below 3.3V?

>> No.2046602

>>2046325
>Have any advice on how to get a research position as an undergrad? No, I promise I won't leak it.
Like I said, most research positions are occupied by at least graduates. I would highly recommend getting a graduate agree.

>> No.2046603

>>2046331
>AoE doesn't teach you anything about how to design the circuits that go in $10k DMMs and 20GHz+ spectrum analyzers and VNAs.
Except for the 'design by masters' analysis parts, which is exactly why I was asking about it. Because AoE only dips the toes in this, but I would love to find more books that do more of this.

>> No.2046652

>>2046602
There was this orientation guy for my college who flexed a little by showing his research card that you use to swipe through electronic doors back when I was a freshman and I've always just been wondering how you'd even start getting access to those. Guess he was a grad student but never told me, thank you anon. Kinda weird that we have state of the art research facilities but for undergrads, we're using old equipment heh. Good luck on your research though, best wishes, make it big anon!

>> No.2046656

>>2046434
>I'd object, but it's probably better than inhaling methanol.
Well, idk. Weirdest thing is that ethanol (denatured with diethyl phthalate, 0.2%) is still cheaper than IPA.
But ethanol sucks as cleaner, takes forever to evaporate.
I mean, for $8 (IPA) per liter I'd rather try methanol (which is $1.3 per liter), or keep watching for 1 liter bottles of ethanol to appear again ($3 per liter).

>> No.2046675
File: 33 KB, 800x800, 1608479732872.jpg [View same] [iqdb] [saucenao] [google]
2046675

>>2046576
use buck-boost
or run your logic supply at even less than 3.3V and use a buck that can run in 100% duty cycle mode

>>2046656
a liter will last quite a long time if you use it sparingly. picrel are very nice, there are cheaper generic versions

>> No.2046737
File: 15 KB, 500x500, push top solvent dispenser.jpg [View same] [iqdb] [saucenao] [google]
2046737

>>2046675

>> No.2046753

hello,

what is the optimal charging rate for a battery to be in? I heard both 100% and 0% is bad, the best is to keep it around 50%, the least stress

for example, my electric toothbrush has a holder so that 99% of the time it is fully charged and connected to the electricity. is this good for the long term lifespan of the battery? this is a normal lithium-ion battery i think. same with my phone, should i never charge it all the way or let it discharge all the way?

>> No.2046763

>>2046753
You don't know the voltage of the battery for certain. There is no guarantee that 100% charge is 100% of the battery's potential capacity.

Around 50% or 3.6V for a 3-4.2V range battery is a good long-term storage voltage. Applies to lithium-ions and lipos.

>> No.2046770

>>2046737
those are very nice too

>>2046753
>optimal charging rate for a battery
depends on the chemistry. shipping regulations call for 40% SoC for Li. NiCd can take abuse from low quality charging, Li+ will just catch fire like a crazy ex when insulted
>toothbrush
a friendly Li charging controller will terminate the charging cycle when the current into the battery is low enough, then wait for the battery to fall back to a certain state of charge (90-95%) before considering restarting the cycle. trickle charging Li+ destroys the battery. you would have to reverse-engineer the battery control circuitry to know whether their controller does that
>phone
I start to charge at 30% to avoid scraping the bottom of the barrel. pull it at 90% if you are nearby and you feel like getting some extra charge cycles out of it. Li batteries lose capacity when over-discharged and can develop faults that make them unsafe to charge again

>> No.2046773

>>2046652
I managed to get two research positions in undergrad, both off campus. The first was a non-profit cancer research center struggling with a microscopic machine vision project. The job would have gone to someone much more experienced anywhere else, but as a non-profit all they could afford was someone halfway through undergrad. After a year there I moved to a different biomedical research company. Lots of research groups need someone skilled with electronics and programming even if on the surface they seem to be in a completely different field.

>> No.2046801
File: 8 KB, 509x422, regenschematic.gif [View same] [iqdb] [saucenao] [google]
2046801

>>2046142
Cheers m80, I'll think on it. Just waiting on some parts before I screw with it again but I also have a feeling I made the board too small and stray capacitances are screwing with it.

>> No.2046818

>>2046675
>picrel are very nice,
Picrel likes to piss solvent everywhere

>> No.2046843

>>2046603
I think it's likely that most masters aren't too keen on sharing their secrets wholesale. A book full of various masterful designs would be an incredible resource to have, but assuming one doesn't exist, the next best option would be to look for highly competent circuit designers, an search for service manuals for products they've been involved with. That or read through a bunch of highly specific and nuanced subcircuits to garner the same understanding that assembled them in the masterful circuits you're looking for.

>>2046675
>a liter will last quite a long time if you use it sparingly
Except he's using it to mix up his own flux.

>>2046770
Everything needs replaceable batteries. After 1000 charge-discharge cycles, phones and stuff all seem to show their age a bit.

>> No.2046855
File: 217 KB, 1785x1509, master thermocouple amplifier.png [View same] [iqdb] [saucenao] [google]
2046855

>>2046818
for me they just kind of dribble. no surprise that decades-old technology would have prostate problems
in any case, I like them for getting under QFPs and QFNs and other such places that can't be reached easily with a swab or tissue

>>2046843
what secrets, they have the same databooks and transistor models as everyone else. they have simply developed an awareness of *everything* that's going on, and arrange things so that the subtleties work to their advantage. the qualitative hacking we often do isn't enough, you have to get on the quantitative level. all that math isn't just a sorting device
>mix up his own flux
flux pens hold less than 10g of product. 1L of rosin solution serves a whole lot of board area

>> No.2046889

>>2045005
It's more complex than "same frequency". It needs to be the same modulation scheme and same data format. no, you can't tell which ones are what, it's all proprietary, unless you set up a test bench and tested the originals with a receiver demodulator and a scope, then tested a sample of each one you want to buy. Chances are even after weeks of work and tons of wasted money, you will find none that match, because each manufacturer makes up their own scheme. Just spend the extra money on the remote.

>> No.2046892

>>2046675
> picrel are very nice
meh. Food container is better. Polypropylene ones are cheap and will withstand most solvents.
As for flux making, I like using glass jars, and then apply them with drawing brushes I stole.
>>2046737
Meh.

Idk, I bought white spirit, let's see if it cleans rosin. $1,3 per liter

>> No.2046901

>>2045815
>Made a ton regen sets but those use vacuum tubes
You can also build them with JFET's
>Coming to the point where I want to make my own superhet though I can't get IF trannys
Wind your own, air coil style, find a tiny set screw to use as a slug, but you'll need some kind of signal generator set up / scope / maybe a sweep generator to get them right (and align the radio after)

>> No.2046907

>>2046892
>drawing brushes
you might be using/wasting way too much. syringes are the shit for flux application
hey, are you the guy who uses bar solder? imagine melting some into very tiny balls, screening them for size, and adding 1 part to like 4 parts rosin solution adjusted to gel consistency. beats paying for solder paste for SMT

>>2046889
>each manufacturer makes up their own scheme
the HT-12D/E were a very popular combo in RF remote controls for quite some time

>>2045815
>IF trannies
I wonder if small toroids and capacitive tuning would do

>> No.2046911
File: 98 KB, 850x638, 1584667802522.jpg [View same] [iqdb] [saucenao] [google]
2046911

Ok, why the fuck has an notebook adapter so much ICs and other shit?

