[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/diy/ - Do-It-Yourself

View post   

[ Toggle deleted replies ]
File: 33 KB, 640x640, 1578777362404.jpg [View same] [iqdb] [saucenao] [google] [report]
1765922 No.1765922 [Reply] [Original] [archived.moe]

>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:

>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:
NI Multisim
iCircuit for Macs
KiCAD (PCB layout software, v5+ recommended)

Mouser, Digi-Key, Arrow, Newark, LCSC (global)
RS Components (Europe)
eBay/AliExpress sellers, for component assortments/sample kits (caveat emptor)
Local independent electronics distributors

>Related YouTube channels:

>Li+/LiPo batteries
Read this first: http://www.elteconline.com/download/pdf/SAFT-RIC-LI-ION-Safety-Recommendations.pdf
>headphone jack noise
>I have junk, what do?
Get rid of it.

>> No.1765967
File: 5 KB, 400x400, tegaki.png [View same] [iqdb] [saucenao] [google] [report]

quick question
i am going fucking crazy from my led bulbs flickering at 50hz with the mains
will adding a big cap across the L and N wires stop the flickering?
since big caps are polarized i would do it like this
Mains--->bridge rectifier--->led bulbs and the capacitor in parallel

>> No.1765974
File: 179 KB, 1600x1200, Xiaomi Yeelight 28W Round LED Ceiling Light.jpg [View same] [iqdb] [saucenao] [google] [report]


your best bet is to buy a diff one that doesnt flicker. i recently bought some 800 lumen ones in the usual E26 format, and they're cheap, bright, and flicker free. the local hardware store has demo lamps so you can look for flicker. i've never seen any.

the problem with caps is you gotta figure out the entire circuit. it's typically some series/parallel combination. and you dont seem astute enough to be able to trace that out. there's also a size problem: high capacity caps at high voltage are big, and may not fit in the housing, or will block the light.

>> No.1765975

>figure out the entire circuit
why? i already know the circuit it's two mains wires and all the led bulbs are in parallel
i don't want to put caps inside of the bulbs i want to put it on the wires that supply the the power to the bulbs which are in the ceiling that can fit even a huge array of caps

>> No.1765976 [DELETED] 

>it's two mains wires and all the led bulbs are in parallel

there's a .00001% probability of you being right.

>> No.1765979

>big caps are polarized
They have these things called "fan capacitors," which are about 2-4uF, unpolarized, and mains rated

>> No.1765982


oh, you're talking about individual bulbs, not fixtures. i was unsure of they'd work in DC but i just tried 2 of them and they do, so i guess your plan has merit.

>> No.1765988

one thing to consider, if you add a cap to smooth out the sine then dimming won't work anymore

>> No.1766102
File: 363 KB, 1062x1375, 1574780193460.png [View same] [iqdb] [saucenao] [google] [report]

this thread's digits brought to you in part by the MAX5922 PoE power switch

thank you fren

>> No.1766109
File: 231 KB, 658x1696, DM74LS469.jpg [View same] [iqdb] [saucenao] [google] [report]

Do these output buffers look okay to you guys?

Also, are there non-obsolete alternatives for this ic?

>> No.1766112

what's the bump limit on this board?

>> No.1766117

I know it says to get rid of junk, but I dumpster scored 40+ laptops 40+ dell lcd monitors and hundreds of electronical gadgets last month.

I’ve taken out the hcfl(?) bulb strips, and controller boards, as well as the actual displays from a handful of the lcds.
Turned them into quick connect displays.

Also the laptops are Fujitsu touchscreens. With those of which are i5 I’ve loaded Linux into and sold, but what to do with rest?

I’ve scavenged a dozen and more lithium batteries as well as lead batteries from backup power boxes etc.

Anyone have any ideas for this?

>> No.1766128

you can do a lot of cool things with old laptop batteries like building portable battery packs for kfz appliances for example using it with an inverter if you are willing to put a lot of time into it

>> No.1766129

>output buffers look okay
wait, for what?


you'd get better results asking /g/ what to do with consumer products. we component-level engineers here

>> No.1766131

For disabling the outputs Q0 to Q7.
Only the buffer of Q0 seems to be connected to nOE. Why?

>> No.1766134

>non-obsolete alternatives
if you want to reimplement that identical functionality, you'll need to combine multiple chips or use programmable logic. e.g. everything to the right of and including the CK vertical looks like a '374, everything to the left can be made of an 8-bit counter (or two 4-bit counters) like the '163, maybe a few loose gates to translate control signals
in the classic 74 series, oddly specific chips like that were usually first to go out of print. not necessarily so with the CD40xx series, maybe there is something closer there

oh, that. seems useless, surely a drafting error, defer to the truth table at top

>> No.1766136

>surely a drafting error
I had a hunch like that but you can never know what weird logic they came up with back in the old days

Oh well ill just pay a billion dollars for one on ebay and hope i get that counter and not a nor gate or some shit.

>> No.1766179
File: 25 KB, 800x480, Untitled.png [View same] [iqdb] [saucenao] [google] [report]

I sorted this out. The diode was not the issue. It was due to the high DC resistance of the inductance being measured. Diode or not, can't measure the time constant accurately if R_l is too high compared to R. So I increased the R value and that did that trick. 1:10 should be good enough.

Now I am getting a fairly accurate result on the output of the comparator. Except I am getting a second pulse since the voltage across R spikes above 5V. Damn. Can try a schottky across the L.

>> No.1766180
File: 23 KB, 881x337, Untitled.png [View same] [iqdb] [saucenao] [google] [report]

So interesting, LT spice doesn't predict that second spike at Vin > 5V. This is the exact schematic I've been testing.

>> No.1766181

a tube of PALs and a programmer will only set you back about a hundred or so

>schottky across the L
this is the most correct answer imo
just know there are several pF in those, in case that confounds your measurements. at 0V where reverse capacitance is maximum, BAT54 is only ~15pF, 1N4148 about ~4pF

>> No.1766183

>expecting a gourmet LT/ADI model to act just like your jellybean chinkparts
that's why you should dump LTspice and use Micro-Cap instead: it has the parts you're actually using

>> No.1766205

Well this is the LT1016 from Digi , I don't think the chinks can fake these? Also it is clearly running outside the specs (the common mode range) so I'd rather expect LT spice to catch that. Many of their models are far from the gourmet quality.

> Micro-Cap instead: it has the parts you're actually using
So how does that work? Is it like real time human captcha solvers, when you run online simulations a bunch of chinks are running around the warehouse and picking your parts, stick them into a breadboard, and report results back to the simulation software and then ship you these same parts they just tested? Awesome. I should give it a try.

>> No.1766267
File: 63 KB, 407x497, Untitled2.png [View same] [iqdb] [saucenao] [google] [report]

yay, the schottky did it, the first screen is with the diode and the second one with the schottky. no more extraneous pulses. just what i needed. i also powered the mosfet from 3.3v to leave more room for the spikes and that seems to work fine.

>> No.1766331

Since shottky diodes have smaller voltage drop it means the produce less heat at same current than regular diodes right?

>> No.1766333

but they have their downsides too

>> No.1766335

other thread is way above that

>> No.1766370

well lads apparently i just graduated from shitposting on /ohm/ to designing aircraft power systems.

Some of you are cool, don't take the airplane over the next 30 years.
(idk what i'm doing pls help)

>> No.1766371

I'm trying to design a pcb that needs two power levels: 12V DC & 5V DC.
Input should be a IEC 60320 C13/C14 plug, so (since I'm in europe) 230 V AC mains input.
The idea was two transformers with appropriate windings, followed by a full bridge rectifier and a decoupling capacitor each. I started looking into part sourcing (mouser was my supplier of choice) and I can't seem to find fitting parts. Even if I did, they seem bulky af. Any ideas how I can better rectify and change the voltage to the two voltage levels in a more elegant way? I have formal training so I don't know how you "should" do it or how most plugs for electronic devices (which contain the rectifier etc) do it.

Should I instead rectify the voltage, add a decoupling capacitor and then build a buck step down converter after? If so, what's better: Two step down converters in parallel or one from 230 to 12V DC and one for 12->5V DC?

I'd appreciate ressources or ways to learn about the topic.

>> No.1766374

Personally I'd be honored to die to some mistake by a fellow /diy/ or /ohm/ tard. Godspeed.

>> No.1766388
File: 44 KB, 847x642, 1580027338071.jpg [View same] [iqdb] [saucenao] [google] [report]

>circuit not working
>spend an hours fucking around with my flukewaifu
>turns out the chink LVR i used was fake
>FUCK that means the others i had are as well and now i have to buy them expensive locally and what two days from them to arrive
>another project some months later
>not getting me again sneaky chinks, the IC before soldering it in to make sure it works
>it does
>work but with bugs
>waifuscope is on the job
>hour later
>the ic chinks sent me sents out correct signals on 3 pins but 2 of the pins are dead and not working, causing the bugs in the circuit
coronachan please kill them all, i love you

>> No.1766394

this is why you buy from digikey, poorfriend

>> No.1766403

>I have formal training so I don't know how you "should" do it or how most plugs for electronic devices (which contain the rectifier etc) do it.

Obviously what I meant to say is I have NO formal training.

>> No.1766421

I'm not poor exactly because I don't buy from digikey

>> No.1766423
File: 916 KB, 1845x2262, clip.jpg [View same] [iqdb] [saucenao] [google] [report]

many choices fren
The simplest one is pic related
buy two ACDC bucks from chinks (component not the band)
and solder them to the board, the one in the pic converts 230V AC to 5VDC and can deliver around 0.5A
You can either buy one 5V and one 12V, or you can buy one 12V to get the 12V and use lm317 to get 5V from the 12V buck.
The bucks only cost about $1 each
It all depends on what sort of current you need

If you need several amps it will be easiest to use 12V wall wart LED power supply which are 12V and can do usually around 3A consistently and then you use a DC-DC buck to get also the 5V

And if you think you are hot shit you can of course buy the expensive heavy as fuck transformer and roll your own extremely expensive, with lots of extra work AC->low DC conversion while reinventing the wheel, but ... why?

So yeah, that about covers it

>> No.1766430

Well I buy from both digi and ali depending on which components. Ali is fine for stuff like passive components/generic LEDs/boxes/wire/switches etc. Never had a problem with anything. I would never buy any power components from ali. Like power transistors/rectifiers/etc. Those are all fake and may only work at like 10% of their official capacity. I even bought some STM32 micros on Ali. They all work fine. I don't know where they get them for half the price but those seem authentic. I mean how do you copy an STM32F4? Maybe they sell some rejected batches. Same with DDS chips. Digi can go fuck themselves with their AD9850 bare chip for $30. You could always buy a complete board on Ali for like $12 even back when it wasn't obsoleted by AD9851.

>> No.1766432

micros are hit and miss
out of 30 ATTTINY85 i got from china 10 were fake so that means i still saved more than like 50% of the price if i purchased from jewkey

>> No.1766434

First of all, thanks for the detailed answer. Using buck converters is a good idea, but I want to design the circuit myself. I know that's a dumb thing to do from a labor perspective (since I can buy a module for far less), but I feel like eI'd want to do so anyways.
>If you need several amps it will be easiest to use 12V wall wart LED power supply which are 12V and can do usually around 3A consistently and then you use a DC-DC buck to get also the 5V

Maybe that's indeed the best idea. Get a 12 V power supply that can supply enough current, then implement a 12->5V buck converter into my PCB. That feels like a fair compromise between me still learning something and doing it myself and me skipping a lot of the rectifying etc.

>> No.1766436

>And if you think you are hot shit you can of course buy the expensive heavy as fuck transformer

That's why I asked, using a transformer didn't seem economical nor get me to where I wanted.

>but ... why?

To have fun? Dunno. I just wanted to do it in the circuit instead of with a module. Again, thanks I think using a 12V LED power supply makes the most sense and then lowering the voltage to 5V as well.

>> No.1766448
File: 8 KB, 252x347, 316afWDYUzL._AC_SY400_.jpg [View same] [iqdb] [saucenao] [google] [report]

I got these speakers.
I want to use them but I don't know how, I know nothing about amplification. Still I tried them but the sound is very low when I connected one to the tv.
The label on them says
- output power 10W
- impedance 8 ohm
They have a regular 3.5 jack

>> No.1766459
File: 4 KB, 272x322, 1.png [View same] [iqdb] [saucenao] [google] [report]

What is the best way to achieve a non-intrusive floating e-switch controlled by a micro? SSR or an optocoupler? Needs to be low inductance/low capacitance low resistance, etc as close as possible to a physical switch. ~100mA.

>> No.1766461

Do you want to know about amplification or do you just want your speakers to work?

If it's the first, read up on audio amplifier circuits.

If it's the latter, buy yourself a standalone amp. Most can be plugged between output of your device and input of the speaker.

You may want to look into mini DACs, depending on what audio quality you want these can be really cheap or really expensive.

>> No.1766483
File: 296 KB, 750x466, 9523AF97-7483-4779-B486-B2E3683DDD6C.jpg [View same] [iqdb] [saucenao] [google] [report]

Is $100 a good deal for this?

>> No.1766494

trash, get rd6006 from ali

>> No.1766503

I can’t tell if you’re meming

>> No.1766507

no joke friend, they are excellent, and very cheap just watch some reviews
but you need a dc power supply for it

>> No.1766518
File: 83 KB, 800x645, maxessential01.jpg [View same] [iqdb] [saucenao] [google] [report]

>do want.jpg

>> No.1766552

you need a power supply for a power supply? how convenient. then you can apply this one recursively: >>1766483

>> No.1766603

you need a power brick to feed say a stable 50V into the power supply. By it self that is useless since none of your projects needs unlimited 50V with no overcurrent protection etc. You can even use an old laptop brick to power it for example.
rd6006 does that, it gives you that adjustable front in a modern digital way with a nice clean safe output and of cource CC, OV, OC stuff as well
It can even charge batteries safely. And it comes with a wifi and usb control if you want that
But listen here son, if you want your two shitty analog knobs for twice the price, you go right ahead, i'm not stopping you

>> No.1766704

kek, no, I mean they have common jellybean parts in their library, like 45000 of them

you can still post on a thread that has reached bump limit

switching power supply transformers are custom parts, with the exception of a few common USB-charger-type transformers that are stock items. how many are you building?
>Get a 12 V power supply that can supply enough current, then implement a 12->5V buck converter into my PCB
this is the most practical and usual answer

you relabel some GD chips

relay. anything not electro-mechanical will require compromises and considerations wrt the load

bretty gud

this is 4chan. everyone's memeing

but it's neither dual output nor tracking

>> No.1766724

>Related YouTube channels:

Not posting The Post Apocalyptic Inventor


>> No.1766729
File: 95 KB, 665x819, problem.png [View same] [iqdb] [saucenao] [google] [report]

I need a sensor that will tell me what X is in this picture, either in sub-millimeter distance units or in a unit of mechanical pressure. There are two of these mechanisms on either side of the machine and I need to be able to reliably set them to the same relative pressure frequently.

