[ 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: 84 KB, 620x425, 1496466351918.jpg [View same] [iqdb] [saucenao] [google] [report]
1215446 No.1215446 [Reply] [Original] [archived.moe]

bump limit reached on old thread >>1209426

pastebin.com/9UgLjyND (embed)

>I'm new to electronics, where to get started?
There are several good books and YouTube channels that are commonly recommended for beginners and those wanting to learn more, many with advanced techniques. The best way to get involved in electronics is just to make stuff. Don't be afraid to get your hands dirty.

>What books are there?

Getting Started in Electronics Forrest Mims III
Make: Electronics Charles Platt
How to Diagnose Fix Everything Electronic Michael Jay Greier

All New Electronics Self-Teaching Guide: Kybett, Boysen
Practical Electronics for Inventors: Paul Scherz and Simon Monk

The Art of Electronics by Paul Horowitz

>What YouTube channels are there?

>What websites feature electronics projects or ideas?

>Where do I get components and lab equipment from?

>What circuit sim software do you use?
This mostly comes down to personal preference. These are the most common ones though:
NI Multisim
iCircuit for Macs

>What software should I use to layout boards?
Circuit Wizard

>> No.1215453

>d-did I do it right?

Looks fine, OP

>> No.1215490

>that smoke
Do they make THC-core solder yet?

>> No.1215494

i have a mystery device that accepts 7v-26v. i hook it up to 8v and it pulls 0.6a. this means it requires 4.8 watts to run.
if i have a battery pack which has 2, 3.7v li-ion cells hooked in series with the capacity of 3500mAh each. the pack has a total 26 Watt hours. which means i can run this device for around 5 hours with this battery pack (or like 3.5h if i want to keep both cells at 20% so they wont die)

or i could hook it up to a 12v PB battery with 7000mAh capacity, that would provide 84 Watt hours and in total keep the device operational for 17 hours

and even though both batteries have 7000mAh capacity, 12v PB battery would still run the device 3 times longer than the Li-ion pack.


>> No.1215499

You don't add charge "capacity" values if the cells are in series, only in parallel. If two cells have say 12Wh each at 3V, this means they have a capacity of 4Ah each. By adding them in series, the voltage across them is 6V, hence their capacity will remain at 4Ah to get 24Wh total energy storage. In parallel since the voltage remains the same, you add the capacities together, because the two cells in series can deliver more current than the cells in parallel, hence giving a larger capacity value for the same energy storage of 24Wh.

>> No.1215509

your maths for the 2 cells is wrong, series adds voltage not current. but lets just say they happened to be 7000mAh.
in that case it depends on the device, if you are correct about its power and the power draw is constant across all voltages (perhaps not, linear voltage regulators are notoriously inefficient because they essentially burn excess voltage as heat) then of course the higher voltage battery will last longer because they power is the same the current draw will be lower, check by calculating the watt hours.

lets say it though worst case that it is a linear regulator or some other voodoo, the current draw might be constant regardless of voltage. in that case its all about the Ah, any voltage over and above the minimum is essentially wasted. but if its a switcher its fine.

if its a resistive load like a fillament lamp or toaster the current will increase with voltage! double the voltage means double the current and thanks to power law the power drawn will quadruple!!! watch out for that bad boy.

>> No.1215520

ah i see now
im a mycologist by trade so i gotta teach all this stuff to myself and confusions happen occasionally
cheers :D

>> No.1215523
File: 2 KB, 439x276, Li-Ion protection.png [View same] [iqdb] [saucenao] [google] [report]

I'm wanting to make a Li-Ion protection circuit, will the following work? I'm wondering how little current I can get it to draw, but I can't find a current/voltage graph of a 3.3V Zener in particular. Would I need to use a different Zener instead because of the voltage dropout of the BJT?

>> No.1215539

What sort of load are you using?

For very simple and minimal current draw using just a MOSFET with carefully selected threshold could work, see http://shaddack.brutowave.com/projects/method_SimpleLiPolyOverdischargeCutoff/

>> No.1215542
File: 105 KB, 1251x784, z.png [View same] [iqdb] [saucenao] [google] [report]

Low voltage zeners are trash and your circuit is rather power hungry even with an ideal zener.

Why not take the protection circuit from some junk battery?

>> No.1215544
File: 30 KB, 598x388, 1485090704925.jpg [View same] [iqdb] [saucenao] [google] [report]

>how little current I can get it to draw
I(load) divided by the beta of the transistor.
>Would I need to use a different Zener instead because of the voltage dropout
Voltage drop, and yes. Your circuit would "switch" off below 3.95V due to V(be) of the transistor.
>will the following work
Looks kinda dodgy to me fampai. I put scare quotes around switch because it doesn't have the snap action you'd probably want. Lithium battery protection is so commoditized today that you can buy the assembled, covers-all-bases solution from China for less than you can build a one-off that just handles overdischarge.