A simple AC to DC is a coil, diodes,some capacitor, LM79XX.

>> No.2046912

>>2046347
>hackaday contributor
My sides have entered orbit, glad I'm not the only one who has noticed that so many of the "ladies" over there, aren't. Should really make up a fake tranny identity and do a project like a light-up LED sign to display my pronouns and see if I can get published.

>> No.2046926

>>2046911
It's a switching power supply, designed for maximum efficiency for the price. Also has to be small, and enclosed usually with no ventilation, so minimal excess heat. Linear supply doing the 65-90W output of typical notebook supplies would be huge and need cooling fan. Also, many notebook supplies (HP, Dell) have one-way digital comms announcing its specs on a third line... so that's another chip.

>> No.2046927
File: 78 KB, 866x422, file.png [View same] [iqdb] [saucenao] [google]
2046927

HOLY FUCK, OHMBROS,
IS V_OUT LITERALLY THE VOLTAGE ACROSS THE ZENER DIODE?
VB = V_OUT???
DID I LITERALLY JUST WASTE A DAY OF THINKING FOR THIS???????????????????????????????????????????????????????

>> No.2046940

>>2046855
>they have the same databooks and transistor models as everyone else
But it's the elegant simplicity of things just like that picture that you can only get with lots of experience.

>>2046911
Looks like a standard SMPS to me, albeit a little lacking in the isolation department. Transformer-based PSUs are bulkier, heavier, more expensive, and less efficient.

>>2046927
yes and yes

>> No.2046951
File: 641 KB, 543x480, file.png [View same] [iqdb] [saucenao] [google]
2046951

>>2046940
>yes and yes
Maybe I'm not cut out for electronics

>> No.2046980

>>2046951
>Maybe I'm not cut out for ________
that's normal

>> No.2047025

>>2046951
A mistake made is a lesson learned.
Important to note is:
>if two components are in parallel, the voltage across them is the same
>if two components are in series, the current through them is the same
In addition:
>if two components are in series, the voltage across them adds
>if two components are in parallel, the current through them adds
These culminate in Kirchhoff's voltage and current laws:
>all voltages in a closed loop add to 0V
Where voltage sources are positive and voltage drops are negative.
>all currents into a junction add to 0A
Where currents going into the junction are positive and currents going out of the junction are negative.

If this is too difficult, try developing an intuition with the hydraulic analogy. Everything is a closed loop of hydraulic pipes, where flow = current, and pressure = voltage.

>> No.2047027
File: 268 KB, 774x962, 1597268563254.jpg [View same] [iqdb] [saucenao] [google]
2047027

>>2046911
>A simple AC to DC is a coil, diodes,some capacitor, LM79XX.

>>2046940
>you can only get with lots of experience.
experience's main input is effort. tools do help

>>2046912
did you just assume my channel doping

>> No.2047048

>>2046980
Bro... you right... Thank you...
>>2047025
No no no, I get it anon, thank you! I didn't really have a teacher or studied the basic circuit fundamentals, kind of like throwing someone into calculus without knowing algebra or trignometry, so a lot of my knowledge is learned on the run.
Just to be sure,
>if two components are in series, the voltage across them adds
So voltage across resistor R1 and R2 both add together right? Just like how I can combine both R1 and R2 since they're in series?
>if two components are in parallel, the current through them adds
The current coming in would be the same current coming out of that parallel configuration right?
Also thank you anon!

>> No.2047056

>>2047048
>So voltage across resistor R1 and R2 both add together right? Just like how I can combine both R1 and R2 since they're in series?
if R1 and R2 are in series and nothing else is connected to the net or node that joins them, then I = I1 = I2, so by the commutative property of addition
E1 = I1 * R1
E2 = I2 * R2
E = E1 + E2
E = (I * R1) + (I * R2)
E = I * (R1+R2)
>The current coming in would be the same current coming out of that parallel configuration right?
if it can't go anywhere else

>> No.2047074

>>2046907
>you might be using/wasting way too much. syringes are the shit for flux application
Not when your flux is liquid. Idk, I don't like paste flux. It is either straight rosin, or rosin in alcohol for me.
>hey, are you the guy who uses bar solder?
Idk, but I melt my own solder, squish it between plywood sheets and then cut it with tin snips, because plumbing solder is cheap, but too thick and wrong alloy (95/5 or 50/50), so I started melting my own proprietary tin-lead-antimony alloy... And then I just found lead and tin ingots for cheap.
>imagine melting some into very tiny balls, screening them for size, and adding 1 part to like 4 parts rosin solution adjusted to gel consistency. beats paying for solder paste for SMT
I've seen dudes making solder paste by crushing solder with salt, and then washing it off. But idk, seems like too much work.

>> No.2047079

>>2045815
>I can't get IF trannys

get 'em from old radios, or clock radios, at the thrift store. i probably have over 100 from years of scavenging.

>> No.2047095
File: 54 KB, 619x491, file.png [View same] [iqdb] [saucenao] [google]
2047095

My practice exam keeps giving design questions so I have 2 I want to ask about.
This one is a parallel-based limiter.
From the name, I can already tell that I'll be using a practical/conventional silicon diode with a turn-on voltage of 0.7v in parallel with the vout, however, because the vout on the positive end has a slope of less than 1, it means that I'll have a resistor on the same 'branch/line' on it as well as a resistor in series with it because I'll be using voltage divider in order to get the 1/3.
However, what I'm not sure about is the 2.5V and the -5V parts. I know how to get a slope of 1 within those points, as I can just turn off the diodes so that V_out would be the same as Vin if I have everything in parallel, but I'm not sure how to make the diodes even begin to have its clamping/clipping properties after -5V and it's normal 1/3 slope after 2.5V.
The reason for this is because the turn on voltage, 0.7v isn't a multiple of 2.5v, like I can make it have this 1/3 slope with 3 diodes conducting, meaning a 2.1V but idk how to make it 2.5, the same for -5V, with -4.9 (7 diodes)...

>> No.2047101
File: 48 KB, 705x233, file.png [View same] [iqdb] [saucenao] [google]
2047101

This is my other design question. I'll probably go with a bridge rectifier, but this question is a bit confusing because I don't understand transformer turn ratio (never really learned it), but the other two I know... I can add in a capacitor (filter) to limit it to within 5% by just using V(ripple) = V(peak) / 2 * frequency * resistor * capacitor, and just solve for the capacitor. This one is probably a lost cause to me, ah I really hope I don't get this for the actual exam. Any advice/tip is appreciated!

>> No.2047135

>>2047095
you have access to any voltage source from 0V to 10V, right? 8 diodes to 0.5V clamps the low end. for the high end, use a 1.8V source

>>2047101
turns ratio = voltage ratio. Vi/Vo = Ni/No

>> No.2047168

>>2047135
>turns ratio = voltage ratio. Vi/Vo = Ni/No
What is ni/no? Sorry... this question is beating me in the weewee very hard

>> No.2047189

>>2047168
N = number of turns

>> No.2047192
File: 661 KB, 974x579, 47088182.png [View same] [iqdb] [saucenao] [google]
2047192

>Jumping down the antenna rabbit hole
What's the best option for covering "long distances" with 2.4ghz wifi?
parabolic dish, antenna gun?
Trying to cover 35m from one metal box to another.
Repeaters aren't really an option because there's an empty field between them.
Are the multiple antennas on the back of a router identical or do they each serve a different purpose, could I extend one of them outside of the house?