The pressure will be considerable, maybe up to 200 lbs, but I don't really know until I have someway of measuring it.

The space around X is not large, maybe 5" tall by 4" between the springs by .75" wide. I can mount things to the outside face of the frame though.

The measurement doesn't have to be absolute, it can be relative. I just need an accurate number.

I cannot use any kind of rotary encoder on the cam screw but I may be able to use a linear encoder to measure X as a distance if its resolution is high enough.

What is the best sensor option for this?

>> No.1766770

/ohm/ RULE 0
this isn't a macgyver shit with wires general

linear encoders are available in very high-res flavors. those cheap electronic calipers, for example
but I'm not convinced continuous measurement is the most practical means, as it would continuously provide its own interference in the action and possibly drift out of calibration. I would think about using a torque wrench or cutting a gage block with stepped benchmarks on it, or making some accommodation for a depth gauge off the side, which you could use with those cheap electronic calipers, or even just weld a metal scale on

>> No.1766791

I like it

If I get a cheap scale like https://www.amazon.com/gp/product/B07K7QPD5G/ am I able to trim off the unnecessary extra length of the rail?

>> No.1766889

How would you implement setting a certain input value, such as a frequency, via a pot connected to ADC? I know one could use a rotary converter but lets say you only have pots and buttons or a rotary switch. Would you prefer a rotary switch to set the range, or click a single button repeatedly provided the selected range is displayed on the screen? How many pots? Just one? Or two, coarse/fine? Or scrap the button/switch and only use two pots?
While I've been trying to reinvent the wheel I ran into some rookie problems. For example smoothing the ADC reading with a low pass filter is a really bad idea in this case. As it seems to me the low pass is only good for something that is changing by itself with a more or less predictable rate, such as temperature. But if it is a pot that is turned by hand, thats not gonna work. I haven't tried a moving average but the simple average works fine. However, is there a way to make sure the setting will never change by itself? Some sort of a hysteresis? Cause no matter how you average, there will be some glitches. Even if it is a single unit, depend on how much weight is assigned to it, that could still result in an unpredictable behavior (in a general case, depending on what we are controlling). One way is to reduce the resolution, say from 12 bit to 10 bit and simply ignore the two least significant bits where all the fluctuations occur.

>> No.1766954

Shouldn't you already know this if you took Semiconductors + lab?

>> No.1766962

I would assume so, but you could also get a shorter scale (150mm) and/or play with offsetting it via brackets

first, precede with a 20Hz LP filter to keep the thermal and wiper noise down
put a software clutch between your set value in the controller and the set value on the inputs. for example, engage it when there's more than A amount of change between the input value and controller value, and only disengage when there's been less than B amount of change over the past y samples. you still have "full" use of all the bits of the ADC you want

>> No.1766965

I need to send data using UART.

Are there separate DIP ICs that you can buy, or do I have to use a microprocessor? I must to be able to send 31.25 kbit/s.

>> No.1766987

Post apocalyptic inventor is based. He made a few videos about recycling washing machine motors for diy projects like electric carts and wind turbines, and had some lecture videos as well for the theory.
All around stand up guy

>> No.1766994

do you think desktop computers can stay alive long enough to survive being switched to back up power via mechanical relay?
relays are pretty fast and computer psus usually have nice juicy caps so i feel like they should

>> No.1766998

the better question is, can you detect a line outage that quickly?

>> No.1766999

>type "relay" into google to look at some schematics
>okay time to procrastinate
>go to youtube.com
>4 videos in my recommended feet are about relays

the big brother gives zero fucks and isn't even hiding anymore

>> No.1767000

Yes, simple AC optocoupler where mains is blinking the IR leds in it, which gives me nice logical pulses for my attiny (100 times per second) and if i detect no pulses for a milisecond or two i slam the relay shut, so i would say about 5ms between the power going out and the relay being fully closed

>> No.1767083
File: 214 KB, 1115x983, Screenshot_2020-02-13_17-21-19.png [View same] [iqdb] [saucenao] [google] [report]

I'm looking for button caps for tactile switches. I have a lot of those 6x6x4.3 tact switches but no button caps. I'd like to find some sort of a list of available caps models like picrel (but a complete one, this seller has only the models which he sells). Not just for my 6x6x4.3 switches (I could buy a complete set of switches+caps after all), but also for future reference

>> No.1767086

get a 3d printer and make your own custom ones for less than cents

>> No.1767087

you could've just said "fuck you"

>> No.1767110

I am serious, 3d printer is a mandatory accessory for electronics diy and you can get them for really cheap from chinks starting at like $200

>> No.1767142

this is a very strongly correct answer

no real value in designing around cap types you can't source

yes, you can! try here https://www.nxp.com/packages

>> No.1767150

In years past you could get a big DIP IC and a crystal and some TTL logic and do this.... these days it is just so much easier to use a microcontroller. Besides, if you have significant amounts of digital data to send, surely you have a microcontroller or CPU somewhere in your design?

>> No.1767204

ah yes, so it is like a long term investment and return? he needs a few button caps or something that he can buy for a few cents. but you are suggesting a 3rd printer. and a decent one is much more expensive than $200. you'd need to print like a thousand caps to justify the investment. i understand it may be a fun hobby by itself, you know drawing your own shit if you are into that. but hardly a cost saving tool.

>> No.1767215
File: 88 KB, 1000x1000, 20JM92_AS01[1].jpg [View same] [iqdb] [saucenao] [google] [report]

>At customer site doing some mechanical service work on a machine they bought from my company
>Suddenly the safety relay stops working on another machine
>"Anon your company sells this machine. Fix it."
>Spend 5 hours trying to make it look like i'm troubleshooting while i teach myself how electrical circuits work
>"Uhhh i think it's broken"
>"Okay good work anon we'll order another one."
>Customer bossman tells his electrician to bypass the whole system safety and start the machine right back up
>3 week lead time for replacement relay
>Probably still won't work when they put the new one back in
Damn i need to learn electronics.

>> No.1767238
File: 461 KB, 3000x2000, IMG_3551 (Custom).jpg [View same] [iqdb] [saucenao] [google] [report]


I have my grandfathers quad 33 & 303 amps and they sound terrible. How viable would restoring them with these kits be for a relative beginner? I've -badly- made some cables before and that's about it.
Also what tools would I want?

The KEF speakers they came with sound fantastic and i'd consider just doing the power amp and running it with a passive pre-amp or something because I don't need a phono stage

>> No.1767240
File: 434 KB, 3000x2000, IMG_3552 (Custom).jpg [View same] [iqdb] [saucenao] [google] [report]

They do work, they just sound like shit, as in, not -ok- but terrible.

>> No.1767241
File: 266 KB, 3000x2000, IMG_3543 (Custom).jpg [View same] [iqdb] [saucenao] [google] [report]

>> No.1767242

Wew lad

Dubious post

What data? How are you planning to get it into a form ready for sending over UART? What is the sending device/what is the receiving device? What baud rate and other UART parameters?

>> No.1767243
File: 252 KB, 3000x2000, IMG_4144 (Custom).jpg [View same] [iqdb] [saucenao] [google] [report]

these are a DIY job from the 60s with KEF drivers they really do sound great already even with a cheap integrated amp so iunno what difference the Quads would even make but i'm curious and they're cool anyway even if it wasn't much and I could probably sell them restored if it's nothing and they're annoying to use.

>> No.1767272

I wouldn't say it would be too difficult. But you'd need to have a good desoldering setup, so a good (probably temp-controlled) soldering iron, a solder sucker and quality wick, extra flux, and tweezers and/or small needle-nose pliers.
Try not to put the iron on a joint for more than ~5s or so, if the solder doesn't reflow after that then take the iron off and let the board cool down. Chances are the problem will be either that your temperature is too low, or that you're not making good thermal contact with the joint. As fluids conform to the shape of things, a small amount of fresh solder on your tip will touch the joint and make far better thermal contact than a dry tip, and will help you get in and out fast. Throw a little flux on the joint before desoldering, it will cause the solder there to bead up nicely and hopefully avoid the hole so you can shove the replacement part in there without any extra work. Higher temperature should be self-explanatory, if you are making good thermal contact and the joint is melting slowly, turn the temp up by 50° or so. This will often be necessary as a temporary measure when working on large parts or ground-planes, which is where the boost function on some soldering stations comes in handy.
1-2 pin components are easy to desolder since you can reflow one leg and lift it relative to the other, but a 3+ pin component will likely be too rigid for this. You can try to use a solder wick and a solder sucker and flux to suck up all the solder around each joint such that you can just lift it out, or turn the temp up a little and run the iron back-and-forth across all the pins to try and get them all molten at once, but both methods will result in probable prolonged heat on the board and potential damage to the pads, so I'd just cut the leads. Hope you don't need to desolder any DIP ICs.

Flux is your friend, but you'll likely want to clean it up after soldering/desoldering, so some isopropyl alcohol would be good to have.

>> No.1767275

hmm ok thanks anon that's helpful advice
I might try to find some project with lower stakes to get a bit of experience first so i can do a better job on those amps that i care about

>> No.1767277

Yeah, it's good to try some other projects just to get a feel for soldering. What I described is just my experience, and everyone has different styles to one extent or another. I'd recommend some little practice boards or cheap kit projects, since soldering on a PCB is somewhat different to wrangling with perfboards. Your amp looks to have single-sided boards (no surprise there) so go for single-sided boards for practicing too. While 2-sided makes for more robust solder joints, it does make them much harder to desolder, so you're lucky in that respect.

>> No.1767302

>yes, you can! try here https://www.nxp.com/packages
No thermal information included, nice source though.

>> No.1767339
File: 620 KB, 2048x1536, IMG_20200213_183957.jpg [View same] [iqdb] [saucenao] [google] [report]

It saves lot's of money. Look at this case i made for my raspi 4. The price of material is like 50 cents. Stores here sell plastic cases like that for around 10 or 15 bucks.
And you can modify it if you need as well.
And no you don't need a "better" printer like the overpriced průša. My printer cost like 250$ from chinks (came fully assembled too and is fully metal) and i have been using it for years
It is so fucking useful to have one

>> No.1767367 [DELETED] 
File: 79 KB, 1500x1500, 12V_5V LaCie Hard Drive F.A. Porsche v.2 External Hard Drive Replacement Power Supply Adaptor.jpg [View same] [iqdb] [saucenao] [google] [report]


i've found a ton of these kinds of 12V/5V power supplies at the thrift store coz they were used in older external HDDs and DVD readers.

>> No.1767372
File: 79 KB, 1500x1500, 12V_5V LaCie Hard Drive F.A. Porsche v.2 External Hard Drive Replacement Power Supply Adaptor.jpg [View same] [iqdb] [saucenao] [google] [report]


i've found a ton of these kinds of 12V/5V power supplies at the thrift store coz they were used in older external HDDs and DVD readers.

cost = $2-$3 each
time cost = 5 minutes

>> No.1767420

pssst, hey, here, listen here boy, use mains capacitive dropper, you're welcome

>> No.1767512

fuck your power factor

>> No.1767637

I see 4/5 pins on that, is it ±12V as well as 5V?

Why would you care about power factor? Unless you're being charged for it, which you probably aren't.

>> No.1767701
File: 10 KB, 251x200, 1576762649094.jpg [View same] [iqdb] [saucenao] [google] [report]

bad engineering must be punished

>> No.1767704

just use an inductive dropper on your other power supply, they'll cancel out

>> No.1767770

Do you guys have any links for info on am and/or fm radios? I've been dicking with audio amplifier circuits for a few years and the idea really interests me. Everything I find on google is unrelated or more step by step with "upcycling" with no theory. A lot of them are just
>get old radio
>cut amplifier and power circuits out
>recreate them
>wa la, you have a bastardized version of what you started with

>> No.1767772

Basically I'm looking for info on the tuner and receiving radio signal

>> No.1767784

What color does the CEM1 base material have?
Is it just white?
This is important.

>> No.1767786
File: 97 KB, 599x396, 1565658525310.jpg [View same] [iqdb] [saucenao] [google] [report]

So lets say I want to take on the task of DIYing myself an inkjet printer...
Assuming I can get everything else figured out, is there any way I can just go and buy a print head module or is that all proprietary bullshit?
I've heard there's some fancy MEMS-based tech out there but that might be too state-of-the-art for me to get my hands on.

I'm not too fond of the idea of trying to fuck around with nozzles and fluid pressure if I can avoid it.

>> No.1767789

Look into the crystal radio crowd, they make AM radios with minimal part-count with maybe 4-20 different circuit topologies out there all based on a diode envelope detector. I'm not sure if a superhet receiver makes sense these days when you can just have an op-amp that works up past 2MHz; with which you'd simply add an envelope detector circuit after the amp, which itself is after the tank circuit. Personally I'd like to make a direct-mixing AM radio (i.e. no IF and no envelope detector) using a PLL (CD4046) running off the RF signal to get a sinusoid at the carrier frequency, phase shifting it if required, and shoving that into a mixer (SA602?) along with the RF signal, followed by a low-pass before the audio stuff. And an AGC circuit I guess.

As for FM, the simplest is to just use a PLL and feed the RF signal to its input, and take the VCO input signal , filter it, and use that as your audio signal. But doing that loses you stereo encoding and other gimmicks. Another simple method is to run it through a filter such that higher frequencies have a different amplitude to lower frequencies causing you to have an AM+FM signal which you can feed through a diode detector or any other manner of AM demodulators.

But the method I'd most recommend is to buy an existing demodulator IC, as dedicated ICs are usually pretty good at what they do. Like the cheapy RDA5807M, that does have to be controlled via I2C, but can receive the FM station information (RDS). Note that most of them are FM only, with a few being AM only, there aren't that many FM+AM ICs. At least not that I remember seeing.

And then if you want to go full ham (heh) you go for SDR, software designed radio. Mix your local ocillator with the RF signal twice (one of which uses a phase shift on the local oscillator), then feed those into a pair of ADCs after filtering. Take your two signals (I and Q) and run them through DSP algorithms to decode AM, FM PSK31, QAM128, you name it, at any damn frequency.

>> No.1767813
File: 71 KB, 406x600, 6292996i.jpg [View same] [iqdb] [saucenao] [google] [report]

Would you guys recommend getting an used power supply (one of these 24V, non adjustable ones)? I intend to couple it with a voltage regulator and current regulator circuit to make a 'lab' power supply for cheap.