>> No.1215556

It looks like it will work in spice, but it won't let me customise the Zener diode, so I'm stuck with a 4.6V Zener.

>I(load) divided by the beta of the transistor.
So what if I use a darlington? When does the Zener stop Zenering?

Could I use a low-dropout 3-terminal regulator instead of the Zener?

>> No.1215557
File: 122 KB, 1084x640, Screen Shot 2017-07-25 at 12.52.22 PM.png [View same] [iqdb] [saucenao] [google] [report]

pic related

>> No.1215587
File: 164 KB, 855x763, protection_module.png [View same] [iqdb] [saucenao] [google] [report]

Transistors have no 'threshold', that's an erroneous concept. Transistors have a transfer characteristic. And all semiconductors have temperature dependent properties. Look at the Z-diode chart above, it says Tj=25°C for a reason. Listen to >>1215544 and get a cheap and reliable protection module, all else is memetech.

>> No.1215594

Working on a car immobilizer. Going to be my first time working with RFID tags, can someone point me in the direction of some good docs?

Also I'm looking towards implementing AES128, how difficult would this be in terms of writing data to the RFID tag?

>> No.1215596
File: 300 KB, 556x460, medium-ATmega32U4-TQFP-44.png [View same] [iqdb] [saucenao] [google] [report]

I want to buy an MC that has two pins which can read analog values, but i am completely confused by these chip pickers. What is the feature? Is it number of ADC channels? Number of Analog Comparators?

Im making a weird joystick. I'd like to make it without an arduino, but im not sure what I'm looking for to find what I need. I'm using the Teensy2.0 right now and I know I could just buy the chip it uses(ATMEGA32U4 ), but wouldn't it be over kill for my application? Plus isnt it impossible to solder by hand?
Am I just being a pussy?


>> No.1215618

>Low voltage zeners are trash
I see that now, I guess I'll look into the voltage regulators instead.

>Transistors have no 'threshold', that's an erroneous concept.
I completely agree, as with diodes and such, but I assumed that Zeners were somewhat of an exception, but apparently that's just the high voltage ones. I might as well use a forward bias diode for all the good a low-voltage zener will do me. I'll mess about with Darlingtons and regulators and see what I can find, but I'm limited in what LTspice will provide.

Cost isn't as important for this project as ease of assembly is, and I'd much prefer designing a circuit to just buying one, especially as the total cost will be under $6 either way.

>> No.1215624
File: 1.36 MB, 1944x2592, IMG_20170724_232800.jpg [View same] [iqdb] [saucenao] [google] [report]

How do nixie tubes cope with high ambient lighting conditions, specifically sunlight? I have a project that used LEDs but they get washed out in sunlight and the display is only readable at night or when there are heavy dark clouds, casual overcastness is too much light. I'm sure I could use higher power LEDs and different driver circuitry but I like nixies and I want to try something different.

>> No.1215628

You want two ADC inputs, on two separate ADCs might be nice if you want to sample really fast, if not, single shared ADC is fine. The Teensy 2.0 has 12 available, that'd work fine.

If you want to make your own board, what other I/O would you need?

The TQFP ATMega32U4 wouldn't be too terrible to hand solder if you've had some SMD experience, but wouldn't recommend it if it's your first time.

For cheapest AVR for a couple of analogue and low speed USB only something like an ATTiny44A would be sensible.

>> No.1215631

For sunlight readable you want something either super bright, or reflective. Nixies are relatively dim, so not great.

>> No.1215635

It just needs to read two analog signals and out put to usb. I dont htink it needs to be super responsive that i'd need 2 ADCs for this application. I'll give the ATTingy44A book a read

>> No.1215647

Alright, I just tried simulating the FET design from >>1215539 and it works a treat, it essentially pulls no load whatsoever. But the trial and error to find the right FET will be difficult.

>> No.1215657

>What is the feature? Is it number of ADC channels? Number of Analog Comparators?

Number of ADC channels specifies how many pins can use the ADC function. Usually, this is a single ADC being multiplexed between multiple pins.

Number of analog comparators actually tells you how many ADCs are on the chip. If you don't have a specific reason for needing more than one, just a single, multiplexed ADC will do you fine. For a joystick, I can't imagine you'd need more than one. Even a "slow" ADC can take a few thousand samples per second, which is more than enough to read the two axes of a joystick.

An ATmega32 would be fine. It's "overkill" in the sense that it might be more capable than you need it to be, but you're not manufacturing a million units where a $0.07 difference is someone's salary for a year. Not really worth worrying about for one-off hobbyist projects. Feel free to scope out some of the <$1.00 chips, though. Probably is at least one thing that will suit your needs.

TQFP isn't really that hard to solder. The QFN version might take some practice, if you're feeling ambitious.

>> No.1215662 [DELETED] 

How do I make cmake compile a subdirectory of files with certain compiler flags without applying those flags to the rest of the build?

>> No.1215670

The only other thing you might have to concern yourself with is a bootloader. There are ATmega16U4s as well. There might be an ATmega8U4 but I'm too lazy to look.
Jarring juxtaposition but I guess I'd better get used to it.