>> No.2047195

>>2047192
don't watch kreosan if you want to know what you're doing
ask the ham radio thread instead

>> No.2047209
File: 37 KB, 800x734, 1521863139109.jpg [View same] [iqdb] [saucenao] [google]
2047209

>>2047192

>> No.2047211

>>2047074
Ok, best solvent for rosin here:
90% white spirit + 10% ethanol. Like dissolves shit super fast, faster than ethanol, and faster than white spirit.

>> No.2047212

>>2047192
Yagi antennas or parabolas with biquad elements that plug into your routers SMA/RP-SMA jacks.

>> No.2047213

>>2047211
That's called aqua vagabonda in Latin.

>> No.2047220

>>2047211
Idk.
Fact is that there are commercial shits that use IPA and hydrocarbons... White spirit is some sort of hydrocarbons, and well, ethanol is substitute for IPA (because it is cheaper). Smells horribly tho, so as flux it is not good at all

>> No.2047254
File: 105 KB, 1595x707, Screenshot_2021-03-05.png [View same] [iqdb] [saucenao] [google]
2047254

Sorry for the newb question, taking a course in entry-level elctropsychics and one assignment is to build a envelope detector as you can see.

What I don't get is that the last capacitor is supposed (according several independently different sources) to remove the DC voltage and make the signal AC voltage again but fucking regardless what values I put in the signal I always get the same result as if I didn't have the last capacitor in the first place.

Even tried with different circuit simulators, it's fucking driving me nuts.

>> No.2047256

>>2047254
See: high-pass filter
You need some sort of load resistance after the high-pass cap. Even if it's 100k. Otherwise there won't be any current flowing through it regardless of bias.

Also it would be more useful to graph V1 against V3.

>> No.2047261

>>2047254
if you want the output to center around some voltage e.g. ground, you need to establish a dc path to a source of that voltage. in the real world there is always a path. in simulator world there are only the paths you explicitly specify

>> No.2047263
File: 81 KB, 1152x768, 091132v7dt6kesqlnoqnel.jpg [View same] [iqdb] [saucenao] [google]
2047263

>>2046911
Here have a smol one

>> No.2047276

>>2046911
>Ok, why the fuck has an notebook adapter so much ICs and other shit?
Get Chinese charger. It has only one IC... IF any :D

I mean. First you probably have PFC (diode bridge -> some 3 pin power shit). Then some SMPS controller (second 3 pin power shit). Then idk, because tl341 or whatever it is called is all you need.
>>2046940
It seems to be compliant. Separation is under goop.
>>2047220
>>2047211
Well, im sold on white spirit/ethanol mix. Stinks, but makes 10/10 flux which coats shit evenly with rosin and dries in 10 seconds, and also is good wash up liquid because of speed it dissolves rosin.

>> No.2047292

>>2047276
aka turpentine? separated at birth is why it removes rosin so well
I bet you could reverse the proportions, just get the rosin workable and well-dissolved first, then add that sap to alcohol as you like

>> No.2047293

How "inaccurate" is using a voltage divider and uC ADC to measure battery charge? When do I need a fuel gauge?

>> No.2047297
File: 90 KB, 474x381, 1602962459364.png [View same] [iqdb] [saucenao] [google]
2047297

>>2047293
the long, shallow slope from Qexp to Qnom over most of the battery capacity is why it's more reliable to count coulombs. if you don't mind being pretty far off over the main part of your battery's charge cycle, and just need to know when you have a certain amount of charge left for an orderly shutdown, ADC and resistor will be fine, though you will have to deal with input resistance and battery drain in your own way

>> No.2047306

>>2047297
I'm assuming that it's not the actual sensitivity of the ADC you're talking about, but rather how the voltage will depend on ESR/current? And that small changes in voltage due to charge will be masked by larger fluctuations in current?

As for coulomb counting, is this usually done by running a differential amplifier on a current sense resistor (probably with a low-pass filter after it), reading it with the ADC, and integrating it in software? I'd worry about potential drift, but so long as it isn't a long-life battery and so long as you charge it identically each cycle, it probably wouldn't be an issue.
If you're trying to account for current drawn/ESR in a voltage reading you'd need to be measuring current anyhow I guess.

Not him btw.

>> No.2047317

>>2047292
Probably.

>> No.2047325

>>2047189
>>2047168
Ayyy I think I got it, probably, thank you anon(s)!

>> No.2047345

anyone have any I2S DAC and ADC recs?

>> No.2047380

>>2047345
AKM, CS, ESS, TI make decent stuff. Look through their catalogs.

>> No.2047611

>>2047306
if you don't use a Kelvin connection, yeah
actually it integrates into a 100pF cap and counts full buckets of charge which it then promptly dumps

>>2047345
on price for a DAC, PT8211 wins hands down. sounds mediocre but that farting novelty gadget doesn't mind

>> No.2047711
File: 1.96 MB, 480x336, mee.gif [View same] [iqdb] [saucenao] [google]
2047711

>learn about new, cool-sounding DSP technique
>look into implementing it
>it makes use of obscure, balls-hard graduate-level topics from 15 different fields of math
>half the required math topics aren't anywhere to be seen in your undergrad Systems and Signals degree flowchart
>rinse
>repeat
it's all so tiresome

>> No.2047739

>>2047611
>Kelvin connection
Well I was assuming most of the ESR is inside the battery, which 4-wire sensing wouldn't help with.
>analog integrator
I didn't expect that. But it doesn't rely on the resolution of an ADC so it's probably better that way.

>> No.2047746
File: 63 KB, 901x534, 1599388703980.png [View same] [iqdb] [saucenao] [google]
2047746

>>2047739
I was exaggerating a little about the battery voltage desu

>> No.2047765
File: 56 KB, 1920x498, 2021-03-04_1.png [View same] [iqdb] [saucenao] [google]
2047765

Maybe this is not the place to ask but how the hell do you do this? I have been trying to work on this for hours but the instructor has not been very effective and teaching digital logic and the book isn't very better. Hell there is no example for me to look at for this that I have found.

>> No.2047770
File: 132 KB, 600x600, 1584448631842.jpg [View same] [iqdb] [saucenao] [google]
2047770

>>2046491
the ammount of time and money you will spend making that and even then it's usability requires an external device to talk to it.
you could have just bought a programmer and then get started actually building custom circuits for projects.

>> No.2047773
File: 109 KB, 1500x1500, 1595604110219.jpg [View same] [iqdb] [saucenao] [google]
2047773

>>2046737

>> No.2047775

>>2047770
I mean it is nice to develop skills doing these things. But I just used my arduino to program these things.

>> No.2047787
File: 666 KB, 1412x749, f9bb1a5ac3129e96dc179c5c8a7eb760.jpg [View same] [iqdb] [saucenao] [google]
2047787

>>2043890
> I have cum in my mouth so welcome to the workschlop

>> No.2047792

>>2047711
Don't give up anon...
What is it about?

>> No.2047810

>>2047711
Don't think I am as far as you are but even at the lower levels it gets hard to get any help. That valley where your skills are high enough that you struggle to find anyone who has any idea and unbelievably impossible. It makes me want to give up but EE is my only passion if I did where the fuck do i go from there?