>> No.1767816
File: 14 KB, 474x234, 1576628834349.jpg [View same] [iqdb] [saucenao] [google] [report]

hmm, I guess I could imagine thermals being die-specific

if you understand the properties of the signal you're receiving, you can work out a way to isolate it, detect or discriminate the changes that encode the information, and convert them to a voltage which you then amplify
but there are a few common configurations for detection. one simple discriminator for FM involves a filtered, limited IF frequency of x MHz coupled into an LC tank tuned to x MHz + m*bw where m is a small multiple (1 < m < 5) and bw is the bandwidth of the channel. the impedance of the tank increases monotonically as frequency approaches the resonant frequency of the tank, which yields a higher voltage output, corresponding to the input frequency
there are still some single-chip broadcast AM/FM ICs around, and some analog front ends capable of other frequency bands

the usual laptop brick should be alright but some laptop supplies have been known to require authentication before releasing their high-energy goodness
if not the usual laptop brick, show picture of power supply

>> No.1767817
File: 304 KB, 491x490, psu.png [View same] [iqdb] [saucenao] [google] [report]

A powerful 156W
I think it is a very good deal, I'm just worried if these kind of things should be bought used.

>> No.1767818

oh, those, ok. how used would be a good question. just know they might have been screwed into a machine that got broken down before it saw power, or they might have had several thousand hours on them
for something like a bench supply I'd think age is not a very big deal. it's like semi-retirement, turn off when not in use

>> No.1767821
File: 589 KB, 1620x930, 5V LED.jpg [View same] [iqdb] [saucenao] [google] [report]

I'm finding it hard to believe that there's no such thing as non individually addressable 5V LED strips but that's apparently all I'm able to find. All I want is an RGB LED strip I can feed off of 5V, I don't need them to be addressable since I just want a single constant colour for the end result I'm trying to get.

Am I going about this the wrong way or something? All I want is to hook up a strip to USB power and some resistors to get the specific colour I need, but everything I can find requires a microcontroller of some sort.

>> No.1767825

>Why would you care about power factor? Unless you're being charged for it, which you probably aren't.

fuck off harambe go hug trees on your own time

>> No.1767830

I bought a 5V RGB LED strip on ali. With a remote.

>> No.1767832

I don't want the remote, though, that's the thing. The ideal case is I plug a USB cable into a wall charger and it turns on; that's it, nothing fancy, no changing lights or patterns or anything.

Also, your link prompts me to log in for some reason and I can't get around that.

>> No.1767834

What the hell are you supposed to do with an RGB LED strip without a remote? How would you control the color? 3 knobs? Well I wish they made those too, but I'd just make your own with a quad comparator IC. And maybe a toggle latch hooked up to an IR receiver if you want to do that too. There might be ones that you control via wifi or BT, but they sound like bloatware.

>> No.1767835

>How would you control the color?
I set it, once, when I install them.
This is for a decorative piece that will be lit up in one single color, permanently.

>> No.1767840

Then why not just buy a single-color LED strip? Or buy one with a remote, chuck out the controller and replace it with a couple of series resistors encased in epoxy?

>> No.1767913

find me a single aquamarine led strip (that isn't jarring toothpaste colored)

>> No.1768029

Yeah, this, pretty much. There's only a few "default" colors of LEDS (usually red, green, blue, purple and maybe a couple of others) but I'd like to tune these things to specific colours.

>replace the controller with series resistors
If it's that simple, I'll do it, but I don't really know how to find out what resistors to use on an entire strip at once, as opposed to individual LEDs like I'm used to. Help in this case would really be appreciated. Plus, again, for some reason the 5V ones I've found are only individually addressable which AFAIK makes it more complicated than "apply power, get color".

>> No.1768126
File: 72 KB, 1170x889, 1550685980701.png [View same] [iqdb] [saucenao] [google] [report]

they're rare

>single resistor
possibly poor repeatability if you're doing a large installation. you'd also need a very large power resistor
you might achieve better color fidelity and blending by getting an aquamarine strip and pulling it in the green or blue direction you want with a second strip
you could also create PWM with a 555, the hobbyist literature is thicc with documentation on this solution and variations, Pic related being just one
or, since the remote control receivers will usually remember their last color value across power cycles, just set it once, (optional) remove the IR receiver, and throw the remote into a drawer

>> No.1768154 [DELETED] 
File: 1.56 MB, 320x180, 6nElBtu.gif [View same] [iqdb] [saucenao] [google] [report]

I want to give my ESP8266 32io pins. What IC should I look for? I want to read key matrix with it.

>> No.1768195

>large installation
I'm going to use a meter at a time at most, so not very high power in this case

>> No.1768197

>poor repeatability
I don't think this would be much of a concern.
>a very large power resistor
It's a USB LED strip, they run at 10W maximum. Meaning each channel runs at 3.3W, so even in the worst case scenario a series resistor will only need to dissipate a quarter of that, or 0.8W. So a 1W resistor will certainly do the trick, and chances are a 0.5W resistor will be fine too.
I'm planning on doing the same thing to my LED strip to get a nice 2000K warm white, and I'll do so by setting the PWM brightness via remote to a low level and using my 1/4W resistors (or a pot/trimpot) to set the color before buying the same value power resistor to install permanently. Then I'll just solder the whole thing up to one of those 2-pin micro-USB ports and encase it in epoxy.
I'm spurred to do this because lately some RF from somewhere has been interfering with the remote control and setting it to flash bright colors at me even when the remote is unplugged.
Wow nice work having to use 3 fucking DIP8s or a DIP14 and a DIP8 instead of a single DIP14 quad comparator. Plus those diodes mean your time constant gets messed up a little, to say nothing of the immense quiescent current, horrible pinout, and poor output voltage.

>> No.1768198

If you end up going through with this I'd love to see what you end up with; never really dealt with things like this before and learning by example for me helps a lot.

>> No.1768205

where could I find a fullframe camera sensor for sale? ie a Sony IMX251
just idly wondering about making my own digital camera back, searched element14 but couldn't find anything

>> No.1768240
File: 613 KB, 730x748, signs of autism opinions.png [View same] [iqdb] [saucenao] [google] [report]

>Get free PC
>a couple caps are blown but I turned it on a while ago without issue
>turn it on one last time so I can verify that it works before I replace the caps
>power on
>hear a faint hissing noise
>magic smoke venting from the PSU
>cut the power
>hissing stops
>open it up
>now 8 blown capacitors, all the KZG brand that blew in the other computers, and 8 that are the exact same type but haven't blown yet
>capacitor plague strikes again
At least my autism supercluster will make a decent space heater when it's done.

>> No.1768252
File: 129 KB, 912x876, 1559063602857.jpg [View same] [iqdb] [saucenao] [google] [report]

>I don't think this would be much of a concern.
>I'm planning on doing the same thing to my LED strip to get a nice 2000K warm white
enjoy your CRI of like 50

>> No.1768254

>enjoy your CRI of like 50
I already use it like this but on the remote. You get used to it.

>> No.1768258

it's not a stroboscope, it's a decorative lighting controller. as you say, precision isn't very important. it sounds like he only needs two primaries, in which case only one PWM is needed, just run the brighter of the two primaries at full duty cycle

I see that 2000K is hard to find. I'd probably start thinking about scooping the phosphor out of one COB and doubling up on the other

by large installation, I meant more than one item in a room. I'd want matched hues in that situation

>> No.1768290
File: 27 KB, 1825x626, screenie.png [View same] [iqdb] [saucenao] [google] [report]

I have a board set up with 4 CB4021Bs reading from 31 switches. I have an issue where it looks like the pin readings are offset by 1 on each chip. Pic related, it's a shot of 2 (zooming out to show all 4 gave shitty pic quality.)

So, with the following code it means that the switch on PI-1 (pin 7) should be the most significant bit 0x80, but it is read in as the second most significant bit 0x40. It also means that the switch on PI-8 (pin1) is supposed to be the least significant bit, but is read as the most significant bit on the next IC in the chain.

My first thought was an issue with clock speed, but it behaves the same way with only one of the ICs attached to the board. Otherwise, I have no idea what the issue is. Any thoughts?

Also, this is the same board as from >>1763570

int _CD4021B_dataPin = 5;
int _CD4021B_clockPin = 6;
int _CD4021B_latchPin = 7;

void setup() {
pinMode(_CD4021B_latchPin, OUTPUT);
pinMode(_CD4021B_clockPin, OUTPUT);
pinMode(_CD4021B_dataPin, INPUT);

void loop() {
int data1 = 0;
int data2 = 0;
int data3 = 0;
int data4 = 0;

digitalWrite(_CD4021B_latchPin, LOW);
data1 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
data2 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
data3 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
data4 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
digitalWrite(_CD4021B_latchPin, HIGH);

if (data1 || data2 || data3 || data4) {
Serial.println(data1, BIN);
Serial.println(data2, BIN);
Serial.println(data3, BIN);
Serial.println(data4, BIN);
delay (300);

>> No.1768298

How are you timing your latch signal with respect to your clock signal? I'd consider adding a delay between it and the clock shifting. Also see if putting the "digitalWrite(_CD4021B_latchPin, HIGH);" before the digitalWrite(_CD4021B_latchPin, LOW); helps, with a digitalWrite(_CD4021B_latchPin, LOW); in the setup, assuming that it's needed.

>if (data1 || data2 || data3 || data4)
I didn't know you could do that kind of weak typing.

>> No.1768306

> How are you timing your latch signal with respect to your clock signal?
What you see is it - shiftIn is an arduino builtin function for working with the CD4021B.

Tried those suggestions, no dice.

> I didn't know you could do that kind of weak typing.

As far as I know, it's basically C.

>> No.1768309

>Tried those suggestions, no dice.
Even adding delays everywhere? Do you have a scope or logic analyser to check the clock and latch signals? If not, adding really long delays between each step (500ms) and looking at the lines by putting LEDs on them should work. The $7 USB logic analyser I bought turned out to be very useful, even alongside my scope, if you're in the market for such a thing.
What's in the shiftln function? Did you try replacing it with sequential digitalreads and digitalwrites?

Though it's not likely due to how repeatable this issue is, you could be skipping clock pulses due to unfortified power rails, you have bypass caps on your IC(s), right?

>> No.1768336

I think the problem for something like that is going to be not only sourcing the components but also finding documentation for stuff like that.

>> No.1768357

Sure, GwInstek is a pretty good company. prolly worth $150 used

>> No.1768358

the clock might be active on the edge opposite the one you think it is (aka an SPI mode mismatch)

>> No.1768362

also check the truth table, while P/~S is high PI-8 passes thru to Q8 immediately

>> No.1768370

Right, I'm a bit of a noob so
Shouldn't both of these be covered by the arduino's shiftin function, unless I screwed up the pins? If not, can you please elaborate?

>> No.1768372

On the assumption P/~S is the parallel/serial control - I'm only reading when the P/S-C (latch pin in the code above) is low so I don't think that's the issue.

>> No.1768375
File: 20 KB, 400x400, WTC-57521.jpg [View same] [iqdb] [saucenao] [google] [report]

I'm autistic and I want to dim my incandescent bulb.
Is there anything to look out for when buying a dimmer? I tried looking up datasheet for a few but they don't have any interesting details, like method of dimming or PWM frequency.
I really don't want my light to flicker.

>> No.1768379

Get trailing edge dimmer, they cost more, but they will also work with LED bulbs and cause less emi

Other than that literally any dimmer on the planet will work with incandescent bulbs

>> No.1768381

You're one of those halogen autists? I envy your CRI and color temp. 99% of mains dimmers will work at mains frequency due to how phase-fired control works, or technically twice mains frequency, so with LEDs you could end up seeing a 120Hz flicker. But you probably don't need to worry about it when it comes to incandescents, because the filament itself has some heat capacity and acts as a buffer from cycle to cycle. You don't notice the 100Hz flicker from them just from operating on AC, do you? But it might be more visible down at lower brightnesses.

To avoid flicker you need to PWM it at a faster frequency, but traditional TRIACs require zero-crossing to turn off again, and so can only work at mains frequency or some fraction of it. Hence you'd need to use an SSR, but most I've seen aren't meant for terribly high frequencies and they're all kinda expensive.
So instead I'd chuck the power through a bridge rectifier and use a high-voltage MOSFET like an IRF740 and run it at 500Hz or more, probably with a heat-sink. Practically speaking, I'm not sure if these exist in preexisting modules, but look for HV DC dimmers/speed controllers I guess.

Trailing edge dimmers work at the same frequency, how would you get less flicker? The only advantage is the lack of that sudden on-spike, which should easily be able to be mitigated with some filtration.

>> No.1768383

most SSR s use zero crossings to turn on or off
if you want pwm you need to use mosfets

>> No.1768400

Damn I actually don't know much about electronics at all, I just want it to look nice.
Working with LEDs would be nice, I might have to swap it for a LED bulb during summer. Any example of a trailing edge dimmer?

I have an arduino maybe I can use it as a PWM source or something.

>> No.1768407
File: 17 KB, 998x511, file.png [View same] [iqdb] [saucenao] [google] [report]

beware when buying led bulbs, they must specifically support dimming or the dimmers wont work

if you just want a cheap dimmer get the normal leading edge dimmers that is how all the cheap ones work BUT don't buy the dirt cheap ones that just use capacitors for timing because they won't work with led bulbs, you need some that uses a micro to detect the edge crossings in the sine and time the triac triggering correctly >>1766423 this pic is an example of a diy leading edge dimmer that uses attiny13
and this is the schematic for it, it is very cheap and easy to make
and if you can into arduino sketches you can add cool functionality like remote control or blinky effects and such

>> No.1768411

>most SSRs use zero crossings to turn on or off
Oh, I swear SSRs usually used optocoupled antiseries MOSFETs, but all the ones I'm looking up show diagrams with TRIACs or SCRs.

If it's a light-bulb that fits into an existing socket, then I'd say incandescent bulb is the best way to go if you don't care about power costs and want to dim it with no flicker. If you have the luxury to have a custom fitting of some sort, then you should consider LEDs. Because getting a variable constant-current LED supply is relatively straightforward. Even if they're not made for it, they'll likely use a resistor or two for setting the current, for which you could simply replace with a potentiometer. There are even drivers with remote controls, and while a lot of them are made for RGB LEDs, you could just as easily use them with warm/cool white LEDs for that nice variable temperature color.
Personally I'd lean towards LED strips only because they need no heat-sinking due to their form-factor and can be manipulated into a variety of shapes, and I never want particularly bright lighting. But COBs come in higher brightnesses and are more compact, and can fit standard reflector housings and lenses.

I haven't actually checked to see how much flicker you get from a dimmable LED bulb though, so maybe buy one and see.

>using a transformer instead of a custom symbol
>those ziggy zaggy traces
pleb schematic design

>> No.1768414

It's for a desk lamp. The bulb is pretty hot (the lamp comes with a 60W bulb for whatever reason) which is why I'm gonna swap it for LED comes summer until it stops being 40C outside.