>analog comparators
You sure they don't mean comparators, as in LM339 equivalents?

>> No.1215671

>You sure they don't mean comparators, as in LM339 equivalents?

I herp when I derp. Yes, it has an analog comparator for, uh...comparing analogs.

But it also does have a 10-bit ADC with 12 channels.

>> No.1215692
File: 2 KB, 439x276, Li-Ion protection 2.png [View same] [iqdb] [saucenao] [google] [report]

Ok, so having a look at MOSFETS, their threshold voltage isn't terribly well defined, and spice just models the maximum. I assume this isn't a rating that often matters to a circuit. But when the datasheet states the threshold voltage as "2V - 4V", what can I expect when I buy one in terms of probability? Does this value change through the life of the FET or under different (possibly controllable) circumstances? I don't really want to put a voltage-drop diode or voltage divider to feed the FET.

>> No.1215703

The datasheets usually have typical Id/Vgs curves.
Vgs(th) naturally has a temperature coefficient and the threshold voltage is defined for some rather small drain current, like 250µA.

You won't have a clear, well defined cut-off point with your circuit.

>> No.1215714
File: 21 KB, 345x201, vaccuum fluorescent.jpg [View same] [iqdb] [saucenao] [google] [report]


we made clocks for airport control towers, where they needed to be seen in daylight, so we used vaccuum fluorescent displays. they were not that different from nixies (powered from 180Vdc) but they can be driven harder and still live a reasonably long life.

>> No.1215719

>You won't have a clear, well defined cut-off point with your circuit.
I'm thinking of grabbing myself a massive (2.2M+) trimmer-potentiometer and a lower-threshold-value FET so I can tune the cutoff, and I don't think the permanent 1.5µA current being drawn from the cell could damage it if I leave the switch on, so that's likely what I'll do. Alternatively, I could put a couple of Li-Ions in series, but smaller is better when it comes to cramming things in a sardine can.

I was planning on removing the cells from a semi-chinkshit USB power bank, and just using the control circuitry on it as a cell charger to charge them one at a time, but I realised that the constant-current portion of the charge would be over-amping an individual cell by a factor of 5, which is perfect for a house-fire. So I could instead build a charging circuit into the circuit I'm currently building, but since it will probably remain fixed to my bicycle most of the time, I figure a USB-in isn't too practical when I'd have to lug my bike up 4 storeys to a USB outlet or bust out the hex-keys. Meaning I'll need to buy another lithium-ion cell (if I'm not scrapping the current power-bank) and construct a charging circuit too. I probably won't scrap the power bank because I just realised that its internals fit perfectly inside a sardine can. So maybe now I'll try and find the simplest way of safely charging a Li-Ion (slow constant-current driven by MOSFET?).

>> No.1215723

Just use a specialized charging IC like tp4056.

>> No.1215725
File: 104 KB, 1126x507, 1493711825948.jpg [View same] [iqdb] [saucenao] [google] [report]

>simplest way
>of safely charging an Li-ion
>everything will be fire
What's wrong with <- this?

>> No.1215742

There's nothing wrong with those in principle, but somebody had to do it the hard way, I guess? It's also more difficult to fit one of those boards in a housing and I hate drawing DIP traces (let alone SOICs or any other damn SMDs). Those are really just half-reasons, the main one being that I'd rather do it myself.

And I'd much rather make something analogue and have to improve it than buy some prefab, digital nightmare and have to make an adjustment to it.

>> No.1215750

Actually, since the total power draw will be up to 3W, I think I'll just package the whole power bank into the can and strap it on the bike, and run everything off the 5V. The shitty linear regulators for the LEDs will waste more power, but the LEDs will only pull 800mW compared to the 2.5W of the EL wire, and It doesn't matter what voltage the flyback transformer runs off, so I'll go with that. No protection circuitry needed, as long as my power bank isn't worth less than the $20 I paid for it.

I'll see about relying on the tiny solar panel in it to charge the cells, though I do anticipate having to take the thing off my bike to charge it every now and then. The circuit also has a built-in single LED "flashlight", which I guess I'll use to power the LEDs, be it through a transistor or not, depending on whether the board looks like it can take 8 times the current through there.

The project is bike lights and EL wire for extra visibility on my bicycle, by the way.

>> No.1215758
File: 923 KB, 1402x1331, chip.jpg [View same] [iqdb] [saucenao] [google] [report]

hi all, was hoping to get some help over here

TL;DR: Looking for the datasheet for the chip on pic related

>Picked up a broken plotter (Canon IPF8000) a while back
>Decided to repair it, easy fix.
>Cannot use it because the maintenance unit is full
>Cannot reset it because entering service mode is only possible when the maintenance unit isn't completely full yet
>Can buy a replacement for $100 or a new chip for $15 but have to wait two weeks

>Wondering if I'm able to reset the existing chip from the maintenance unit
>Chip counts the amount of times the print-head has wiped clean
>After X amount of times it will ask for a replacement
>Looking for the datasheet for the chip on pic related

Any help is appreciated

>> No.1215759
File: 39 KB, 800x800, s-l1600.jpg [View same] [iqdb] [saucenao] [google] [report]

Image semi-related, a replacement chip from China

>> No.1215775

I have an analog output that has a .5V range. How can I get 8bit accuracy from it without breaking the bank. I could shell out another $10 for a new device that outputs with a 2V range.