>> No.2047818
File: 180 KB, 1053x578, 1586277505681.png [View same] [iqdb] [saucenao] [google]
2047818

>>2047765
you're in the right place
what's the Boolean equation for a mux, and can you apply it to any part of the equation for f?

>> No.2047840

>>2047770
Well 1: I already have all the parts in my possession or on order for other projects, 2: buying one of those would take even longer to get shipped here than making it will take, 3: this is another practical project to test and improve my PCB fabrication ability, and 4: lmao bit-banging USB

>> No.2047842

>>2047840
>unironically thinking tty faggotry is better than bit-banging USB

>> No.2047866

I've casted the rest of my life to electronics, what am I in for?

>> No.2047868

>bought an oscilloscope
>wave generator doesn't come until a month later
did i fuck up, i feel like i did

>> No.2047870

>>2047866
Being poorer than some codemonkey freak faggot

>> No.2047871

>>2046252
What about GBL ? I think it is chemically more complex and due to abuse probably way more expensive than methanol. No one sane or non-dumb would abuse it knowingly. But seems it is used in PCB cleaning.

>> No.2047880
File: 342 KB, 1881x1120, 1609444010873.jpg [View same] [iqdb] [saucenao] [google]
2047880

>>2047810
Not the anon you're replying to but since you mentioned that at even lower levels, it gets hard to get help, what level are you at exactly? Are you in college?
I'm currently beginning electronics so there's still a decent amount of resources, though they're not as good or as detailed as I would like them to be.
>It makes me want to give up but EE is my only passion if I did where the fuck do i go from there?
Anon, why does it sound like you'd give up on a passion because at the later stages, it gets harder and lonely? Admittedly, it does get scary and even a bit frightening, but I don't think you should give up on a passion because of it.

>> No.2047886

Anybody know how to make teardrops in KiCad? I try to route my boards with a sort of late 80s to early 90s style as I like the aesthetics but they all have teardrops on their vias/PTHs which I can't work out how to do.

>> No.2047910

>>2047868
if you haven't applied all hacks and keygens to your scope, please do so now
there are plenty of things for you to look at while you wait. pop open some easy consumer device and look for serial buses, maybe

>>2047871
GBL's produced to 185kt/yr and an easy precursor to other things
flux remover secret ingredient: 1% or so "banana oil" aka amyl acetate. cuts a hard flux like magic. you don't need much

>>2047886
you're gonna love KiCAD 6
also curved trace corners~
>teardrops for A E S T H E T I C S
>not because of the wave soldering machine in your garage
you can find plugins that produce configurable teardrops, if you don't want to fuck around with a KiCAD beta. also google TopoR

>> No.2047911
File: 35 KB, 818x279, electrical ratings.png [View same] [iqdb] [saucenao] [google]
2047911

Why is the DC thermal current rating for these switches so much lower than the AC thermal current rating?
I understand why the DC interrupt rating is so much lower, but thermal current is the current the switch can carry without having to break the load. Why should that be different for DC?

>> No.2047914

>>2047910
>if you haven't applied all hacks and keygens to your scope, please do so now
Hacks and keygens...? What do they do?
I got a siglent 1104x-e so I don't know if it is hackable, is it one of those game type of hacks where you can get free software?

>> No.2047915

>>2047914
some scopes have shitty protections on them that let you unlock features for free when normally they cost extra.

>> No.2047920

>>2047880
Tell me every possible way of making a vector image like this
Guessing the best way is to use a stereo audio DAC, but it would be cool if there was some analog method too

Need a lainscope 24/7

>> No.2047932

>>2047920
>stereo audio
>analog
In theory you could generate this vector image with an audio tape loop.

>> No.2047939

>>2047920
https://youtu.be/hjyrb9xthas?t=713

>> No.2047951

>>2047915
Kind of weird that companies would still lock extra add-ons behind pay walls as if they don't expect people to just hack through them all

>> No.2047956

>>2047920
a /jp/fag does it all the time apparently, you can probably try to find him there, best of luck and bring the secret home to us

>> No.2047961

>>2047932
That’s a really interesting idea, nice and simple. A lot of common tape standards aren’t infinitely looping, and require kinda complicated mechanisms, and make noise and wear. So maybe I should try to write analog audio data on a disassembled floppy. With any luck one of the two junk stores on the block have them.
Don’t think an HDD disc will work, but I guess it might be worth a try if I can get a spare one.

Not going full oscope-music just yet.

>>2047939
That’s some neat software, but why Blender instead of Inkscape? Don’t really want to spend t he time converting pixel images to vector either, since I lack a good workflow for it. I’m considering writing my own, converting a B+W pixel image into one connected line, then re-interpolate it at a particular framerate. Probably a fixed framerate, but there might be effects to get from having it be dynamic. Might want to add an option to raster scan through an area to colour it in, but baby steps, I’m not a very good programmer.

Ideally I’d get the output and use it to train a neural net, hopefully getting something optimised enough that I can run it on unfiltered images, or even run it live. But while I’ve wanted to use neural nets for a while, I don’t have any idea how I’d use one.

>> No.2047963

>>2047956
Noted, I’m willing to dive in there for a while. Is that where the lain image came from?

If I do write some shitty python software, I’ll share it somewhere.

>> No.2047965

>>2047961
>A lot of common tape standards aren’t infinitely looping, and require kinda complicated mechanisms
Actually it's much easier than you think to make a tape loop. You just record the section of tape you want to loop, slice it out, tape it in a loop, and package it back up inside the cassette. Then it plays in a loop.
This was real common for DIY music effects in the 80s and 90s. There's a lot of info on it still out there.

>> No.2047967

>>2047951
standard practice in the industry actually.
Rigol is more of a hobbyist/entry level scope so they don't really care if people hack the features for free.
Companies like tektronix who charge 10's of thousands for a GHz scope and thousands extra for each SW package get away with it because it's company money paying for it not joe in his garage.

>> No.2047973
File: 1.99 MB, 1354x1062, lain.png [View same] [iqdb] [saucenao] [google]
2047973

>>2047963
Yeah, I shared it with my friend and they said that a jpfag did it, the model is C1-93 so you can probably search through the archives for any keywords like oscilloscope, lain, etc. I'll go ask my bud since he frequents jp a lot more and will get back to you on it. I just did some Google search and this might be his stuff, don't quote me on it though, I don't know how to code for my life (though I should learn)

https://github.com/topics/oscilloscope?l=c%23

>> No.2047974

>>2047973
https://github.com/wwwMADwww/FT2232ImageOutput
Correction*

>> No.2047979

>>2047967
>joe in his garage.
You guys have a garage? The only space I have is my desk heh.

>> No.2047982

>>2047961
>converting pixel images to vector
the first place I'd look for an easy workflow would be some KiCAD integration. you know how circuit boards have gotten lately

>> No.2047994
File: 278 KB, 1755x1240, Sr8x5_XY16_Z8.png [View same] [iqdb] [saucenao] [google]
2047994

>>2047965
Oh, so in exchange for loopability, they vastly shorten the length of the tape? I wonder if they had some sort of indexing method for stopping them before momentary effects replayed themselves.