>> No.1768454

a simple dimmer solution would be a simple on/off switch which sets the brightness to 0% or 100%

>> No.1768493

i'm trying to seriously minimize power consumption of a microcontroller algorithm that detects a sustained (say >30s) vibration event via an accelerometer. the vibration isn't very high above the noise floor, but it is large enough to be detectable based on my data collection. the accelerometer will sometimes give nonsense samples that need to be filtered. my idea right now:

>wake every 32 seconds
>read (n) samples
>calculate median
>if (%) samples exceed some threshold relative to the median, begin continuous sampling
>continuously sample, maintain a rolling median
>increment (or decrement) an accumulator for each sample that exceed (or does not exceed) the above threshold
>set a "vibrating" flag if the accumulator reaches an upper bound
>clear the flag if the accumulator reaches a lower bound, and stop continuous sampling

that seems awfully complicated though. is there some obvious way to do this that i'm not thinking of?

>> No.1768552

Search for foxhole am radio to understand the basics, then search for superheterodyne receivers.
This guy videos might help

>> No.1768648

I'd like to get started with formal verification in (system)verilog.
Are there any resources that you would personally recommend?

>> No.1768684
File: 54 KB, 1636x646, Capture.png [View same] [iqdb] [saucenao] [google] [report]


Please improve your schematic layout. I've done a rework of how I would personally lay out that design. Below are some recommendations on how to improve readability.

1. Absolutely no text over symbols, ever.
2. Ground should point downward unless you really really need the space
3. Traces should make the minimum number of bends to get to their destination logically. Rearrange components if necessary to make it easier to route traces logically. Traces should be inline with each other or symmetrically offset when logical.
4. Absolutely no 4-way junctions, ever. They become ambiguous if you meant to join all traces at that point. There is no ambiguity in a 3-way junction. If you must join 4 traces at one point, offset by 1 and create two 3-way junctions. This problem becomes apparent when photocopies are made and it becomes more difficult to see the junction dots.
5. use proper symbols. If they don't exist, modify existing symbols into a new symbol library or draw your own.
6. Use global labels or power symbols to avoid long traces across the schematic between unrelated parts. (you can do your own global labels for VACL/N, but kicad also has line/neut for you as well as power symbols
7. PWR_FLAG symbols can fuck off to a corner of the sheet with a note that they are for DRC use only as they are specific to kicad. Only place as many as you need when you run the DRC, depending on your schematic symbols they have power input/power output set on their pins and tell the DRC what the net is. If you get a pin not drive error in the DRC it probably means you had some passives or other components between a power net and the net in question, and the DRC doesn't know that it should be power.
8. Do not join symbols directly from their pins; there should always be a trace between the two. This is important for cases where you have symbols with multiple components on them, some transistors have drawn-in resistors with them. Separate by at least 1.

>> No.1768689


Personal autism from here on:
9. I prefer to have connectors separate with clearly defined net labels and power symbols connected to them, but there's no great reason for this other than making it easy to immediately read what pin position is what net later on, especially during assembly/testing.
10. Do not connect a trace to a pin at a right angle. This is hyper-autism for me.
11. For cases of >3 repeated components, you can get away with the 5X, 10X, etc. notation. But for quantities of 3 or less just repeat the symbol. It makes it easier for dealing with
12. Whenever possible, box logical parts of a schematic in their own contained area. This isn't relevant for what you posted since it's so small and specialized.

Mistakes on my schematic: R1 not sized, I forgot to add the fuse F1, labels for U2 should probably be on top to be consistent with the rest

>> No.1768725
File: 371 KB, 540x405, Please slow down.png [View same] [iqdb] [saucenao] [google] [report]

I want to make some things with microcontrollers similar to an arduino, but smaller and cheaper so I can make a bunch of them, but I have no idea where to start. I know arduinos use an ATmega chip for the "brains" of it and if i'm understanding correctly that's most of what I need. I don't need a lot of headers or reset buttons buttons or LEDs or whatnot that come on them. Is something like this even feasible or is it too much to ask? Even the cheapest Arduino I can find is somewhere around $10, which adds up.

>> No.1768745

You could use an analog circuit in order to integrate the values coming out of the accelerometer, and if they're over a threshold after 30 seconds it will feed a comparator that sends an interrupt to the MCU, if not the integrated value gets reset with a mosfet pulling the voltage back to ground. Assuming it's an analog sensor that is.

If it is digital, you could write a more formal DSP algorithm.
If you're looking for vibration in a system with resonances (it probably does) you could put a filter on the signal (either on the wire if analog or in the DSP algo if digital) and use that to filter out noise and possibly use a simpler method for detecting whether it's detected vibration or not.

How much power is it using anyhow? Could you implement energy harvesting of some sort? Like one of those self-winding watches, or a little amorphous solar cell, or a thermopile somewhere, or even an antenna stretching the length of whatever it is that's vibrating.


>> No.1768753
File: 90 KB, 746x419, Curiosidades_aerogenradores_746x419.jpg [View same] [iqdb] [saucenao] [google] [report]

I'm curious. How do wind turbines synchronize their outputs to feed the grid? I know small ones tend to work on DC and generate AC through an inverter, but how do large ones (In the range of MW) do it?
How would you design a turbine capable of generating some kilowatts of power?

>> No.1768756

if it's just cost you're concerned with then you can get chinkduino nanos for something like $1-2 each on aliexpress. they work perfectly well with two minor catches: you'll need to install the driver for the onboard chinese ch340g usb controller. and the voltage regulator is fake, which just means it won't take 16Vin or dissipate 1W like the real thing would.

if you want to use a standalone uc then I recommend PIC microcontrollers. the external components required are a cap, a 10k pullup on mclr (both "optional"), and a 5 pin programming header. the pickit programmer is something like $10 off ali. my chink copy is called a "kit 3.5" and it mostly works. the chips themselves are <$1. programming standalone microcontrollers is not that different from arduino except the toolchains are annoying and all of the hardware functionality is controlled by writing to special registers. Serial.write('n'); on arduino will be TX1REG='n' or likewise. there aren't any hand holding built in error corrections so you'll have to RTFM or drown in bugs.

>> No.1768758

I had the same question and looking at online resources it seems that also the bigger ones just rectify to DC and then invert back to AC.
I'm just a dumbass but idk how else you could handle the constantly varying frequency of the rotation.

>> No.1768762

Thanks for the tip. Yeah, the cost is what I'm concerned about; the standalone microcontroller sounds like it's a bit much for me to get into at this time.
>install the driver
On the computer I'm connecting it to, I assume?
>fake voltage regulator
>won't dissipate 1W
Only planning on connecting it via USB power, so hopefully that won't become an issue since I'm just going to be driving a bunch of LEDs (unless there's something else that I'm overlooking, which is possible).

>> No.1768763

Absolutely feasible.

You have two options here. You can select a chip that has an arduino configuration all done for it already, or you can write code native to the microcontroller. There are advantages and disadvantages to each. Here is the process:

Native route:
1. Select a microcontroller. I would start by going to digi-key and going through their parametric search for what you need in terms of GPIO and other peripherals. https://www.digikey.com/products/en/integrated-circuits-ics/embedded-microcontrollers/685
You may want to filter by particular manufacturer if you know you want to use an MSP430, or a PIC, or an ATMEGA for example. Otherwise, if you don't have a preference, that's ok too; you'll just have to learn how to program whatever one you select.
2. Read the datasheet. You can skip over sections that you won't be using, but you will need to be comfortable knowing how to set up connections to get a basic working micro--do you need an external oscillator, is there a reset pin that needs to be held high, etc.
3. If you want to see how programming one of these will go, you can get a devkit and try with that first if there's one available.
4. Look up appropriate IDE and compilers if shipped separate from the IDE--I know microchip has separate C compilers that you need to download in addition to MPLAB. More likely than not you'll be writing in C but there are different options out there.
5. Identify required configuration for your application--what registers must be set. E.G. pinout config for GPIO, ADC setup, timers, oscillator control.
6. Hardware design and layout. This is after step 5 as you need to know which pins go to what. Make sure to include a programming header!
7. Write code to get to blinky. This proves you have a working setup. Repeat for each module you use to verify functionality.
8. Write application code.
9. Debug and test.

Pros: Full control over uC peripherals, endless controllers to choose from
Cons: It's a lot of work

>> No.1768765

>On the computer I'm connecting it to, I assume?
>Only planning on connecting it via USB power
you're good then. the atmega328p only needs a 5v main supply and a 3.3v secondary supply that it generates internally. the fake voltage regulator is only if you're powering the board from a >5v supply.

>> No.1768768


Arduino route:
1. Find a list of microcontrollers that have arduino configs set up. Based on your needs, select one that works for your application needs. Alternative: write the arduino config yourself for the micro. That's also a lot of work.
2. Read the datasheet, again you need to know what is required for a minimal connection standpoint.
3. See if you can get a prototype board set up or do it on a breadboard to prove you can load the arduino bootloader onto the chip. Get to blinky.
4. Do full hardware layout
5. Write application code
6. Debug and test

Pros: Streamlines application code (can use C++), don't need to fiddle with individual registers to set things up. Getting to blinky is trivial. You get access to a plethora of libraries to easily integrate into your project.
Cons: No or limited control over low-level functions of the uC, extremely limited number of uCs you can work with out of the box, bootloader takes up space in in flash.

-Prototyping: Arduino or arduino-based solution. Fast and easy but limited.
-Production: Native uC code. Slow process but robust and high performance.

>> No.1768775

How did you get that carbon fiber looking finish?? That pattern looks great with the shine of the filament btw.

>> No.1768842

>Mistakes on my schematic: R1 not sized, I forgot to add the fuse F1, labels for U2 should probably be on top

that's very nice but there's plenty more mistakes
- using ground symbols everywhere would remove a lot of superfluous wires, with fewer cross-overs. (that's just cleanliness, which is next to ''cleanser'' in the dictionary, not ''godliness'' as many think.)
- reverse the optoisolator so emitter faces down. NPN emitters point down, PNP emitters point up. (that's the expectation, even in Australia.)
- the circuit exists to drive a load, yet no load is visible in the drawing: just two wires exiting into the cosmos. (thats plain stupid coz it leaves you wondering.)

>> No.1768855

the problem is that one bit of data is already on the bus before you start clocking. most people use chips with a different clocking paradigm, where the slave device places data on the bus after the clock rather than before the clock. the function has no idea what chip you put there
the upshot is that you just need to read that bit of data before you start clocking

a good 18W LED light engine with Ra>80 can be purchased for $2 or so. paste it onto a heat sink with some thermal plaster and use a boring constant-current driver to adjust LED current to your taste et viola, desk lamp

is a dead meme 7-bit barely-a-micro. leave it die

>cleanliness is next to godliness
means don't wash your junk

>> No.1768867
File: 33 KB, 1662x550, Capture.png [View same] [iqdb] [saucenao] [google] [report]

I'm torn on using grounds everywhere to remove wires. That one I think is personal preference. Generally if ground or power rails can be arranged to have positive voltage at the top and ground at the bottom, readability is the same as replacing all the grounds with individual ground symbols or better.

No; that's not right. I'm assuming that he's got internal pullup resistors on PB1. So when line is crossing zero, the optocoupler U2's phototransistor pins 3 and 4 are not conducting, setting PB1 high. That's how the microcontroller knows where zero is and doesn't get out of sync with the mains line. Otherwise, PB1 is pulled to ground.

Practice from our EEs at work is to only place symbols that are actually going to be components on the board; hence in this case it is a connector J2. If you think it warrants it, place a note by the connector describing the load that it's driving, including voltage and current draw or other important things. Some people may place the load symbol on the schematic itself even if it is connected via a cable and connector; this should definitely be noted next to whatever component that is, again to avoid ambiguity.

I don't usually work on mains stuff but I think that the varistor is usually close to the input connector and fuse.

>> No.1768868

Aren't there symbols for mains plugs?

>> No.1768877
File: 335 KB, 526x457, screwable-wire-thingies.png [View same] [iqdb] [saucenao] [google] [report]

What are these things called? I want to buy some but can't for the hell of it figure out what they're named. It's those things where you can clamp/screw a wire onto a board to fasten it without soldering the wire itself.

>> No.1768878

Screw terminals. There's a few varieties, but those are the most common I've seen outside of mains terminal blocks (which are just wire to wire, not wire to PCB).

>> No.1768880


>> No.1768935

This looks like it could lead to some interesting projects:
I wonder if such devices could be used by the Police alongside DNA and fingerprint evidence?

>> No.1768936

Oh this one has a section on people

>> No.1768955
File: 16 KB, 800x435, i can smell your cunt.jpg [View same] [iqdb] [saucenao] [google] [report]

you probably wouldn't use them on a schematic destined for board layout. instead, you'd just lay down pads for whatever the cord will attach to. if that happens to be plug prongs for a wall wart of some kind, kewl

hiring Pic related is a better idea

>> No.1769013

fren, X2 doesn't men two capacitors it is a x2 rated safety capacitor

>> No.1769015

Pfft didn't notice that, thanks for a laugh at someone else's expense

>> No.1769017

> the upshot is that you just need to read that bit of data before you start clocking

Oh man, yeah that works, cheers. I mean, that's annoying but it works!

void loop() {
int data1 = 0;
int data2 = 0;
int data3 = 0;
int data4 = 0;

int first = digitalRead(_CD4021B_dataPin);
digitalWrite(_CD4021B_latchPin, LOW);
data1 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
data2 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
data3 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
data4 = shiftIn(_CD4021B_dataPin, _CD4021B_clockPin, LSBFIRST);
digitalWrite(_CD4021B_latchPin, HIGH);

if (first || data1 || data2 || data3 || data4) {
Serial.println(first, HEX);
Serial.println(data1, HEX);
Serial.println(data2, HEX);
Serial.println(data3, HEX);
Serial.println(data4, HEX);
delay (300);

>> No.1769043

Build your mum a 20000 watt power supply

>> No.1769044

Oh yeah I forgot about that feature of PISOs.

>> No.1769063

>No; that's not right. I'm assuming that he's got internal pullup resistors on PB1.

i didnt mean that the wiring was wrong, only that the symbol was drawn upside down.
for the same reason we make the signal flow from left to right, current should flow top to bottom. (so an NPN emitter should point down)

>> No.1769092


>> No.1769094

And this is why I don't work on mains stuff. Ty

>> No.1769097
File: 33 KB, 1808x500, Capture.png [View same] [iqdb] [saucenao] [google] [report]


>> No.1769100
File: 36 KB, 567x559, 1581637116177.jpg [View same] [iqdb] [saucenao] [google] [report]

Where can I buy a kW power range/ MHz switching range laser, and just how many paychecks will it cost me?

>> No.1769119

Ceramic capacitors are microphonic, right?
Does that mean if I speak near an RF oscillator with a cap in the signal path, and it radiates off some wire, that anyone with a reciever can hear me?

>> No.1769148


no, but you can hit with your dick to send Morse.