>> No.1215777


Opamp. perhaps instrumentation variety. But don't respond to me, wait for a guru to show up later today.

>> No.1215791

What's your ADC input range?

>> No.1215800
File: 10 KB, 273x317, lipo_charger_s.png [View same] [iqdb] [saucenao] [google] [report]

>I'd rather do it myself
Ages before the TP4056..

>> No.1215805

Not directly ohm related but I want big buttons on a thing. Snooze button size (impossible to miss). I don't quite know what to search for.

>> No.1215808

something like a mushroom switch?

>> No.1215810

Yeah, that seems good. Ideally it'd just be a tactile button cap. But I haven't found that.

>> No.1215819

looking at buying this one
Since it has 12bit res and a gain amplifier. Was using the one built into an arduino.
Not sure how exactly to answer your question. Im a little confused reading the datasheet.
But I am worried that the amplifier would defeat the point of having high accuracy.

>> No.1215877
File: 130 KB, 805x586, ads1015 pp 10 24.png [View same] [iqdb] [saucenao] [google] [report]

I really know nothing about that stuff but I got me the datasheet and spotted this. Maybe there are more application notes elsewhere. It seems to me that the PGA (programmable gain amp) is exactly what you need to full scale your 0.5V signal. You may lose 1 bit if it is ground referenced (0V..0.5V) rather than differential.

>> No.1215905
File: 436 KB, 1280x720, lcd.webm [View same] [iqdb] [saucenao] [google] [report]

Sup, my LCD is not working with my attiny2313 despite having all the pins connected properly.
I'm trying to solve this by myself, but I'd like to know why does this happen, what are the differences between a conventional arduino and the attiny2313 that allow an LCD to work?

I'm running the attiny at 1MHz, while my arduino runs at 16MHz, maybe that's something to do with it?

Webm related, it displays "Hello World" for a fraction of a second and then the first row goes blank. Maybe it has something to do with refreshing, since it can display Hello World just fine, but the screen goes blank when writing into it too fast.

I'll also might have to make my own library for the display, since the original uses too much program memory (only 2048 bytes in the 2313)

>> No.1215909

Calibration doesn't help? It kind of seems like what would happen if the voltage input increased

>> No.1215922
File: 1.17 MB, 2048x1536, nomemory.jpg [View same] [iqdb] [saucenao] [google] [report]

The contrast is fine and stable. It might've just been a problem with the example with refresh rates that the attiny can't handle, I added a 1000ms delay and that seemed to solve the problem, although there's still the problem of too much memory usage, just this tiny bit of code uses 91% of the 2k, add another print and it goes above 100%.

I would've tested the display displaying a counter, but if I do that there's not enough memory. I'll have to use a lower level programming language and create my own library. I don't know if I can use the arduino IDE to do this or if it doesn't matters as long as I use low level language in it.

I didn't consider the memory issue when I bought the 2313, and running the display, reading a encoder and a button and powering a couple of MOSFETS will be tricky.

I'll rewatch some videos on character displays starting with the 8bit guy.

>> No.1215927


Is only that your LCD is on full brightness, you only need to put a resistor on the LCD brightness controller.

>> No.1215928
File: 1.48 MB, 3024x3024, IMG_2715.jpg [View same] [iqdb] [saucenao] [google] [report]

Trying to fix a radio controlled relay/lamp switch.
This is the radio transmitter. I was 100% positive that the buttons just wore out from mechanical use, but after putting OHM probes on the buttons my probe would beep when pressing those buttons. Does that mean the buttons are fine? WTF is wrong with this thing now?

>> No.1215938

Maybe it's the receiver that has a problem, not the transmitter.

>> No.1215941

>The shitty linear regulators for the LEDs will waste more power
A switch-mode LED driver from scratch could be a fun little project.

Check a datasheet to see how slow the HD44780 is. Some commands require tens of microseconds to milliseconds of delay.

Buttons are probably fine. If the battery is up to voltage, I'd check the receiver end, or test the whole system in a different radio environment.

>> No.1215947

>LED driver from scratch
Discrete design please.

>> No.1215954

A fun weekend project, rather.

>> No.1215961

Found a fuse in the reciever that wasn't wired in, or maybe the solder broke off. Still doesn't work.

What's weird is the device has 2 buttons. Both buttons went slowly over time. I could make it work by pressing REALLY hard, then it'd stop working all together. Then the 2nd button died shortly after. I'm still convinced it should be the buttons.