>>2047973
>>2047974
Thanks, will do. That github looks really promising, I'll have a look at the hardware side first to see what it does.
>add DAC to Z channel to using it as blanking and b/w color output
Now that's something I hadn't thought of, I'm 90% sure my scope has that feature. It's crossed out too, so I can only assume it's functional. Hope my FT232H will work, though I'd need to buy new parts either way.

Also how did it take me this long to realise you can chain R2Rs

>>2047979
Same, and even that's cluttered and tiny. I solder atop 2 inches of forgotten components.

>> No.2048007
File: 369 KB, 1276x1134, the horror.png [View same] [iqdb] [saucenao] [google]
2048007

>>2047994
wait what the fuck his R2Rs are discrete
and tuned with trimpots
and the only ICs are generic SOIC16s
and why are his 74HC595s only using 4 outputs each
and why the fuck did he use EasyEDA to draw the circuit diagram, only to manually select and place components in KiCAD's Pcbnew
he better have used EasyEDA's netlist somehow

>> No.2048010
File: 699 KB, 1984x1120, mudac%20photo%202.jpg [View same] [iqdb] [saucenao] [google]
2048010

>>2048007
IT GETS EVEN WORSE
IS THAT FUCKING CARDBOARD

>> No.2048012
File: 129 KB, 912x876, lolcaps.jpg [View same] [iqdb] [saucenao] [google]
2048012

>>2048010
friendship ended with mudac

>> No.2048017

>>2048007
I have no idea what's going on (maybe in 3 years when I git gud) but I'm happy I posted the lain pic hehe, that's my only contribution to the community so far...

>> No.2048021

>>2048007
There were some meme audio DACs made like this
They all measure like shit

>> No.2048048
File: 780 KB, 1286x1737, ?.png [View same] [iqdb] [saucenao] [google]
2048048

So this module has several powering options, there's a table below that shows connection state of some components. But since they're all inside the module and I can't touch them, is there any meaning to the table? Or does it mean something else?
Also what's that little network between VDD, VDDH and VBUS (bottom left corner) for?

>> No.2048064

>>2047994
>Oh, so in exchange for loopability, they vastly shorten the length of the tape? I wonder if they had some sort of indexing method for stopping them before momentary effects replayed themselves.
Yes, it only loops a few seconds. You'd need a lot less than that to draw a vector graphic on a scope.
No indexing. There were a number of specialty tape formats that did have indexing like that, but the classic "splice the tape into a circle" method is crude and effective.

>> No.2048068

>>2048007
but interest was quickly lost, the project

>>2048048
>But since they're all inside the module and I can't touch them, is there any meaning to the table?
it's not unusual to draw the circuit and design the board so that parts can be placed or omitted to implement different variations on the product theme
consulting the table for your module's option variant, you can determine how the module pins for your module are connected internally
>little network
it's for 0 ohm resistors or not, placed by manufacturing according to the table

>>2048021
it was just how you did it before the SoundBlaster came along

>> No.2048090
File: 111 KB, 379x381, MODBLE - MS88SF2.png [View same] [iqdb] [saucenao] [google]
2048090

>>2048068
There are no options, just one model.
All parts are inside the module.

>> No.2048097

>>2048090
no options for you, but if you wanted 1000 of one of the others, contact the factory

>> No.2048117
File: 131 KB, 1280x1280, gondolacap 2.0.png [View same] [iqdb] [saucenao] [google]
2048117

>>2048012
picrel

>>2048017
Keep learning, keep asking questions and challenging yourself. And KiCAD is a pretty good piece of software, both for drawing out circuit diagrams, and for making PCBs. Also it's got 90% of all ICs you could ever use in it, meaning you don't have to constantly cross-reference the datasheet for the pinouts. I recommend snooping about on local 2nd-hand websites for an analog oscilloscope, they're. And thanks for that post!

>>2048021
I imagine if you make an array of low-tempco, low-drift, matched SMD resistors, the resultant quality wouldn't be bad. But that would be more expensive than an integrated SPI or I2S DAC, and a lot harder to assemble. Buying monolithic R2R arrays however, that's still viable. I am accepting part recs.

>>2048064
Yeah of course it's more than enough for an oscope vector. Though I'd definitely consider running the tape slower if it would improve frequency response.

>>2048068
>but interest was quickly lost, the project
Are you the anon who made the project? Judging by your images, it works really well for what it is. But there's a lot of questionable design decision rectifications between its current state and it being something optimised for a limited production run. If nothing else, a monolithic 12-bit stereo DAC would be able to handle the X and Y channels pretty well. Z could be done with something else. I'm also not sure why you'd choose to use an FT2232 instead of a dedicated MCU with native USB.

>> No.2048123

>>2048117
This is apparently the best known discrete r2r audio DAC https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-holo-audio-may-probably-the-best-discrete-r2r-dac.10161/
It's $4k

>> No.2048125

How do I design my PCB to allow for external uA-level current measurements? Or even in-system measurements if possible.

>> No.2048126

>>2048125
proper shielding and grounding to prevent surrounding electric fields from introducing phantom currents, and a bunch of other typical noise-immunity techniques. then just read the datasheets of anything that touches the current traces very carefully.

>> No.2048142

What's the best way of making a fast (10kHz - 1MHz) triangle wave for a class-D amp or other PWM use? I think sawtooth waves will give worse transients and power rail ripple. I'm looking at circuits online, but a lot seem to use 555s or other relaxation oscillators, and one used what I think was supposed to be an op-amp integrator. Now as far as I can figure, an integrator-based oscillator is a smart way of doing it, as the output amplitude is dependant on the adjustment of the positive feedback resistor, and not of the frequency. I suspect the wonky wave of a relaxation oscillator like a 555 would give you a bunch of THD at the output.

If a comparator+op-amp integrator oscillator really is the best choice, I suppose standard audio-grade components won't be fast enough, and I'm not fond of my shitty 339s and 393s having open-collector outputs. Any recommendations? RRIO would be handy too I guess. Be nice if there was a list that recommended common ICs.

>> No.2048151

>>2048142
>integrator-based oscillator
integrators are fucky

>> No.2048170
File: 233 KB, 1607x867, smps.png [View same] [iqdb] [saucenao] [google]
2048170

>>2043890
Anyone ever made their own switching power supplies? What do I do about the transformer? Do I wind my own, or are there some that I can buy from the interwebs. I need 12V DC for my circuit to work.
This is strictly a learning exercise so I will buy an off the shelf power supply but I'll buy any number of components to make one
I'm an CE and I don't have much experience with analogy stuff like Transformers etc

>> No.2048195

>>2048151
Over a long time sure, but when it's a triangle wave generator running at over 1kHz there isn't much wrong with it.

>>2048170
I tend to avoid designing my own line switching converters, since they're not that well optimised (often bang-bang like that TNY) due to the necessity to power the switching controller on the primary side, and the lack of an analog feedback path from the secondary side. The requirement for a low q-current doesn't help. A 12V power brick or laptop brick will have a significantly more complicated board, which I assume means it's got variable duty-cycle and that kinda stuff to its advantage. There are reasonably linear analog optocouplers out there that run with an op-amp on their input side, which would work if your output voltage is anywhere from 3V to 36V. Getting a line switcher that accepts such a voltage is another thing entirely. See if you can find the most optimised line-switching converter schematic possible, and strip out features until you've got a compromise that suits you. TNYs are cheap and easy to design for, but if you only want a line switcher that goes to 20W or so, you'd have an easier time just buying an off-the-shelf one. Be it a power brick or a PCB-mounted jobbie.