>> No.1769157

Look for an optical lab supply store, though those guys don't mess around with their prices. If you're lucky then some yobbo on eBay will be selling a 2nd hand one, or maybe you'll find a cheaper one on Alibaba or wherever, but both are somewhat untrustworthy.
What do you need one for?

That's why you use two ceramic capacitors in parallel, but are physically positioned in opposite directions so the microphonic character cancels.

>> No.1769304
File: 31 KB, 540x508, 1558358312559.jpg [View same] [iqdb] [saucenao] [google] [report]

>current should flow top to bottom
which current? conventional or electron flow?

class I ceramic dielectrics are not, and their capacitance varies so much less with voltage and temperature, but those top out in the nanofarads or so. that's perfectly alright for a tuning reactance
class II dielectrics are both microphonic and temperature- and voltage-dependent

>tfw benis not sufficiently physically endowed to send a dah

>> No.1769369
File: 21 KB, 750x547, 1581983549278.png [View same] [iqdb] [saucenao] [google] [report]

How can I power and connect several peltier modules and regulate their temperature on the bottom floor of a chinchilla cage? I looked at a tutorial for making a mini portable fridge using Arduino but it was a bit too much for my brain. I know I need an Arduino Uno, a relay, a DHT11 or LM35 temp sensor, and something called a breadboard. I am completely new to this and don't understand most of the terms. For reference my mains power is 240v at 50Hz with British plug sockets

>Pic related is an explanation a kind anon drew for me

>> No.1769378

conventional obviously, electron flow is never even mentioned unless you're dealing with vacuum tubes or your circuit needs to pass through the ionosphere

nice pic. you don't necessarily need an arduino at all, but some sort of microcontroller to read the temperature and provide PID-style control on the data and handle the PWM for you would definitely make things easier. an arduino is a microcontroller board btw, and you can get away with a nano instead of a massive uno and they're functionally identical, or even a digispark but those can be a bit of a pain to program. as for a DHT11, it will measure the temperature of the air, not of the floor, and I'd be most concerned about freezing the chinchilla feet. the lm35 will work fine for measuring the floor temp, but there are digital temperature sensors that might be easier. as for a breadboard, fuck no, those are for prototyping not for permanent circuits, and especially not for the kind of power that a bunch of thermoelectric coolers will require; solder is the only way to go for much of this circuit. you'll also need a dc power supply that's rated at the voltage of a single peltier and rated with a current carrying capability at or above the maximum current draw of a peltier multiplied by the number of peltiers. then you'll need a transistor (MOSFET) to turn the peltiers on and off, and a massive heat sink on the bottom of the peltiers. i can draw a circuit diagram if you'd like.

>> No.1769382
File: 685 KB, 1080x1647, Screenshot_20200218-000532_Samsung Notes.jpg [View same] [iqdb] [saucenao] [google] [report]

So what I'm getting from this is that it's a pain in the arse to do, at least for a beginner. The idea was to make a cooling system with piped water passing over the peltier cells and travel through a bigger copper pipe that blows cooled air via small holes in the copper pipe throughout the cage. Kind of like the pic. Which microcontroller do you recommend? I don't even have a soldering kit Jesus Christ

>> No.1769385

Also the peltier cells and pipes would be underneath a solid wooden floor so to avoid freezing the chinchilla's feet. I want to keep the cage at 18°C constantly and found an automated fan with thermostat on Amazon that can supply the air for circulation

>> No.1769388

on the electronics part, since you just want to get the job done and don't want to learn the intricacies of control theory, just get one of these
that's a lot of mechanical pieces, fren. I can't really advise you on that part

>> No.1769394

Thanks for the link I didn't know those were so cheap. I'll read a bit more about how this stuff works. I'm useless at this :'(

>> No.1769400
File: 275 KB, 1727x845, all you need to cool.jpg [View same] [iqdb] [saucenao] [google] [report]


dont believe the hype. you dont need no stinkin' arduino/MOSFET/temp sensor or any of that satanic microchip garbage. you just need 3 things and 5 minutes to wire 'em up.
- cooling module from Amazon, TRHS in pic. could be a single cooler, dual, or whatever you need. the one in the pic needs 12V at around 10A
- power supply capable of supplying 12V at 10A. in the pic shown, it comes as a kit, so you dont need to look for one.
- a thermostat is optional, coz it's likely you'll need the thing to be going full throttle all the time to get any kinda cooling.
- if you must have one, you can choose an electronic one like the one on the right, or a mechanical one like the one on the bottom. very easy to wire up: just cut all the red wires for the fans and coolers and feed them thru the thermostat.

>> No.1769407

> at least for a beginner
That's a pain in the aris to do even for someone who isn't a beginner, because it requires work from all sorts of disciplines. If you're going to flow liquid coolant about the place anyway, I'd just use a conventional refrigeration compressor, they're far more efficient than peltiers (those peltier mini-fridges use about as much power as a full-size refrigerator on average). Then you'd just run the copper coolant line around the enclosure, probably no need for running an air line over the top of it so long as you have air circulation inside the enclosure itself.
So basically, go to the junk yard, buy a trashed refrigerator, use its compressor and coolant loop and heat sink with little modification, and chuck a fan and temp-control module and you're good to go.

>incubator temperature controller
Holy shit, I didn't know these existed, and all this time I'd been looking for a cheap PID. Does it support cooling as well as heating? What temp range? Does it do proper PID?

>> No.1769474

>cooling as well as heating
maybe. some just have a Form C relay with normally-closed contacts, some may invert the sense of the relay coil programmatically
>proper PID
very probably not, more likely just bang-bang, but could still be rigged to work well enough for OP's pocket fren. PID isn't very useful without PWM control anyway, not very compatible with mechanical relays

>> No.1769476

if you want cheap PIDs, the panel-mount devices seem to be where it's at

>> No.1769488
File: 769 KB, 2048x1536, IMG_20200218_113822.jpg [View same] [iqdb] [saucenao] [google] [report]

Dimmer autist here, ended up getting a cheap $2 dimmer.
Forgot to take internal pics, there's a triac inside, and some passives alongside a huge pot. Can't take the whole thing off.
Is that enough to tell which type it is?

>> No.1769500

>PID isn't very useful without PWM control anyway, not very compatible with mechanical relays
I was about to say, if it's relay driven it would be hard to have it PID.

To be honest I'd rather have a dedicated low-latency PID IC to just slap on whatever I please, with a few programming resistors (and maybe capacitors too) to set the P, I, D coefficients and the PWM frequency. Could maybe fit on an 8-pin IC, but would likely need at leat 10, depending on what sort of temperature measurement it does.

Ordinary bog-standard phase-fired control using a DIAC for triggering, almost certainly. In other words, potential 60/120Hz flicker, poor electrical noise, potential humming. But probably nothing you need to worry about. It's rated for the power/current that your lights will draw, right?

>> No.1769502

It's rated for 700W
My desk lamp is 60W
There's some kinda noise at lower brightness but it comes from the lightbulb.

>> No.1769504

hi mechanical retard here.. are inverters just a series of mosfets that make DC behave like an AC signal?

>> No.1769505
File: 17 KB, 774x382, pwm inverter.png [View same] [iqdb] [saucenao] [google] [report]

Sounds pretty alright, you could put a small X2 cap across the bulb and that might stop the noise. Or should it be a small choke in series with it?

Yeah basically, but because AC is a sine wave, you often want to imitate that somehow. So they do this either with a variable duty cycle (picrel) or maybe in expensive cases throwing a big low-pass filter on it also and using feedback like some sort of class-D amp.

>> No.1769508

I got a 32V max input buck regulator along with a 36V power supply, If I increase the buck regulator's capacitor voltage, will I be able to use it with 36V?

>> No.1769537

Good news to me!
The chink module uses a rebranded LTC1626, which supports 36V so no problem at all, specially if the shitty 36V power supply sucks and gives out something like 35V. Still changing the caps tho.

>> No.1769540

Check all the components, not just the capacitor. The diodes should be able to handle it, but the switching controller (and mosfet if it's an external one) and ceramic caps are what I'd be worried about, after the elecrolytic caps after all. Non-destructively measuring the ceramic cap voltage might not be easy though.

>> No.1769541

If i have a some sine, like a sound wave for example and I have some carrier wave like a 10MHz sine
What is the easiest hardware way to combine them together (for transmission) and following that, what is the easiest way to split them apart?

>> No.1769546
File: 11 KB, 400x400, yizzle.png [View same] [iqdb] [saucenao] [google] [report]

What you want is called a mixer; it multiplies the two signals together. There's a few different topologies, notably the diode ring mixer: a circuit that uses 4 diodes and a couple of transformers and usually has to be made by hand, and the Gilbert cell: a transistor-based circuit that you can buy in relatively cheap packages like the SA602. The SA602 and similar mixer ICs are small and have an optional built-in oscillator for connecting to a crystal of choice, and I assume they provide a relatively sinusoidal carrier wave, but they only really work at a particular input voltage range, so they can be fiddly to get going. Another option is a transconductance amplifier, which IIRC can be used as a mixer, but they're likely not suited to 10MHz operation.

Furthermore, if you want to transmit an audio AM signal, you'll need to include a DC component to the audio wave, otherwise you'll end up with a signal envelope of |f(t)|, pic related. This results in a carrier wave in between each sideband in the frequency domain.

Also it's easier to modulate and demodulate an FM signal if you just use a PLL, plus it's better for noise reasons.

>> No.1769548
File: 143 KB, 2136x318, cascode.png [View same] [iqdb] [saucenao] [google] [report]


>> No.1769554

Hmm make me think if it wouldn't be much simple to simply use a beefy micro that does all the math and combines the sines in software, and then outputs then i just hook up some simple DAC to turn the software sine into actual electric sine

>> No.1769555

Not at 10MHz it isn't. You'd be lucky if you could get that sort of thing going at a few hundred kHz with a bargain bin avr. Gilbert cell mixers aren't that awful to work with, I bought a dozen myself since they're still better than the alternatives. All you need to do is ensure that your audio signal is attenuated and biased properly before entering the mixer, chuck a 10MHz crystal on the two oscillator pins, and throw an RF amplifier of some sort on the output. The inputs are balanced, so you could just have an audio signal on one input and the bias on the other, probably.

>> No.1769558

>Gilbert cell mixers
okay, will look into that, some ready made ic like you mentioned will probably make it relatively easy

>> No.1769560

The main ones are the SA602/SA612 and NE602/NE612 I think, there's some slight differences between the 02 and 12, but I think NE and SA are just manufacturer markings.

>> No.1769562
File: 1.64 MB, 480x270, Neopixelx8.gif [View same] [iqdb] [saucenao] [google] [report]

I want to power an Arduino Nano and some amount of Neopixels (around a meter or two) off of the same power supply (wall-pluggable USB, so 5V and somewhere around 2A). I can't think of any reason not to just fork the end and have 5V/GND running to the Neopixel strip as well as the Nano. I feel like I'm missing something - this seems too easy. Any thoughts? I know I'll have to be careful about the amount of neopixels to keep the power draw under what the USB socket/wall wart can provide.

>> No.1769567

a lots of pixel strips are 12V so check the voltage
other than that there is no problem with powering then both from the same power source
just make the micro can handle the amount of leds you want to program. the neopixel library buffers all leds into an array so it can send the colors for all of them all at once, so with something like attiny i think you can do at most like 80 pixels or some shit like that, it is dependent on the ram size of the micro, nano has i think 32kb? so you should be fine tho

>> No.1769571

A metre of neopixels (especially if they're as densely packed as your image) will take more than 2A. My RGB LED strip runs on USB and is about 1m long with a ~30mm spacing, and it uses 90% of the maximum 2A. Look at the neopixel datasheet and see how many watts they pull maximum, and hence determine how many you can fit in the 10W spacing on a USB supply, assuming they're 5V that is.

>> No.1769619

>kW power range/ MHz switching range laser
crashed alium spaceship

>> No.1769634

>throwing a big low-pass filter on it
How feasible is it to build a pi-network LPF using the transformer as the inductor element?

>> No.1769648


the transformer will complain to the union, "I wasn't built for this, my job is to transform, to simply use me as an inductor is insulting, abusive, and creates a hostile work environment."

>> No.1769883

That was a fun read, looks like they're all the same save temperature tolerance. I guess by buying the SA602 I ended up opting for the most expensive possible model for no reason. Not to mention they were only available in SOIC8 when I got them.

Might work, try simming it.

>> No.1769911

Are there any good kits or sets that would come with the parts necessary to build simple circuits for powering a few small LEDs? If anyone could point me in the right direction I would appreciate it. I don't have any electronics experience except for a small introduction class.

I've never built something without a breadboard and instructions. Eventually I want to read all of those books in the OP, but I don't have the time right now.

>> No.1769923

A lot of the kits these days are arduino-based kits, basically little USB-programmable boards that you tell to read sensors or buttons and turn on LEDs and such. Personally, I'd avoid these as they don't really teach you about the basics of electronics. Gotta know analog before you go digital. But the theory and practice for electronics are really interlinked; you really need to get a strong comprehension of each segment before you move to something more complicated. This requires learning the theory so you don't blow up your parts and know what to expect, and testing and measuring such that you can learn how real parts differ from ideal parts and what circuits are practical.

>> No.1769949

Dimmer autist again, it's weird but I just checked with my camera and the light only flickers at higher brightness, if I set it to 60% it stops flickering.
I thought there's supposed to be less flicker at higher brightness?

>> No.1769967

$12-$16 elegoo kit, then look up "astable multivibrator"

>> No.1769974

that's just your camera slowing down its shutter to get the correct average scene brightness

what, just light them steadily? ledcalculator.net

>> No.1769990

can you set the framerate of your camera to be different? At higher brightnesses the shutter speed will be faster so you'll get more distinctive time-domain aliasing.

>> No.1769999

I have a string of 6 LED lights which flicker horribly.
I had assumed that they'd just be across the mains with a capacitive ballast, possibly a bridge rectifier, but when I measured the power factor, it blew 0.95.
This implies that every lamp has a tiny switching converter in it. Tiny.
I'm not asking for anything, I'm just marveling at how tiny and cheap power electronics have gotten.

>> No.1770003

oh fuck i accidentally reprogrammed SWD/JLINK pins. does anyone remember how to connect under reset or some trick like that? i hope if if i could reset it and hold it reset the code won't run and it should bring the pins to the default state. i think.

>> No.1770008

never mind, segger is smart, i pulled nrst low so basically connected under reset, and jlink interpreted that as a "locked chip" for some reason, and warned me it is going to erase the chip to unlock it. just what i needed. sweet. but i am surprised jlink has no explicit "connect under reset" option like in st-link.