>> No.1215964
File: 1.43 MB, 2048x1536, 20151011_163235.jpg [View same] [iqdb] [saucenao] [google] [report]


They work great. Despite what >>1215631 says you can up the voltage slightly and get a much brighter display. They may not last 15 years, maybe more like 10 but they will still last quite some time.


Sexy. Typically Nixie styled VFDs can go up to about 30V safely and are very bright.

>> No.1215966
File: 1.68 MB, 4128x2322, 20170101_003416.jpg [View same] [iqdb] [saucenao] [google] [report]

Pic related.

>> No.1215970

It's possible the buttons could have gotten extra bouncy over time. You could try tacking on some cleaner push buttons on the transmitter and see if the receiver works better. Are you getting good voltages around the receiver?

>> No.1215974

I'd remove the battery and take out the PCB. Look at the traces around the switches. Resolder them. Do they still click? Next a drop of alcohol on the button. Push a few times and let it dry. No wait, first lake a piece of wire or a paper clip and short the switch(es) to see if it's the contact or something else.

>> No.1215983


one possibility is that it's not so much the buttons, but the radio range has decreased as the circuit components age and it gets de-tuned. this is easy to test: just move closer, or place the transmitter under your chin when pressing the buttons.

>> No.1216052

How can I find stores that sells components in my area? I'm at the corner of southwest detroit if anyone in the area might have a lead. The closest I can find is in westland, 30 minutes away, and I've never been there. There's a microcenter 30 minutes a dofferent way but from online it looks like radioshack. Am I using the wrong google fu or is it really just that hard to find B&Ms.

>> No.1216053

There are a few suppliers that run will-call windows, such as Galco Electronics in Mad Heights. Palco Electronics in Southgate has a web presence but I don't know if they're still open.

>> No.1216066

>A switch-mode LED driver from scratch could be a fun little project.
I've been thinking about that too, though at those low voltages you're losing a significant amount of power through the rectification diode. I'd just do it with a flyback transformer anyway because I've got no clue how to make a buck or boost converter vary its duty cycle in response to the output voltage.

I'll just see about finding a fixed ~3V voltage rail on the power bank, though I will need 3.6V to run my super-flux (5lm) LED, 3.3V to run the other two white LEDs, and 2V to run the red rear light LEDs. I'll see about finding multiples of LEDs to make the linear voltage drop from a rail as low as possible, but since the EL wire will be using 4 times the power anyways, I'm not too fussed.

>> No.1216084

Can anyone tell me if
Is capable of level shifting or would I have to make that on my own. I worried about doing so because of voltage drifts and other things that make the readings for the input inaccurate.

>> No.1216094

What levels would you want to shift?

>> No.1216104
File: 12 KB, 921x524, 1481481763389.png [View same] [iqdb] [saucenao] [google] [report]

>I'll just see about finding a fixed ~3V voltage rail on the power bank
There probably won't be one that won't drain the cells unevenly, and you don't want that.
>how to make a buck or boost converter vary its duty cycle in response to the output voltage.
For LEDs you're probably interested in current, not voltage. It depends on your particular circuit, but here's an oversimplified illustration of negative feedback. The more current through the LEDs on the secondary side, the more voltage developed across R1 and therefore Q1's B-E junction. Q1 eventually starts to conduct enough to pull down the sawtooth oscillator and reduce the amplitude on Q2's gate, thus decreasing the duty cycle of Q2 and the transformer. The nice thing about current mode is that you can add or exchange LEDs into the series string without characterizing or recalibrating much, as long as it's still within the abilities of your electromagnetics and they all demand the same current. If your LEDs are to run on different amounts of current, they will probably need separate converters.
The same general principle applies for a boost or buck converter: convert current to voltage and use that to retard the oscillator somehow.

>Vcc: 2.0-5.5V
As a rule, inputs should never exceed Vcc, and Vcc should always stay within the specified limits. Don't try to feed it 5V signals on a 3.3V Vcc and you'll probably be fine.

>> No.1216130

2.5-3V to 0-5V.
This is my first time dealing with sensors that dont output 0-5V so Im somewhat lost.

>> No.1216133

2.5-3 from your analogue sensor into the ADC board you mean? Easiest way would be capture it directly and do the offset in math.

>> No.1216141

You've got differential inputs on that DAC. You could just supply a reference equal to the midpoint of your expected signal range to AIN3 and put your signal to be digitized on AIN0..AIN2, then use the programmable gain amplifier to get the precision you need. Downside is you have to dedicate AIN3 to your reference.

>> No.1216142

i don't always derp, but when i do, i herp

>> No.1216156

Well I think id lose the precision of the adc doing that. Id like at least 8 bit res for this sensor.
Im a little confused here. Wouldnt using a gain amplifier cause the 3.0V to go to 6 or move volts and destroy the adc

>> No.1216157

The ADC is differential, not single-ended. It digitizes the difference between two inputs. When doing single-ended conversions, AGND is implicitly the negative end of the differential pair. A differential amplifier amplifeis the difference, and in any case the amplifier can't amplify beyond the voltage or current provided to it.