But if you have to make one, transformer selection is indeed the biggest issue. Ideally you find a standard ratio, which should usually be a possibility if you're doing a standard conversion factor, and it's not like the ratio is critical. So long as you're within a factor of 2 of your calculated ideal ratio, the efficiency should be passable. Not that I'd have any idea how to calculate a transformer winding ratio, nor how to size the inductance to match the switcher's frequency. If you're stuck on the matter of calculating those values, I'd hope there are online calculators out there, if not a nice appnote or spot in the switcher's datasheet.

>> No.2048311
File: 96 KB, 427x600, 4444.jpg [View same] [iqdb] [saucenao] [google]
2048311

https://www.youtube.com/watch?v=8Veobksi3pI&ab_channel=ZackFreedman
???????

>> No.2048334

>>2048311
>retard destroys a free oscilloscope.
kek

>> No.2048335
File: 333 KB, 887x863, 1606064820121.png [View same] [iqdb] [saucenao] [google]
2048335

>>2048170
>I'm an CE and I don't have much experience with analogy stuff like Transformers etc
Bob Pease has a story just for you
there are common transformers for common applications on aliexpress, including the ubiquitous 5V 2W USB "charger" with shite load regulation, and there are companies that will custom-wind them to your spec in lots as small as 5 for pretty reasonable prices. creating the spec is more difficult, esp for flyback transformers, as there's a lot to specify about the core
off-line SMPS are a fun challenge, no doubt, but they are no joke. you are going to need some power electronics background and great workmanship in order to reduce the likelihood of your circuit blowing up in your face below 1, even if you take one of the Power Integrations cookbook designs or wizard designs and follow instructions and BOM precisely. at the very least grab the canonical book on SMPS (Pressman) and try to understand what you're signing up for on this ride
important: shield your face and eyes against small explosions and flying objects, and use an isolation transformer. a plexiglas cover for the circuit would be wise too
potted, isolated power supply modules such as CUI's VSK series or many chink clones solder right on to your board, in sizes from 1W to 25W

>>2048142
LM311 should be fast enough for the lower part of your range and has an uncommitted transistor output, though an open emitter output has its own problems when it comes to speed and stability
at the upper end of your target range, LMH6600 series are decent, cheap, fast CMOS RRIO major-brand op amps that don't consume ridiculous amounts of power, but read the datasheet before trying to apply them as comparators. there are chinky opamps like GS8091 that are 1/5 the price with about the same level of performance and treachery

>> No.2048360
File: 17 KB, 346x257, IEEE.png [View same] [iqdb] [saucenao] [google]
2048360

Trying to join IEEE and I came across this particular section. Am I allowed to change this later on? Currently I'm thinking components/circuits/devices etc because that sounds like /ohm/ and I like the stuff we do here but it also sounds like computing and processing (hardware and software).

>> No.2048364

>>2048360
>geoscience
hmm

>> No.2048476

>>2048360
>join IEEE
why? these """professional""" associations are always full of corrupt bureaucratic kikes

>> No.2048483

>>2048311
absolutely disgusting

>>2048360
if I were you, I'd choose general topics for engineers until I have chosen a proper specialty

>>2048476
on the CV, it signals professional investment, which can open doors
also, helping out libgen with another library account with its quota of downloads
finally, IEEE seems to be among the less corrupt, even if they stole their logo from ARRL kek

>> No.2048490

>>2048476
>he doesn't want to become one of the corrupt bureaucrats, flying around the world to sham "conferences" set in massive hotels + golf courses on tropical islands

>> No.2048536

>>2048311
The integrated computer hooked up to the scope's data output could be pretty useful for more advanced math functions. No clue if the water cooling actually improves anything though. The aesthetics could do with a lot of work, I'd stick to a dark matte grey with amber lighting.

>>2048335
I think it would be smart to use one dedicated op-amp and one dedicated comparator, instead of trying to use two op-amps. I do want class-D operation, so being able to work at 1MHz is reasonably important.
Actually I'll take a few minutes to calculate the ripple with respect to the effective DAC LSB size at varying frequencies, using a 2-pole LC filter.

>> No.2048546

>>2048117
> Though I'd definitely consider running the tape slower if it would improve frequency response.

It's the opposite, actually. The faster you run the tape, the greater frequency response and fidelity you get.
Professional studio tape machines run at 15 inches per second for example, while cassette tapes run at 1 7/8". Dictation tapes run even slower.

>> No.2048557
File: 267 KB, 1039x559, ooga.png [View same] [iqdb] [saucenao] [google]
2048557

Bouta start learning how to laplace transform on circuits and I have no idea what the fuck I'm doing kek

>> No.2048559

>>2048490
>he thinks they'll accept him into their cabal, rather than just milk him for membership fees

>> No.2048568
File: 335 KB, 2286x1180, driver board.png [View same] [iqdb] [saucenao] [google]
2048568

>>2048536
>calculate the ripple
Ok with 20kHz cutoff, 1MHz oscillation, that's -68dB or 1/2500 times attenuation. Getting the logbase2 of that gives 11 bits. So if I've got a full range sine wave on a 10 bit DAC, it will have worse ripple than a 1MHz class-D with a perfect signal source. Changing to 5MHz bumps that 11 to 16, which I think is definitely worth doing, but is starting to get into the "my MOSFET won't switch fast enough" range.

Actually I'll be using these IR2184 FET drivers, which have a dead-time of 500ns, which makes switching even at 1MHz a stretch. Wonder if it's worth making a 4-pole LC filter? 4-pole with 250kHz would give me 15 bits of difference and a period of 8 times the dead-time, and let me use more common MOSFETs. 4 times the total dead-time, since it goes dead twice a cycle. But at this point I'd probably be better off looking for other FET drivers.

Pic related is the half-finished project schematic, a multipurpose board that can be used to implement PID control, audio DSP, USB HID, motor driving, etc. Slots for your own MOSFETs and ability to set the gate drive voltage are important. This one is being designed with an ATmega32U4 as a first prototype, but once I get my STM32s to work like I want I'll likely use one of them instead.

>> No.2048571

>>2048568
Oh and of course it has the ability to act as a class-D amp by the MCU pushing PWM directly to the FET drivers. But since the feedback would need to be via an ADC, I suspect the latency will make it inferior to using the DAC to feed an op-amp that feeds the PWM generator, with the output of the amp feeding back into the op-amp for feedback. That's something I'll be able to test once I've got a prototype done.

Also I'd like to use the full-bridge for driving a peltier cooler/heater with temperature feedback. Maybe I should include a thermocouple reading IC, but I think that's going overboard. Using the audio-quality ADC I plan to put on the board anyhow, alongside use of the multipurpose op-amps, should let me read any kind of thermocouple or thermistor. Should add a temperature sensor to the board too I guess, for cold-junction compensation.