>> No.1770011

but this is funny. i made a small correction in my code, i noticed the line where the portA is set to 0xffff and then reset to 0000 carefully via a mask. and i was like hmmm why only 8 bits, i need to correct that and assigned 0xffffffff to the entire port without any masks. but totally forgot that i really only needed to touch the first 8 pins which is my parallel interface to LCD. but instead i set all 16 pins on portA which of course include SWCLK/SWDIO. and everything worked fine except the debugger disconnected immediately. then it dawned on me... wow that was dumb. a near brick condition. (if you can brick an STM chip at all)

>> No.1770015

>switching converter
Or a linear dropper with a high-voltage LED array, there are multi-chip LEDs these days after all. The linear ones usually have no filtration after the current regulator IC and hence flicker somewhat, but not as much as a capacitive dropper. There's no real reason for a switching converter's output to be flickering, especially not at a visible frequency.

>> No.1770023

shutter speed ≠ frame rate
source: can operate a manual SLR camera

don't turn it on, take it apart. I think you'll be disappointed

I'm a big fan of dev boards tying BOOT0 to ground through something other than direct copper, so that you can remove a solder jumper or override a resistor to get BOOT0=1

>> No.1770027

I'm making a schematic for a numpad in KiCAD, the last time I so much as soldered anything was in Boy Scouts like eight years ago. I'll post it when I'm done, any more advice for literal retards? I'd like to follow decent practices while I'm going to take the time to do this anway.

>> No.1770031

>shutter speed ≠ frame rate
Yes I know that, I own a DSLR and always operate it in manual mode and have even messed about with night-time photography. Considered making a light-sensor+timer MCU circuit in order to get a perfect exposure each time, but I think it would also need a display and a rotary encoder/buttons since it would be dependant on f-stop and ISO. Not that I'd ever increase my ISO.
Setting his framerate differently (or syncing it to the mains frequency) could get rid of the aliasing, which is why I asked about it first. Because both cameras and mains electricity quite like to run at 60Hz, and since mains frequencies get varied in order to compensate for grid loads and clock timing, you'd always be really close but not quite at 60Hz, causing really bad aliasing.
As for the shutter speed, I think my explanation stands.

>a solder jumper
>a resistor
>not using a reed switch or hall-sensor so you can only program it with a magnet

>> No.1770032

As an addendum, should I really really take the time to put in a block diagram for how the IC I'm using functions? I've copied the pinout down as a symbol from the data sheet, is it strictly necessary to also try and fit in a block diagram for all the actual functions of the chip?

>> No.1770040

No skip the block diagram; it's not needed. Group pins by logical associations. Take a look at some of kicad's microcontrollers for example, it would be silly to do a block diagram for those.

>> No.1770050

I don't want to take it apart because it would be a destructive process and I can't get more of these nonstandard lamps. I can just look down through the glass and see a bit of a PCB.
I didn't think they were switching converters at first, but I'd be surprised. I've taken apart filament-style LED bulbs and found a little switching supply in the base. that one also suffered from flicker, I think just from having not enough filter capacitance on the HV side.

>> No.1770062

>not enough filter capacitance on the HV side
Oh yeah, that could do it. Can't imagine the switching MOSFET would like that sort of treatment though, probably getting some intermediate voltages from the controller as it browns-out each twice cycle.

>> No.1770096 [DELETED] 

I figure it's time for me to get my own multimeter instead of bumming the one at work or borrowing a friend's. Looking to get something from Amazon so I don't have to wait a couple of months to get it delivered. Any specific recommendations or ones to stay away from?

>> No.1770097
File: 12 KB, 282x178, Multimeters.jpg [View same] [iqdb] [saucenao] [google] [report]

I figure it's time for me to get my own multimeter instead of bumming the one at work or borrowing a friend's. Looking to get something from Amazon so I don't have to wait a couple of months to get it delivered. Any specific recommendations or ones to stay away from?

>> No.1770101

Get an ANENG AN8009

>> No.1770104

Any recommendations for aftermarket multimeter test leads? Fucked around with a disgruntled start capacitor and vaporized the tips on my only pair.

>> No.1770106

get the ones with silicone leads and rubberised handles (like the ones on the ANENG store) and chuck an inline fuse on it. DMM fuses are expensive, automotive/5mm fuses aren't.

>> No.1770110


>> No.1770116

>Considered making a light-sensor+timer MCU circuit in order to get a perfect exposure each time
there are times when it is correct and useful to operate the camera in Tv or Av mode
>setting his framerate
generally not a user-exposed feature on a potatophone. maybe there's a flicker killer feature which doesn't seem appropriate to this measurement
but a good strong light, enough to keep the shutter speed up, will help. the setup I'm thinking is to cover about 5% of the frame with a strong white light, like a tac flashlight against the wall, and let the rest of the frame show the pulsations of the shit lamp
>he doesn't know about 59.94Hz
your think pictures make my eyes rain
>using a mag sensor on a bodge
this is a galaxy brain meme waiting to happen. go forth, young faggot

keep them, and fabricate pogo pin adapters for the tips

>> No.1770148

>this is a galaxy brain meme waiting to happen. go forth, young faggot
Actually it stemmed from the idea of shoving a USB flash drive inside a computer mouse or keyboard but having it disabled unless you hold a magnet up to the right part of it. You know, for storing banned books or whatever. But toggling a USB line with a single switch happens to be difficult. Could probably do it with diodes and resistors, but it sounds like a pain.

>> No.1770164

optocoupled circuit to trigger the relay.

>> No.1770178

Yeah I totally forgot about that pin. This is my DIY board so I hard wired it to ground since I didn't have any plans to boot from anything other than flash. But this could be a good alternative to connecting under reset.

>> No.1770194

>2.You need to use For Raspberry pi system, and if the client doesn't configure it, you need to use a mirror.

>> No.1770213

You just need to tie BOOT0 to ground for a moment right? Could you have two pads to short, or a button? What is the motivation to solder it or wire it directly to ground with a via or something?

>> No.1770223
File: 46 KB, 566x161, renditionDownload[1].jpg [View same] [iqdb] [saucenao] [google] [report]

You tie it to ground to boot from flash. If you don't ever need to boot from a boot loader or SDRAM, you can permanently solder it to ground. But in my case when I fucked up and needed to prevent my flash code from running, I could either hold the reset low or pull BOOT0 high to avoid executing bad code from flash, before I could erase/reprogram the chip. Thats how you unbrick it.

>> No.1770247

What is the situation with mini LCD TFT boards for micros? Looking at the adafruit webstie, they are still stuck in 2015 offering cheap chink LCDs with a huge markup. I can buy those on ali for like $5-$10. And the next step up is something like 7" for raspberry that sell for $70-$100, yikes. Is there anything in between? I'd like to experiment with cheap chink stuff but make it work faster than a typical parallel 8-bit interface, to try to achieve flicker free refresh rates of at least 60fps. Just for the heck of it. Possible? I can buy just a controller and wire it myself but I am not sure where to start. Looking at their datasheets, there are a bunch of different LCD controllers available on ali that support RGB / HSYNC/VHSYNC. Do I really need to worry about SYNC signals in 2020? Is there any specfic type of an LCD glass that I could buy to take advantage of any STM32 built-in interfaces like DVI? or maybe possible convert it to HMDI.

>> No.1770370
File: 46 KB, 485x484, 78121bfc71-cd4043.png [View same] [iqdb] [saucenao] [google] [report]

If I have an SR latch like the CD4043, is there a way to get it to toggle based on the output that's on.
Like if I tied S1 and R2 together and S2 and R1 together and then had Q1 and Q2 going into two separate AND gates and a single push button connected to the other input of both and gates and the output of the AND gates fed into S1/R2 and S2/R1, I assume pressing the button would change the state but as long as the button was pressed it would toggle back and forth or what?

>> No.1770392

Most modern SSRs already have optical triggering.

Use JK flip-flops instead? Or do you specifically need (gated) latches for whatever reason?

>> No.1770406

if you weren't interested in hacking the MCU on the pen drive you could use said Hall sensor to control one of the address lines to the flash chip

may all anons laugh at my pain and learn from it

there are internal pulldowns on that pin but I don't trust them in a noisy environment

HDMI is a pretty big barrier and an 8-bit parallel interface is plenty fast as long as it's properly memory-mapped, which apparently is not the case on the Pi series, but is possible with some larger members of the STM32 series
>ordering from adafruit
come on now
>sync signals
yes you still need to worry about these in current year, at some level or other. nowadays they're just muxed onto a high-speed serial stream
start by looking up the DS90C363 LVDS transmitter, for a tiny bit of the flavor of modern display links. you are unlikely to find very much HDMI information or components in the white public domain but you might find some work products scattered around on github and some specs if your search fu is strong. generally, anything significant to do with video implies a specialized chip or an FPGA, and 100MHz signals are an adventure all their own
>convert it to HDMI
HDMI is a superset of DVI which adds audio, HDCP, and some other crap mostly useless for experimenters

could, but watch out for glitches, propagation delays, and stuff like that

>> No.1770441

I'm getting an assortment with a few power resistors for general use, should I get 1W or 2W ones? I'm only buying higher power ones for specific uses.

>> No.1770455

>one of the address lines to the flash chip
Or even better, if there happens to be an enable or chip-select line I could use that. But either way I'd be messing with some really fine-pitch soldering, as opposed to just slapping something on the Vcc, D+, and D- lines out of the USB connector.

>> No.1770480
File: 331 KB, 444x446, Screenshot_12.png [View same] [iqdb] [saucenao] [google] [report]

>8-bit parallel interface is plenty fast
So i ran some tests using a cheap 320x240 64k color LCD. I re-wrote some critical bitbanging parts in asm and ran on STMF4 at 168Mhz. The entire GRAM is now updated in about 10ms which would give me about 90-100fps which is faster than I'd ever need. Problem is the flicker since data are written directly to GRAM and I feel like there is no way to avoid it.
> properly memory-mapped
what do you mean by memory mapping? I haven't used FSMC. I ported/hacked some LCD drivers I found on github that bit bang all the strobe signals to emulate the 8080 mode (8-bits plus control signals). I don't know if FSMC would give me any improvement in speed and/or flexibility.

However as a next step I'd like to try something bigger than 320x240 and thats where the 8-bit parallel mode might not be fast enough. I am not looking to play videos, at least for now, just wanted to get a refresh rate close to 60 fps. Which is why I was wondering if it would make sense to buy some of those cheap LCD controllers (like in the pic) but wire it for RGB instead of 8-bit. I don't know if thats a popular approach and how to best drive it. I don't want to try to bitbang RGB+sync signals.

thanks this is interesting, need to read more on that. may graduate to LVDS panels as the next step.

>there are internal pulldowns on that pin but I don't trust them in a noisy environment
not just noise, may lead to glitches due to high/low level thresholds affected by a slow VDD ramp-up.

>> No.1770484
File: 1.17 MB, 827x1168, 91f8a71c8a7e502123222115ddc35a9d.png [View same] [iqdb] [saucenao] [google] [report]

Also, are metal film resistor worth it over carbon film when dealing with 1/2W?

>> No.1770492
File: 123 KB, 1110x825, Screenshot_13.png [View same] [iqdb] [saucenao] [google] [report]

Even if I don't buy anything from them, I enjoy reading their manuals.
Don't hit and rain when transportation!

>> No.1770493

Look at this: https://www.buydisplay.com/7-tft-lcd-touch-screen-display-module-800x480-for-mp4-gps-tablet-pc

7" only $16.
and unlike "random driver no docs" on Ali they also provide a bunch of datasheets and code examples. wow.

>> No.1770496

Just measured my soldering iron temp with a thermocouple, when I run it around 330C apparently I'm only running the tip at 300C. When I set it to 410C boost the tip shows 370C. When I set it to 150C standby the tip shows 130C. So I think it's just thermal resistance between the thermocouple and the tip. Using a knockoff hakko T12 tip, btw.

>> No.1770502

you could try waiting for the VSYNC pulse if your controller allows that sort of thing
>what do you mean by memory mapping
ideally graphics RAM would be exposed with all its address lines so that you're not bouncing back and forth between command and data modes, so that you could literally just read/write a chunk of RAM. I have read about memory controllers that do something like that for you through the command/data interface to some types of memory, but I don't remember where I saw it. might've just been some serial flash shit from Freescale
>speed or flexibility
use a DMA channel, ofc
>larger panels
if you can, convert it to an RGB mode and bypass the internal GRAM so that you use the internal TFT controller with onboard RAM. that would be lit
good point about power sequencing, and thanks for the reminder that there are some STM32 MCUs affected by such issues (on the ~RST pin, not the BOOTx pins, but still)


not bad, especially if the viewing angle specs are legit

>> No.1770562

>soldering over my fabric mousepad
ok maybe this wasn't the best idea

>> No.1770567

Yeah I think I need a flip flop rather than a latch.
I basically just want the outputs to toggle once when I press a button but with tbe Sr latch they will flip back and forth as long as the button is help where as the flip flop with only flip on the rising edge

>> No.1770571

Yes. An SR latch has the property that when both Set and Reset inputs are held high it's put into an invalid state, so to put this extra possible input to some sort of use, people invented the JK latch. The JK latch has the feature that when both Set and Reset (J and K) inputs are held high, the output toggles, but this only really makes sense for a gated latch or flip-flop. A T-flip-flop is an even simpler device, it simply has a single clock input and a single output (maybe an inverted output as well), and whenever clocked it toggles the output from high to low or vice versa.

I implore you to look into the topology of SR latches, SR gated latches, and SR flip-flops, and the JK, T, and D type gated latches and flip-flops that came from them. There's some particulars about master-slave vs (a?)synchronous latches, but I never remembered that so it probably isn't important.

Since it sounds similar to what you're doing, I once got one of those IR remote controls and an IR receiver that spits out demodulated data, and I shoved that data through a low-pass filter and a schmitt trigger to just get a single pulse out of it each time I pressed any button on the remote. Then I wired it up to a 4000 series JK flip-flop IC and with both J and K held high, and wired the output to the PWM comparator of

>> No.1770572

to the PWM comparator of my LEDs, so I could turn my lamp on and off remotely. It worked wonderfully, except the horrible LED flicker and the fact that it tripped a breaker and wouldn't turn on again. An excellent learning experience as why COB LEDs are bad and a future part donator (I'm not letting three metric potentiometers and a heat-sink hit the scrapheap).

>> No.1770579

I need some sensor that can detect a human in the room
But the problem is that all sensors i found are motion sensors where if you stop moving they stop detecting you
Is there any sensor that can detect simply a human presence in the room without the human having to move?
(and something better than like a laser diode and a laser reciever, working on a beam interruption principle, since i want it to cover the whole room and not just one spot)

>> No.1770581

PIR sensors aren't intrinsically motion sensors, they just use a high-pass filter on their elements and differential measurements in order to filter out all the DC/low-freq noise. While it might not be possible to remove the filtering and differential measurement, even if you did you'd end up with a signal that's impossible to amplify unless you calibrated your room's temperature. Which technically is possible with a few temperature sensors about the place, but it sounds like a rather difficult plan. You could try doing away with the high-pass filter and keeping the differential measurement between the ~3 PIR pixels, but the plastic lens is designed for motion detection and diffracts the IR strangely, so you'd need to ditch that. I'm not even sure if there is a high-pass filter, so you could just try removing the lens and testing it. I'd put a piece of plastic bag over the bare sensor though, to stop direct light from shining at the PIR elements.