>> No.1216161

Doing the simple way of just sticking the 2.5-3V in with ground as negative reference gives 2mV resolution, 7.9 bits effective over the range.

Doing it with 2.75 reference subtracted gives a +0.25 - -0.25V range, multiplied internally, giving a 0.125mV resolution, 11.9 bits.

>> No.1216162

>There probably won't be one that won't drain the cells unevenly, and you don't want that.
The cells themselves are all 18650s in parallel, so I shouldn't have to worry about them going out of balance, but after checking I don't think there's a constant logic rail at all, the circuitry just runs directly off the cells. In addition, the USB out outputs cell voltage when not active, so I can't just use a constant 5V rail either. I'll have to make sure the USB remains active, so I'll experiment with how to get that to happen.

When it comes to switched-mode, I'd simply tune the transformer coil ratio until I get the correct voltage for just below the correct current, that way I wouldn't have to worry about regulation as long as I pull it from a regulated output. Technically changes in temperature will change the forward voltage of the LEDs and the rectifying diode, so I'd experiment with those in boiling and icy water, but regardless I think it would be more efficient, assuming the changes don't drop the LEDs by more than a couple of mA.

Also I'll be using 3 different LEDs, all with different forward voltages and currents, so putting dropper resistors is what I'll do in the end to stop things from getting too complicated.

>> No.1216184

I was reading the arduino code it gives as an example and its pretty clear that you shouldnt set the gain to exceed the source. But what you are saying makes sense but to make sure, the gain amplifier doesnt amplify all the pins all the time but only the reading you asked for? meaning it only amplifies the difference. Amplifying the differential reading between 4.5 and 5 would be the same as amplifying .5v and not damage the adc
Yeah I was originally looking for a 12bit adc because its close with no gain amplification. But Id like some wiggle room in case I dont use the full range of the sensors output in some scenario. But using a reference voltage and amplifying the difference would give plenty of room.

I am still interested in level shifting 2.5-3v to 0-4v because I can use faster 8bit adcs, but I can put that off and learn that later after Im done with the prototype. Thanks for your help

>> No.1216186

Shifting and amplifying with an opamp is an option, but it's easier to just do it in the ADC with a 2.75 reference if you don't need all 4 channels.

>> No.1216192

>so putting dropper resistors is what I'll do in the end to stop things from getting too complicated.
Unless I find an elegant, discrete, constant-current source that takes up not much extra power. It would be easy to do if BJTs had a factory-set hFE/beta/gain, but I'll look a little at FETs.

Both circuits will be losing the unwanted voltage through resistive loss, so it all comes down to whether the extra power drain from the current source is worth getting the extra brightness from the LEDs.

>> No.1216227
File: 29 KB, 807x558, 1500540532617.png [View same] [iqdb] [saucenao] [google] [report]

>When it comes to switched-mode
I haven't seen anything about your flyback driver circuit. Will it maintain the appropriate duty cycle over temperature?

>its pretty clear that you shouldnt set the gain to exceed the source
Naturally, so that your readings fall on the scale and don't "bury the needle" so to speak. There's no warning in the datasheet about damage by exceeding the scale, as long as both inputs are between VDD and GND. You'll just get a -FS or +FS reading.
>because I can use faster 8bit adcs
Are you sure the 8bit ADCs don't have differential inputs or references?
So you want fast, do you?
At 30ksps you get almost 20 effective bits, with programmable gain and true differential inputs available.
Anyway, your other problem is that you need to level-shift and amplify. It can be done with one op-amp and a handful of resistors. I used CircuitJS to """"""engineer"""""" pic related. The voltage divider at the top centers the input on your desired output voltage, 2V. The bottom divider sets the bias point and programs the gain of the op amp to 8 * the attenuation of the top divider 1.37 = 10.96 - 1 since noninverting = 9.96 ~= 10. The 2V reference source should be low-impedance, or you should increase the bottom divider resistors proportionally.

>> No.1216229

The warning is in the example code
>// The ADC input range (or gain) can be changed via the following
// functions, but be careful never to exceed VDD +0.3V max, or to
// exceed the upper and lower limits if you adjust the input range!
// Setting these values incorrectly may destroy your ADC!

>> No.1216232

That adc and your diagram look awesome btw. I really need to get a circuitry built to figure out how to make circuits like this. Thanks

>> No.1216234
File: 3 KB, 400x200, EL wire waveforms.png [View same] [iqdb] [saucenao] [google] [report]

>flyback driver circuit
Literally a multivibrator feeding a hand-wound 20:1 toroidal transformer with a stupid amount of filtering. Chances are I'll try to get something sinusoidal on the input side so it won't actually be a "flyback" transformer at all, but this is for my EL wire which runs off AC anyway. I'll also have to leave the transformer floating so I don't get negative voltages fucking up my 0V reference or whatever, or messing with the "duty cycle" of the EL wire.