>> No.2048614

>>2048536
those opamps aren't so bad, they just happen to be designed for speed. permissible differential input voltage suffers as a result of the emitters of the input trannies being very sensitive and only one stage deep. they are still RR inputs in common-mode terms. in this circuit the diff limit's main effect is to limit the amplitude of your carrier wave

>>2048568
>>2048571
steeper cutoffs are feasible, but leave margin above 20kHz or you'll step on the high end
the 2184 on/off times are severely mismatched, you'll get major dc bias
the STM32F302 has some nice analog facilities onboard and it's a bretty nice core. you could switch off a comparator output and send a timer output instead
>cold-junction compensation
I doubt you need such accuracy that you wouldn't just go with the cheap thermistors. they build them small these days

>> No.2048646

>>2048614
>leave margin above 20kHz or you'll step on the high end
6 or 12dB cutoff at 20kHz wouldn't affect much, would it? I thought barely anything was ever done up that high. It looks like I'll be going for 12b or less, so I'm not hugely concerned with high fidelity, but I'll still consider pushing my cutoff up by a few kHz. Might be able to get a better cutoff than a simple butterworth too, maybe.
The LC filters will be off-board anyhow, so for now I just need to consider what's going on the dev-board itself. But I want to have the op-amps (and probably some comparators) on-board since they'll be handy to have around ADCs and DACs anyhow. Not sure whether to go for strategically placed solder jumper pads to wire them up, or just give some perfboard. I hate perfboard so I'll try the former.

>cheap thermistors
I'll use them where possible, but if I have something that goes too hot for a thermistor I won't have any other option than a thermocouple. Like, say, powering a soldering pencil without its station.

>> No.2048720
File: 50 KB, 491x505, 1610821879564.png [View same] [iqdb] [saucenao] [google]
2048720

>>2048646
>6 or 12dB cutoff at 20kHz wouldn't affect much, would it?
"air" starts above 10kHz. cymbals, plucked strings, percussion attacks, that mosquito ring tone meme, things like that
cutoff is the 3dB frequency. how bad it would sound depends on the material. you could certainly tune the stages of your filter differently. half-power at 20kHz isn't terrible, plenty of worse consumer gear exists, but I can still hear when it is switched in or out. it's subtle but noticeable when it's missing. older people will hear it even less
>Not sure whether to go for strategically placed solder jumper pads to wire them up, or just give some perfboard
I like the former. or pic. someday I'm going to design the perfect SMT breadboard
>I won't have any other option than a thermocouple
that is true sometimes. not sure you'd want to do it on a board that might be dissipating quite a bit of power and causing a lot of CJ drift, and is it going to be more of a PITA compensating it than it's worth

>> No.2048763

>>2048720
>someday I'm going to design the perfect SMT breadboard
Problem is, lots of SMD footprints don't have the same pitch as each other. You need at least one IC package (better be SOIC), one transistor package, and one passive package to fit with one another without danger of shorting. So probably not relying on diagonal placements. I haven't looked into it, but I suspect it won't be too difficult. Fitting some common castellated via pitches would be nice too.

Wait do you mean breadboard or "perfboard"? Because having a bunch of pre-existing connections would make your board a lot less flexible, and I'm not to fond of cutting connections like in strip/vero board. Especially since the narrower pitch of a 2mm or 1mm perfboard would give you the ability to solder bridge between pads even easier than 0.1" perfboard. But it would still be good to have 2 or 3 existing continuous power rails. 3 because I'm an analog fag.

>not sure you'd want to do it on a board that might be dissipating quite a bit of power and causing a lot of CJ drift
Probably right, I'd be better off using a thermocouple sensor IC off-board.

>> No.2048844

Should I use a boost converter for a DIY soldering station?
The tip requires 24V but I only have 20V.

>> No.2048846
File: 54 KB, 1132x890, supposed to be a circle.png [View same] [iqdb] [saucenao] [google]
2048846

ok this vector drawer needs a little work

>>2048844
not advised for efficiency reasons, you'd be better off getting a higher voltage and just bucking (or PWMing) it down

>> No.2048877

>>2048763
the SOT23-5 opamp pinout is nicely standardized. at minimum, bring power, put the "usual" couple of placements for your lab's standard size passives between each lead of the opamp, a few single pin headers or wire wrap posts in strategic locations, maybe some jumpering accommodation for unused op amps like cutting the power or voltage follower configuration
I've thought of ordering a few hundred little "adapter plates" of a similar design for breadboarding

>>2048846
1/10 trollface. asian dad is disappoint with love

>>2048844
you're only losing 1/3 of the power by running at 20V

>> No.2048903
File: 84 KB, 1090x836, something.png [View same] [iqdb] [saucenao] [google]
2048903

>>2048846
>>2048877
It's getting slightly better, but I've still gotta work on getting it to ignore superfluous rogue points better.

>> No.2049010
File: 175 KB, 1100x844, help.png [View same] [iqdb] [saucenao] [google]
2049010

progress is slow
sleep is needed
fuck corners having overlapping pixels

>> No.2049052
File: 63 KB, 307x239, conn.png [View same] [iqdb] [saucenao] [google]
2049052

What is this type of connector usually called?
Looking for breakouts but don't know the term.

>> No.2049054

>>2049052
mezzanine

>> No.2049076

>>2049054
Turns out it's a Hirose BM30B(0.8) 40DP-0.4V(53)
No breakouts for it and I suck at soldering

>> No.2049207

I'm looking for a component with the following characteristics but I don't know what they're called:
>behaves like a normal rotary potentiometer
>NO detents
>i can push it like a button
>the button is spring loaded
Think of a car radio volume control where you push it once to turn on the radio, can change volume, and push again to turn off the radio.

I want to use two of such devices for a colorful lamp project:
One for on/off function and brightness,
and the other for mode selection and to manually cycle through and select a color.
I want to achieve this with only those two UI controls so that I don't need extra switches/buttons/etc.

>> No.2049216
File: 86 KB, 942x457, push on push off pot.jpg [View same] [iqdb] [saucenao] [google]
2049216

>>2049207

push on push off pot

>> No.2049220

>>2049207
your search term is "momentary switch potentiometer"
they're rare and a bit expensive, if you're not ordering 100k of them, and their rotational lifetime is pretty limited, so not suitable for intensive use like a gaming application
https://www.digikey.com/en/products/detail/bourns-inc/39LB-1PB-103/954391
Mouser has a cheaper Alps part number RK0971114Z07, which is out of stock for the next 3 weeks or so

>> No.2049240

How do I find out the max current that a choke can handle? I would like to use one that I got in a free parts box that I would like to use in a power supply for a project. I tested the inductor using one of those cheap multi-part testers and it measured 8.3H, but I don't know how it will actually perform in circuit. I tried googling the part number, but had no luck. The coil has a DC resistance of roughly 330 ohms.

>> No.2049246

>>2049207
There's probably some rotary encoders that apply to this, ones that are optical or magnetic. Assuming you have no issue with using 2 digital inputs as opposed to a single analog input.

>>2049240
Firstly, there's a difference between chokes and inductors, which you should look into. Testing whether your inductor is a choke or not should be possible with a scope. Secondly, the maximum current through any magnetic circuit element will be governed by both the magnetic saturation of the magnetic core, and the thermal limitations of the wire wrapped around it. 8.3H is a massive inductance to use in an SMPS, usually they'll use 10µ-10mH or so, depending on frequency. 330Ω is also pretty high, but that's to be expected of something with that many turns.