Methods like calibrated 1µm IR reflectance or calibrated ultrasonic echo testing could pick up minute changes in brightness or echo time/content respectively, but a person standing against the wall would be somewhat difficult to detect.
Another less direct method that could work well in conjunction with the PIR sensor would be a CO2/VOC sensor. By picking up the VOC content in the air, you'd know with significant latency whether a person was in the room or not. Combine this with the PIR motion sensor's data of when people moved (into or out of the room) you could get a good retroactive picture of when people were in the room and when they weren't. Provided the window wasn't open, and a host of other temperature and multi-person calibration issues.

>> No.1770588

easy, use raspberry pi with a camera and install tensor flow on it, you can then use neural network to recognize objects in the room, even specific people
it sounds complicated, but it is very easy to do with premade packages

>> No.1770593
File: 88 KB, 816x554, LM2576 typical application.png [View same] [iqdb] [saucenao] [google] [report]

Got a question regarding ripple current after a DC-DC step down converter I want to build:

Using a LM2576 in its typical application (pic related) I'd get a regulated +5V output.

Since I want to use it for measurements, changes in this voltage should be as small as possible and with the LM2576 at max output (3A) I'd get about 50mV output voltage ripple (more when changing the load during runtime, but that's not important for the application), so ~1% variation. I want to reduce this if possible.

The datasheet suggests a small LC filter at the output (20μH and 100μF) to reduce this further by factor 10 (so I'd get 0.1% variation). This might be enough for the application, but I looked online and found this suggestion by EEVblog (ep. 1116) [https://www.youtube.com/watch?v=wopmEyZKnYo] which suggested a capacitance multiplier (also pic related). Alternatively this is mentioned in chapter 8.15.1 of the Art of electronics. Any idea about effectiveness of this compared to the LC filter?

>> No.1770596

Basically, get one to output ~7V and use a linear regulator. Those transistor circuits will drop a few hundred mV, which may or may not be acceptable for a low-noise power supply.

>> No.1770601

If I understood the video on EEVBlog correctly, (3:20 minute mark) he precicely states that using a linear regulator would be a bad idea. At 5:20 he shows it in practice with a 500mV input ripple. Obviously that's a factor 10 increase from my case, but the problem should be the same or am I misunderstanding somehow?

>> No.1770607

I don't have the time at the moment to watch the video, but linear regulators offer excellent ripple rejection for the price of dissipating that energy, provided they're rated for the full range of current that you plan on using (upper and lower). They're basically a voltage follower strapped to a voltage reference, which contrasts to your voltage follower strapped to a capacitor. At some point, your ripple will be high enough that you get better ripple rejection with a linear regulator, but as the regulator's own noise starts to get more significant as the input ripple is lowered, you'd likely see more benefit from a capacitance multiplier. Just a haphazard guess however.
How much power are you drawing anyhow? It was enough to decide on using a buck converter in the first place.

>> No.1770609

>I don't have the time at the moment to watch the video
Understandable, thanks for answering anyways

>How much power are you drawing anyhow?
I'm actually not 100% sure as of yet, as I don't ahve all the information about how much load the whole thing has to handle. Probably makes sense though to get at least a rough estimate before proceeding.

Either way, thanks already. This already helped me by putting the idea of a linear regulator back on the table and I'll just look into it a bit more and compare them.

>> No.1770612

Yeah I'd encourage you to bust out the oscilloscope and do some tests for yourself. One of the circumstances where 10x mode on the probe really helps.

>> No.1770613

That's what I figured. Nothing beats just measuring after all.

>> No.1770624

I've a control circuit that uses a cmos 4000 series chip to base resistor of an npn that's driving a solenoid.
If I added a manual override switch that connects the base resistor directly to high (12v), could this damage the cmos chip of the output is low?
Aka if the cmos chip is trying to get the output low but I've it held high?
Would it be necassary to use an optocoupler to isolate them?

>> No.1770629

If i have a transformer that transforms 230V to 12V, does the secondary winding have to be shorten if no load is connected?
I read somewhere that if the primary is energized but the secondary is not connected to anything the secondary will generate crazy high voltages

>> No.1770636

Now that I think of it, I could just use a diode from the output of the cmos to the base transistor and have the manual pull high on that side.
The voltage drop across the diode shouldn't matter as I'm driving the transistor well into saturation amyway

>> No.1770722

>wanting to use an LM2576
it's current year. you should use something with a higher frequency and therefore less ripple
and you shouldn't be drawing much current from a reference voltage supply anyway. if you just need a voltage reference, use something like a TL431

yes, and maybe even if it's high, depends on your logic Vcc
no, do this >>1770636

>> No.1770745

I'm >>1770593. I'm absolutely open to using something with higher frequency. Any suggestions for a 12V input, 5V output, Best if I don't need too many external parts.

I need the 5V for some sensors and their ADCs, thus I wanted the Voltage to be constant if possible. I'll look into the TL431 and reference voltage supplies in general, thanks for that.

>> No.1770751

>ideally graphics RAM would be exposed with all its address lines so that you're not bouncing back and forth between command and data modes, so that you could literally just read/write a chunk of RAM. I have read about memory controllers that do something like that for you through the command/data interface to some types of memory
Not that I know of. There are different types of controllers so who knows. But the ones I am familiar with (ILI*) are basically a SoC with a controller/MCU with GRAM, gate/source drivers and DC-DC circuits. So you either write to GRAM via registers or bypass the controller/GRAM and talk directly to the panel's timing controller/gate driver via RGB/Sync interface. Or some of them are just glass + gate drivers and have no GRAM at all, and sometimes no power circuits either.

>> No.1770753

Why would you ever use carbon when MF is an option?

>> No.1770793

>12V input, 5V output
there are thousands of buck converters in SOT23-5 or SOT23-6 or SOIC-8 that will serve nicely. most of them use the same external part ensemble connected in the same way: feedback voltage divider, inductor, freewheel diode, input cap, output cap. some omit the external freewheel diode and use an internal MOSFET instead for improved efficiency. some require an additional bootstrap cap for improved efficiency thru higher gate drive voltage
>for some sensors
how much current out? for up to 12V in 5V out regular-tier chips, MP2307 (MPS) or MP1584 in SOIC-8 are good for up to 3A, MP2357 in SOT23-6 is good for 500mA but has higher fSW so smaller inductors can be used, AP3211 (Diodes Inc.) likewise but up to 1.5A out, HT7463 (Holtek) is similar to the MP2357 but accepts up to 52V in (you really want a high max Vin if your 12V source is a car). TI has a bunch of mids that would be suitable, and some really good application information in their datasheets. and of course LT/Analog Devices have a full menu of gourmet switchers for special applications. see manufacturers' or distributors' interactive parametric selection guides to optimize your selection such as

>> No.1770805

Is it bad to have a current sensing track running under a Mosfet that switches at around 200khz?

>> No.1770806

first I hear of it. What's the physics behind it?

>> No.1770825

Thanks for the detailed answer. I'm not sure about specific current out requirements yet. Once I know I'll reevaluate what I'm doing here.

>> No.1770826
File: 259 KB, 828x720, 962CC055-A41A-4A0E-9517-51C410DD673B.jpg [View same] [iqdb] [saucenao] [google] [report]

What do they mean by this? What is the resistor doing?

Also that cable with the resistor claims like 10x faster data speed, is it related?

>> No.1770835

No you do not. These crazy high voltages come from shutting off the current within an inductor or transformer, which is called an inductive spike. A transformer being fed a sine wave or even a square wave will not face these issues so long as it can feed current back into the source.

10 times faster data speed is likely just USB 3 vs USB 2. As for the resistor, different resistors on the data pins of power-banks or power supplies are used to tell the device being plugged in how much current it's able to draw without the need for anything smart on the other end. However I'm not sure if or how this applies to USB 3, especially USB C.
Looks like it's used for telling it how much current it can draw, same as USB 2, but inside the cable itself I think.

>> No.1770840

Not sure if you are still here, but when I was doing a prototyping thing for a big tech company in Japan, I stumbled upon a website of people DIYing their own printheads. Pretty interesting stuff. Replace X with dot obviously.
reprap X org/wiki/Scratchbuilt_Piezo_Printhead
reprap X org/wiki/Reprappable-inkjet

>> No.1770849

That’s kinda what I figured when it said “for protection” so you don’t kill your phone battery in 3mos by accidentally charging it at 5A off a sketchy power bank. But any device worth preserving, like a phone, won’t the phone figure out what kind of current it’s going to take anyway?

Sounds like it’s protection in case both the charger and device have issues and want to dump all the juice real fast.

>> No.1770859

It's about protecting the power source. If the device tries to pull 3A out of a power bank or wall wart that can't handle it, the output voltage will drop or it will trip a fuse or something. The power source needs to tell the device how much current it's allowed to request.

>> No.1770865
File: 369 KB, 828x990, EB348933-60F5-428F-A666-2343EC7095D5.jpg [View same] [iqdb] [saucenao] [google] [report]

Hmm ok, I figured it was impossible because of some physical limitation, like if an iPad capable of charging @3A was charged into a cheap 800mA car charger, the car adapter wasn’t going to overheat or anything and it’s going to put out whatever it can un less the device decides it wants less.

I need to get on more Big Clive videos.

Somewhat related, I saw that pic related was on sale, I didn’t know Klein even made those. I kind of want to grab it just to play with, I’d be curious to see what the new iOS is doing with the “smart charging”, it knows your schedule and will do a slower charge if you plug it in at bedtime, it knows you wake up at 6am so it will hit 80% an hour after bed and then wait until 5am to hit 100%.

There is a $39 model too where you can cap the output and send a max of 1A, 3A, or 5A.

>> No.1770883

Doesn't look too bad, but there are ones with more features. My one is USB 2 only, but if you ask/look about on /g/csg there are a few decent USB C ones, and they're probably popular enough to be found on amazon or from US warehouses on ali, in case you don't want the kung flu. Even my cheap one can measure battery capacity and the like. Also of note are USB loads, which can be pretty bulky but are nice for testing the I/V curve of USB power supplies.
But that all implies doing power electronics with USB, which I'm just not into; I'd rather just grab a cheap 2nd hand 12V power brick.

>> No.1770892
File: 239 KB, 463x527, D9308C46-2A03-4D4E-96FE-9F946596F484.png [View same] [iqdb] [saucenao] [google] [report]

That $10 Klein one measures capacity too, there is a mAh measurement on it.

There are 3 version, the $10 one I posted, the $39 one that will test max draw from a power source (0.5-1-2-3 Amp), then there is a $29 version with USB-C and regular USB...which I’m confused looking at, I thought Type C was the output like Micro USB and Lightning connectors. Is that something new, Type-C to Type-C cables? Never seen them on anything but I‘m a boomer.

>> No.1770897

>Is that something new, Type-C to Type-C cables?
Yep, USB C is being pushed as the new USB standard to replace both USB A and mini/micro USB. Like how 8P8C or BNC is the same on both ends, with male on the cable and female on the devices. A bunch of newer thin laptops have no USB a ports at all, just USB C. The new chargers are smart enough that you can plug a USB C cable from the power supply to a laptop and the laptop will tell the charger what voltage it wants. It's also USB 3, so much faster than USB 2 and good for backup drives and the like, making it a good connector for both power and data on devices of any size. But since everyone has USB A flash drives, mice and keyboards, wireless dongles, etc. it's a somewhat unwelcome change. The connector itself is also somewhat more delicate than USB A IIRC, and all those data pins are kinda useless and a pain to solder on devices that just need it for power. Thankfully complete USB replacement is only really catching on with the laptops, where all those things are somewhat optional.

We should be using coax instead, with filters separating DC for power and different frequency bands for different data formats. A more modern 8P8C replacement with PoE wouldn't be unwelcome either, but that's just bringing up another divisive standard.

>> No.1770901
File: 6 KB, 266x189, download.png [View same] [iqdb] [saucenao] [google] [report]

I have a car stereo without line outs, and am trying to connect to the original car amp

Can I take the line outs from before the headunits amp?

>> No.1770903

nah it's just a MOSFET. might be a bad idea to route it under an inductor tho

maybe, the signal level will not necessarily be consumer line-level and may require massaging

>> No.1770906

I decided to make Signals and Systems/Signal Processing my sub-specialty of my EE degree.I thought I was just gonna be fucking around with filters and doing cool transformations on signals, but apparently it's one of if not the most math-intensive EE specialties out there short of RF engineering, and apparently the math is really crazy

What kind of math are we talking here?

>> No.1770911

Depends on how deep you want to go. Most complexity comes from stuff like the Fourier transform and especially the Laplace transform which is a more general case of the FT. The actual math of LT is fairly complicated if you need to know WHY and HOW it works. It requires complex analysis, series, convergence, complex integration, contour integrals etc etc. But most likely you won't even need to go down to that level and it will be just plug and chug. You will be solving DE of your typical RLC circuits but the input voltage can be something funky like the unit step so instead of jwt you will be using st whatever that s is depending on FT or LT it is either a complex function of a real variable or a complex function of a complex variable which is just a four dimensional case. You will be drawing poles and zeros, etc but most of the time you will using the ready made formulas based on well know signal types so nothing to worry about. You will also need to understand stuff like convolution and window functions, and the sinc function (sinx/x). It is not hard but it is something you've probably never seen before.
And you definitely need to be able to solve 1st and 2nd order ODE. And need to understand the concept of improper integrals. And you also need to have a solid background in Calc 1: exponentials, logarithms, trigs, complex numbers, the Euler formula and all kinds of function transformations: time shifts, squashing, stretching. It is not that hard but is very important.

>> No.1770921

I was thinking it was some sort of hardcore abstract algebra/analysis level stuff. Honestly, none of that sounds much more difficult than (desu it sounds significantly easier than) my current ODE class

I'm a bit of a math (graphs, visualization, and transformations really) sperg so I'm actually well familiar with both convolution and the cardinal sine function. Convolution is one of my favorite mathematical operations just because I think it's so damn neat

I'm so interested in signal processing because I've been writing electronic music since I was 10. Now/since that I've learned that all of the software synthesizers have their basis in electronics, mathematics, hardware, and also digital signal processing, I've been seriously hyped about this specialty. Going back to being a graph sperg,there's something about being able to control the graph or output of a periodic function by varying its coefficients (am/fm/pm synthesis) or various other transformations. That, and also the various relations between functions that alter their output or create entirely new outputs like convolution, multiplication, integration, differentiation, laplace/fourier transforms, composition... you name it

I could babble on forever but I think I've found the right degree for me

>> No.1770933
File: 42 KB, 1145x622, 1578444160445.png [View same] [iqdb] [saucenao] [google] [report]

good, you are golden then. sounds like a perfect match. funny that you mention abstract algebra. of course it is even more optional than complex anal or fun anal (fourier specific stuff like the hilbert space etc) but i think there is a grad level material that generalizes both FT and LT into something called gelfland transform: it something about banach algebra that can be the group algebra and therefore homeomorphic to R, and thats your FT, or if it is the half-line and homeomorphic to C then is the LT. once you start learning something there is really no finish line in mathematics. or rather it is an asymptote.