In pic related, if I leave ground of the secondary attached to ground of the primary, I'd either get the top image if I had only one secondary, the bottom image if I attached the other terminal of the EL wire to another out-of-phase secondary winding, but I think if I leave it floating it will be the image on the right, which is easier.

The LEDs will be powered by some manner of linear power supply, because smart filtering is whack.

>> No.1216237 [DELETED] 
File: 64 KB, 853x480, Sketch.png [View same] [iqdb] [saucenao] [google] [report]

This is how I'd approach it, current mode hysteresis control buck for your main headlight, will keep running full power to ~3.65V if you use a nice schottky. The lower power ones I'd just go LDO configured for constant current.

>> No.1216241

That's interesting, to put it mildly. If I had one on hand I'd risk it to test that warning.

Aliexpress, fâm. It's the maker's paradise. As for the circuit design app, see
It's bretty comfy and really fun to watch. I implore the next OP to add it to the sticky.
When I actually know what I'm doing and don't need to simulate, I use KiCad nightly and its built-in schematic capture module because it's a short hop from there to a board layout and thence to OSH Park and physical boards. Then out comes the solder paste and the tweezers and the heat gun, whee!

Ah, cool. Winding your own magnetics is pretty punk, anon. Cheers.

>> No.1216246
File: 48 KB, 892x436, spectrumAnalyzer.png [View same] [iqdb] [saucenao] [google] [report]

I'm designing an audio frequency spectrum analyzer but there's a bit of a twist, I want a swept tuned architecture rather than a real-time or FFT based analyzer. I obviously want to be able to view the entire audio range so 20Hz to 20kHz and I want a resolution bandwidth of 20Hz or less. To avoid high Q filters I plan to utilize several different mixing stages like in pic related. What I am not sure about is the various local oscillator frequencies I nor do I really know the bandwidth and the cutoff for all the required bandpass filters. I also do not know the sweep time so I don't know what frequency to set the ramp oscillator at.

The mixers will utilize NE602 Gilbert cell mixers which are designed for VHF but it seems like they'll work at low frequencies too. I did test mixing audio with a 30kHz LO using a Gilbert cell multiplier using discrete 2N3904s and it worked okay, I observed what I believe to have been DSB-SC modulation but it might've just been AM and I'm retarded. Either way, I'd figure the IC would work just fine if discrete transistors did, even better due to temperature compensation and matched transistors.

Also do I go with active filters or LC filters for the bandpass filters? Even with the multiple mixing stages I still need relatively high Q of <=10 (which is better than 10,000 but still).I also want a sharp roll off. I should be -60dB down within 1/5 of an octave or less. I don't want to have 10-pole Butterworth or Chebyshev filters but it's kinda looking that way.

>> No.1216247
File: 154 KB, 627x536, latest[1].jpg [View same] [iqdb] [saucenao] [google] [report]

hey, how hard do you think it would be to recreate Light's drawer from Death Note?

looks to be a simple circuit in concept, but I don't know what could make the plates that act as the security switch. also finding a way to secure them to the base of the drawer without giving away that they're there could be tricky, but possibly done with solder or super glue

the only problem I can see is securely putting the top back down without setting it off could be tricky

>> No.1216258

I meant to ask, but is a particular opamp needed here

>> No.1216261

Not for the low-bandwidth sensor stuff I imagine you're after.

>> No.1216263

Sounds like it'd be fairly hard for you.

I suggest you don't burn your house down.

>> No.1216268

>Winding your own magnetics is pretty punk
I fucking hate it. I've got a small reel of 33AWG magnet wire halfway around an iron core that I was going to use to broadcast audio as radio and then receive it with a solenoid hooked up to a high-gain amp that I plug headphones into (an Elektrosluch) but I am so god damn fed up with winding that thing that It's in the "to be continued" bin. I'll buy another reel of wire because I don't want to give up on that yet. Maybe I'll try mounting one of my microwave turntable motors and gearbox assemblies to hold the ferrite and put a variable frequency voltage to it attached to a foot-switch that controls the frequency, but it's well and truly on the back-burner for now.

So when dealing with a few watts at 400-3000Hz, how many turns should the primary be? Looks like 28AWG is the way to go, maybe thicker. I'm hoping to use one of my existing 25mm OD, 15mm ID, 10mm width ferrite toroids. Does this depend on my waveform (hopefully sinusoidal, maybe triangular)? Also since this is 100V at ~1kHz, it's going to hurt when I touch it isn't it?

Oh shit, my power bank has my Li-Ions at 4.23V and it's still displaying 3/4 charge lights, should I be stopping it?

>> No.1216269

Not Death Note, but Hentai Note. Maybe CP Note if you're into that, I don't judge.

>> No.1216272

what do you think would work for the plates that connect the circuit?