I'd simulate any design before you build it, especially with SMPS designs. Make a crude square-wave powered no-feedback buck or boost converter, and see what frequencies it works well at, and whether there are significant thermal issues.

>> No.2049250

>>2049240
>The coil has a DC resistance of roughly 330 ohms.
so it's a 330Ω resistor
>How do I find out the max current that a choke can handle
start with Ohm's Law, namely, V=I*R. at a current like 10mA the voltage drop across that 330Ω is 3.3V
>in a power supply
the only realistic task is to protect a zener voltage reference and perhaps other linear control circuitry from high-frequency, high-power noise that might exist on the input rail. it has no rational place in the power path

>> No.2049284

General question. Most of my lights at home are LED. I notice some of those lights flicker when I run a 1,500 watt appliance. An electric heater for example.

Does that flickering cause any problems for any of my electronics? Modem, router, etc..?

>> No.2049295
File: 3.16 MB, 4032x3024, 20210307_193219.jpg [View same] [iqdb] [saucenao] [google]
2049295

Is there a name for these connectors other than IDC connectors?
I'm looking for a panel mount female version of this.

>> No.2049344
File: 370 KB, 891x668, 34-40-pin-ide-cables-5ac639ed3128340037a96ce7.png [View same] [iqdb] [saucenao] [google]
2049344

>>2049295

get IDE extension. use glue to make it panel-mounted.
GPIO extensions for the Raspi are similar.

>> No.2049346

>>2049284

brown-outs can cause gadgets to crash and do unpredictable things. also, turning gadgets on/off decreases life-span. i'd guess turning them half-way on/off causes the same kind of current spikes.

>> No.2049364
File: 92 KB, 897x916, 61GVKfbjaXL._AC_SL1000_[1].jpg [View same] [iqdb] [saucenao] [google]
2049364

Any recs for a beginner repairing handheld electronics as far as vise/board holders? I've pretty much been using a silicone mat and its friction as the place to hold my boards and I'm feeling like it's time to upgrade to something more reasonable.

Was thinking of one of these, any good?

>> No.2049373

>>2049364
also, I hav question about capacitors. I'm looking at recapping a sega game gear. It looks like the ones to get as replacements are aluminum electrolytic capacitors. I have a list of caps needed (by voltage/capacitance) and was going to go buy some from digikey, but then I pull up nichicon and it lists this further breakdown of product lines with a whole list of 3-character series designations. Am I needing to worry about the series or product line or that kind of info, or will buying correct capacitance/voltage rating be enough?

>> No.2049395

>>2049295
>I'm looking for a panel mount female version of this.
You could just get a female header and put it on a PCB with mounting holes.

>> No.2049451

Hello ohmbros.
Threadly reminder that I love you guys.

>> No.2049484

>>2049364
Not sure about those, probably not too good for soldering heavy things like transformers, or in general if you need a substantial amount of force out of your tip.
I’ve heard Panavice vices getting recommended here.

>>2049373
The higher specifications will have better temperature handling capacity, lower tolerances, lower ESR, and/or lower thermal drift. These are features that might be useful for an audio or radio circuit, but for something digital like this it probably doesn’t matter. Unless the caps you’re taking off are marked to indicate that they’re low-ESR or whatever, in which case replace those particular ones with a matching spec. You can always use higher-spec caps to replace lower spec ones, but not the other way around.

>>2049451
*ohmlets*

>> No.2049487

>>2049484
>*ohmlets*
Wait is that how we call ourselves now? I swear there are so many different versions. Kinda makes us sound dumb, hey wait a min--

>> No.2049491

I wonder, will feeding an audio signal into a transformer that powers EL wire give me a usable result? Will likely need a reasonable amplifier on the LV side, but I think it’s worth a shot. 100:1 ratio or so.

Also I want to use the EL wire as the antenna for an RDA5807M circuit, but baby steps.

>> No.2049504
File: 150 KB, 788x453, conn.png [View same] [iqdb] [saucenao] [google]
2049504

>>2049344
>>2049395
I found these listed under "backpanel connectors" so I'm gonna give them a go, just bolted onto my panel

>> No.2049506

>>2049487
ohmosexuals

>> No.2049513

>>2049506
Would that imply everyone here is gay...
Y-you're all males right...?

>> No.2049514

>>2049220
This is what I needed, thanks.
>>2049246
>>2049216
thank you for the responses.

>> No.2049516
File: 615 KB, 1280x1500, tatsu.png [View same] [iqdb] [saucenao] [google]
2049516

>>2049513
We're all cute girls here anon...

>> No.2049518

>>2049516
O-Ohmnee-chan...

>> No.2049523

>anime posting
I thought you guys were all old and experienced boomers

>> No.2049526

>>2049373
>a whole list of 3-character series designations
yes, ideally you would look up the old series code, usually printed on top of the can, and try to locate an equivalent in the Nichicon line. recapping's starting to get a bit harder now that equipment of a certain age is starting to use switching converters

>>2049491
>usable
maybe if you filter it to the appropriate frequency. can't say what transients in the music will do to your EL wire tho

>>2049504
I'm not sure if you mean backplane connectors but those are good. maybe you can find those in a right-angle configuration

>>2049523
4chan is almost old enough to buy cigarettes in some states

>> No.2049561

>>2049506
>ohmosexuals
>(ohmo)sexuals
>(homo)sexuals
WHAT THE FUCK

>> No.2049573
File: 68 KB, 152x233, 1553523805316.png [View same] [iqdb] [saucenao] [google]
2049573

>>2049561
ohmbre

>> No.2049576
File: 53 KB, 600x591, 1596156145147.jpg [View same] [iqdb] [saucenao] [google]
2049576

>>2049573
I hate you so much

>> No.2049579
File: 63 KB, 158x230, 1553523805317.png [View same] [iqdb] [saucenao] [google]
2049579

>>2049576
that's raycis

>> No.2049581

Never change my ohmmies

>> No.2049591

>>2049526
>maybe if you filter it to the appropriate frequency
I have heard that they’re best around 1kHz. I don’t think they can be damaged at below that, since night lights have run them on mains for decades, but I’m not sure about above that. My understanding of the mechanism behind them suggests not, they’re just capacitors with a lossy dielectric after all, but I’ll be sure to read some data sheets. Also I’ll look at Applied Science’s EL driving code to see what kind of frequencies he’s using with his dot matrix. I don’t think the ESL will limit their current since they wire is basically coax, hope it doesn’t resonate with my transformer.

A quick bit of research and I’ve yet to see someone driving one with an AC transformer. Doesn’t help that “transformer” in common vernacular means “something that converts voltages”, so I’m ending up with a slew of kHz HV switched-mode inverters in my search results.

>> No.2049610
File: 1.09 MB, 2000x1183, fluke.jpg [View same] [iqdb] [saucenao] [google]
2049610

"kill me"
-fluke 8840a

>> No.2049649

>>2049648
>>2049648
>>2049648
NEW THREAD

>> No.2049668

>>2049207
>One for on/off function and brightness,
>and the other for mode selection and to manually cycle through and select a color.
you can do this with an encoder and parts will be cheaper.

>> No.2049671

>>2049295
have you tried a parametric search on your favorite catalog site for the connector parameters
>double row
>x pins
>x pin pitch
guarantee you will find what you're looking for