>> No.1770935

>complex anal or fun anal

I'm going to refresh myself on what the difference between the laplace transform's and fourier transform's uses are, then come back and ask what the gelfland transform is useful for.

>> No.1770968

Ok never mind it doesn't make enough sense to ask about. At least now i have a better idea of what I was doing when learning poles and zeroes, I think. I wonder if the laplace transform will help me understand friction in a non-oscillating system, where it has a v^2 term.

>> No.1770993

Who makes the best 3.5mm stereo jacks? I'm working on a synth and want to swap the current jacks with ones that won't just get all loose and shitty over time.

>> No.1770994

can't go wrong with Neutrik
or maybe Switchcraft's higher end stuff

>> No.1771018

>feed current back into the source.
what does that mean? i need to connect the primary and secondary somehow?

>> No.1771035

No. A resistor's current is directly proportional to the voltage, and shown by the relationship:
> I = V/R
But reactive components, inductors and capacitors, behave differently. For a capacitor:
> I = C*dV/dt
and for an inductor:
> V = L*dI/dt
> I = (integral of V dt)/L
Looking at these equations, you can see that a very fast voltage transition across a capacitor (high dV/dt) will cause a big current spike, and a very fast current transition (high dI/dt) will cause a big voltage spike.

For a sinusoid, this means that the current will lead or lag the voltage by 90° depending on the component, and will be a sinusoid scaled as a function of frequency (higher frequency means more current through capacitor or less current through inductor) and of inductance/capacitance (higher capacitance means more current, higher inductance means less current).

But the important takeaway is that for two of the quarters of each AC cycle, the current will be of the same polarity as the voltage (i.e. energy being stored into the inductor/capacitor), and for the other two quarters of the AC cycle, the current will be of the opposite sign (energy coming back out and into the power supply). This latter situation means that if the voltage is positive, the current will be flowing back into the live terminal. The net power consumed by this is ideally zero, even though there is a current flowing, and this is what we refer to as reactive power. Practically, the energy going back into the grid is stored as kinetic energy in the giant rotating flywheels of wind/hydro/gas turbine generators.

So if you have an inverter or amplifier or whatever that's outputting a sine wave but won't let current flow in the opposite direction of voltage through an inductive or capacitive load, you won't see the same result as with a proper power supply. For example, an emitter follower won't work, and I don't think a common-emitter amplifier will work either. But totem-pole outputs should work.

>> No.1771223
File: 232 KB, 373x693, 20171231080436!FS_Shinmyoumaru_Sukuna.png [View same] [iqdb] [saucenao] [google] [report]

Best T12 tips? Looking to get a pair for general soldering, thinking about getting a K along with a C4/BC2.

>> No.1771253
File: 1.30 MB, 790x570, counters.png [View same] [iqdb] [saucenao] [google] [report]

Counters gonna count!
Trying to design an expandable version of my 4017 switch encoder circuit. Unfortunately expanding becomes more complicated than I expected. It seems I have to ditch this design and go with a string of CD4021 parallel->serial shift registers.

>> No.1771260
File: 12 KB, 609x109, pure sine wave battery power.png [View same] [iqdb] [saucenao] [google] [report]

what did /g/ mean by this

>> No.1771287

D12 looks good for SMD stuff, while the K serves me well for THT.

>> No.1771296

idk the context but he's referring to the output signal of an inverter. Cheap ones use some kind of multileveled DC trying to emulate a sine wave. Getting a pure sine wave out of an inverter is expensive

>> No.1771300

Those 4017s are decade counters, right? That's an unusual way to decode buttons. With PISO shift registers you'd be able to simply chain the Q8 from the last into the Serial Input of the next. Though the 4021 is a bit odd in that it has Q6, Q7, and Q8, unlike the 74HC165, which just has Q7 and NOT(Q7) outputs (starting at 0). The 165 also has a seperate asynchronous parallel load, which may or may not be helpful in your case.

>multileveled DC
Isn't it just PWM?

>> No.1771313

It's /g/, so they probably meant "we're absolutely retarded" with a few extra words thrown in.

>> No.1771315

In the trade, it's called "modified square wave"; it usually has two or three steps at most - and yeah, it's shit.

>> No.1771328

>and yeah, it's shit
>implying anything needs a perfect sine anyhow
>implying anything needs AC in the first place
unless it's a microwave or has an induction motor, in which case you probably shouldn't be running it off an inverter anyhow

>> No.1771329

>Getting a pure sine wave out of an inverter is expensive
What's the reasoning for this? Square waves induce all kinds of nasty garbage spikes and noise with inductors.
Meanwhile, a high frequency PWM synthesizing a 60Hz sine is easy af to design, cheap (works with MC's and MOSFETS), and doesn't require a fuckhuge transformer.
How are sines more expensive?

>> No.1771331

>CAUTION The buzzer needs high voltages of the order of 100 volts to work. While experimenting, your hands will get wet with whatever you use for ink. Do try not to electrocute yourself...
Yeah... this is exactly the kind of thing I was trying to avoid.
I know this is /diy/ but I was hoping to get away with DIY-ing everything BUT the printhead.
I am finding a few other resources on this wiki about working with commercially made printheads though, so thanks for that.

>Not sure if you are still here
With the power of technology (browser extensions) I can add a thread to my thread watcher and completely forget that it exists until it turns red about a week later.

>> No.1771348

It means switching at a higher frequency, so more switching losses, and hence more heat sinking and more complicated MOSFET drivers and more expensive MOSFETs. Also probably a negative feedback circuit. Just chuck a small filter after the square wave and it's good enough for 90% of electronics. Heck, 100VDC is good enough for 80% of electronics.

>> No.1771349

>What's the reasoning for this?
It's true, getting a pure sine wave IS expensive. If you don't believe, you can take apart a cheap "modified sine" inverter or UPS and a "true/pure sine" inverter and see how they compare. True sine inverters have a lot more components for switching and filtering. I believe it's because the high switching speed requires better FETs and more complicated gate drive circuitry, and a lot of filtering is needed.
Modified sine is very simple by comparison, it's just a single DC voltage switched into either polarity, and switched much more slowly.
SMPSs, found in computers and everything else, don't care. They would even run fine on DC, except maybe overworking two diodes in its bridge rectifier.
Electrical noise is, right, nasty. But it's worth it for the lower cost, size, and weight.

>> No.1771366

I thought the negative feedback was purely a buck/boost converter thing? Since the length of the duty cycle can nonlinearly change the output voltage. Why do sine inverters need feedback?
Maybe we're looking at different things. The two sine inverters I'm seeing involve three matched diff pairs and some mosfets, and another that uses like 10 pairs of mosfets. Neither of those looks like more than $10 worth of parts.
>high switching speed requires better FETs
The trouble with square waves is, the first noise harmonic you have to filter off is at 180Hz. You need a pretty complex filter for that, and the inductors would get pretty yuge. But if you bump that up to say 1-2kHz, the PWM frequency becomes much easier to filter away from 60Hz, possibly requiring only a single pi network.
What's the typical switching speed of a mosfet, anyways?

>> No.1771374
File: 13 KB, 577x76, IRF1404.png [View same] [iqdb] [saucenao] [google] [report]

>What's the typical switching speed of a mosfet, anyways?
IRF1404 is a typical mosfet I think.

>> No.1771377

>Why do sine inverters need feedback
For non-ohmic loads that don't draw a sinusoidal current, like a bridge rectifier feeding LEDs. In this example, the current will be seen as spikes at each voltage peak/trough. The output voltage will get skewed from the ideal sine, assuming the PWM is filtered by an LC filter that is.

Just like how class-D amplifiers use negative feedback.

>> No.1771416
File: 61 KB, 403x307, 1574384132898.jpg [View same] [iqdb] [saucenao] [google] [report]

No, no, IRFP460 in the chad to-247 package.

>> No.1771551
File: 20 KB, 444x266, R7545336-01.jpg [View same] [iqdb] [saucenao] [google] [report]

>to-247 package

>> No.1771567

Where do these books rank in the 'shit-to-learn' reading list?
I have just finished Mims and am looking for next steps.

>> No.1771571

Just looking at the titles they look introductory to intermediate. Probably worth reading through, and just skimming past areas that you're already familiar with, they're free after all. They're also being updated regularly, so they're probably quite up to date, more so than conventional textbooks.

>> No.1771580

If i have a micro powered by 5V DC and i want to feel 5V AC into one of the ADC pins, i do it through a diode so the micro only recievies the positive top of the sine. But how do i connect the ground wire for the AC? do i just connect the DC ground of the micro with the AC neutral wire of the 5V ac source? (they are both isolated power sources)

>> No.1771581


>> No.1771590

what do you want to measure the AC for, frequency? Linearity? You could use resistors to bias the signal so you could measure a full sine from 0V to 5V if you wanted. As for which wire is which, if they're both isolated from each other it doesn't matter. Swapping "live" and "neutral" wouldn't make a difference. If you just want frequency, you'd be better off chucking it into a common-emitter/source transistor and feeding the result into a digital input, no need for analog.

>> No.1771596
File: 32 KB, 709x782, max0419-09-msgeq7-mono-circuit[1].jpg [View same] [iqdb] [saucenao] [google] [report]

it's for audio signal i have this chip and the show where to connect the live wire from the audio jack (the audio _in)
but they don't show where to connect the ground from the audio jack

>> No.1771600

It's AC coupled with a capacitor, it literally doesn't matter

>> No.1771603

so i don't have to also couple the audio ground and can just directly connect DC and AC grounds together?

>> No.1771708

Fucking chinks. They lure you with low prices and I am not even talking about quality, I don't care about quality at this point I just want to know what I am buying. Why does it have to be so complicated? They list an LCD with a certain controller, then link a datasheet for a totally different controller and also link some driver code for YET ANOTHER controller. WTF. Every fucking time. I mean what does low cost have to do with this mess? If you sell cheap shit that might not last or have inferior parameters, I get that. But why can't you at least provide the correct specs and part numbers. I mean it is not just a single case, it is the pattern, thats how they operate. I don't get that.

>> No.1771739
File: 134 KB, 1106x830, DSC_0133.jpg [View same] [iqdb] [saucenao] [google] [report]

>able to simply chain the Q8 from the last into the Serial Input of the next.
Yes, indeed this is the great feature for this application. With CD4021 expanding is much easier than with 4017s. For my application (a small game) I can easily make two boards with dual 4021 registers. If I ever add more rotary switches, I can just continue the string with yet another 4021 board.

>> No.1771803

yes, tying the dc and ac grounds together is fine in this case. the cap blocks dc, and presumably this chip re-biases the signal on its side of the capacitor to a dc level it likes. take care that your input levels are within the MSGEQ7's recommended limits

fleabay dropshippers can't be trusted to know what they're even dealing in

>> No.1771806

I mean, there are dozens of sellers on alibay. you can choose one that knows what they're selling and knows what they're talking about. a discrepancy like that would be enough to convince me to move on to the next seller. you get what you pay for, so survey the market and skip sellers with prices too low to be even close to reasonable

>> No.1771875

They probably all copy each other because they don't have a good understanding of what they're selling or how to speak english, resulting in compounding issues.
You've seen how they all use the same images, right? Chances are they don't even have the product itself on-hand and are all just shipping it from the same warehouse. Instead look at websites like Adafruit and Sparkfun to get the real specs, and buy them from the chinks for cheaper if the pictures match. If what arrived doesn't match the picture, that's pretty good grounds for a dispute.

>> No.1772061
File: 12 KB, 300x200, bh_49.jpg [View same] [iqdb] [saucenao] [google] [report]

How do you measure core loss in a low mu toroid (like 18mu->35mu)? I need to trace those graph like you find in core datasheet where the losses in W/cm3 are function of B for different frequency.

Is there a good alternative to the Iwatsu SY-8218?

>> No.1772113
File: 30 KB, 391x265, 50d770cd99c0efcfca9ae87183154d50.jpg [View same] [iqdb] [saucenao] [google] [report]

IGBT is gigachad. Or Big Mommy if that's someone's cuppa. Pic unrelated but I'm considering putting this together to annoy a couple neighbours.

>> No.1772118

Wouldn't you just compare theoretical impedance (wire resistance+freq dependent reactance) with measured impedence, using the Q of an LC resonator?

>> No.1772215

I guess it might work, but how do you drive the coil? You around 20A between 1-100kHz, any good amp recommandation ?

>> No.1772253

A normal class-D might work for that, 100kHz isn't that far above audio. Though I can't say I know how good they are at driving reactive loads.

>> No.1772260
File: 25 KB, 1359x646, wtf.png [View same] [iqdb] [saucenao] [google] [report]

How can I not fry my power supply and inductors? Am I simulating it wrong?

>> No.1772267

Tips on how to get that sort of job? I'm a pretty average student

>> No.1772317

Simulate some ESR in the inductors. Also what topology is that? Pretty sure most power supplies don't have DC through inductors, looks like you're just shorting Vcc to ground through your inductors and FETs.

>> No.1772333
File: 96 KB, 1342x2409, InductionHeater.png [View same] [iqdb] [saucenao] [google] [report]

Thanks man.

>> No.1772402

real components won't be perfectly balanced. also maybe unbalance the sim a little, like nudge any of R1-R4 by 0.5-1% in either direction

>> No.1772408

Adding "startup" after your spice command will start your voltage source from 0V at the beginning and introduce that add a bit of unpredictability the start of the sim, which is usually enough. Takes like a microsecond though, so I often just use a pulse voltage source that starts at 0V and jumps to Vcc in 1ps.

>> No.1772597

don't really have a lot of tips to give. That particular "job" was just what a consulting company said they were going to put me forward for, but I ended up not signing for them because they were a bit sketchy. The good thing about consulting is that you get to see different things but that also means that if you have a clear idea of what you want to do you'll have to compromise sometimes.
I was a pretty average student as well, what I saw made an impact in interviews was personal projects. In your case, everyone who comes out of uni with a degree can say they know power circuits, but if you built a power supply by yourself (one with feedback and all the trimmings) for example, you will show passion, initiative, practical experience, etc. that few others will have. I put my hobby projects on my CV (as a new graduate of course) and interviewers ate that shit up.

>> No.1772648


Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.