>> No.1216273

>Aliexpress, fâm. It's the maker's paradise
that reminds me, everyone ITT watch this if you haven't already: https://www.youtube.com/watch?v=SGJ5cZnoodY

(You can also find it split up in smaller parts if you have ADD)

>> No.1216274

I'd suggest spring-steel sheet, so you can be sure that the contacts will keep a force on the match, but there is a better way, which is to use an normally closed reed/magnetic switch and a magnet positioned such that lifting the cover turns the circuit on, or just use a photodiode so that it goes off when opened under light, which is much more convenient for opening it yourself. But seriously, don't start a fire indoors.

>> No.1216275

don't worry. where i'm putting this, it's practically all non-flammables (no carpets, furniture, etc)
I'm just putting something inside that I'd rather keep secret, even if I can't keep it
and I'll have a fire extinguisher in the room

>> No.1216279

Anybody here know enough science or geometry to tell me how to figure out the solid angle of light (for calculating lumens from candela) coming out of a thin, luminescent wire? I've got its intensity/area (lm/m) value, and want to compare its flux to some LEDs, but I haven't a clue where to start.

>> No.1216282

Alright really appreciate the help. Im going to research more about opamp circuits and how they work, but this helps me plan what to buy now and continue with prototyping.

>> No.1216286

I thought about using a gaussian surface, but luminous flux going one way doesn't cancel luminous flux going the opposite direction, so I can't cancel out the light going in different directions and treat it as a 2-dimensional problem. Also I meant I've got its cd/m value. How the fuck did they work out its intensity in the first place?

>> No.1216307

why not use spice

>> No.1216308

what the fuck you. Just think of it as infinite and do a cylinder arround it. Then you have the flux per lenght. Jesus fuck people

>> No.1216309

Also what kind of stuff uses candela? It's retarded. Just use lumens

>> No.1216475

Madison Heights is still a bit of a drive, but the willcall option would be incredibly convenient. Definetly going to try them out. Sameday is better than 3-4.

Southgate is 10 minutes down the road though, but all I see on their website is HAM, CB and soldering irons. Would be cool to have something that close though, i'll call them up.

>> No.1216478

Also, forgot to say thanks. I appreciate it.

>> No.1216528

2D angle is not the same as solid angle, I cannot compare the luminous flux to that of LEDs if the units aren't at least intuitively similar. I'm using this EL wire from Adafruit:
Candela are useful when you want to know how bright a light source is from a single direction, which matters if you've got LEDs with different viewing angles and hence different solid angles.

>> No.1216530

Why you don't calculate how much power the light gives and then calculate the intensity at the desired direction?

>> No.1216574

That's probably what I'll have to do, except the other way around because luminous power = luminous flux = lm, and intensity = cd. I'm wondering if I can use the solid angle from a point being irradiated to the EL wire itself, but it would be dependant on distance and on the length of the wire. Wait, could I just say that its 4π steradians because the viewing angle is very nearly 360°? I'll use that approximation for now.

>> No.1216576

fucking hell are you trying to induce my MK ultra training or something?

>> No.1216602

Ok it's giving me 232000 lm/W, which is inconsistent with conservation of energy. I'm calling bullshit on this datasheet.

>> No.1216603

>4π sr
I don't think so, it's not a luminous point (or sphere), it's a line. If you look at it from an angle there should be some cos(phi) effect, similar to a flat LED.

>> No.1216652

> how to figure out the solid angle of light (for calculating lumens from candela) coming out of a thin, luminescent wire?
You can't calculate it. Either it's stated by the manufacturer, or you need to measure it.

For an omnidirectional source, divide flux (in lumens) by 4π to get intensity (in candelas). For a wire, you can probably assume that it's independent of radial direction (perpendicular to the wire), but you'd also need to know how the intensity varies between radial (perpendicular) and axial (parallel).

>> No.1216721 [DELETED] 

Kinda late but I tried replacing buttons, the battery, and even compared it with an intact working unit. I can't find anything wrong, but I'm thinking the relay is what's dead. If the relay was having trouble clicking over, then it might take several presses of the buttons, or several tries for it to stick, which is what made me think it's the buttons. I don't know how to test the relay, Probably will just junk it and buy a new one.

>> No.1216722

This is kinda late reply, but I tried replacing buttons, the battery, and even compared it with an intact working unit. I can't find anything wrong, but I'm thinking the relay is what's dead. If the relay was having trouble clicking over, then it might take several presses of the buttons, or several tries for it to stick, which is what made me think it's the buttons. I don't know how to test the relay, Probably will just junk the entire thing and buy a new one.

>> No.1216743

>So when dealing with a few watts at 400-3000Hz, how many turns should the primary be?
I never that into electromagnetics, fama. Last time I tried it 20 years ago my freshly minted EE friend rolled his eyes and sent me some samples of his new employer's TO-220-5 simple switchers. daycounter.com has a load of neat little browser-based electronics calculators. Coil calculators may be among them.
>Oh shit
"Never trust anyone, not even your favorite meter."
The charger's probably in constant voltage mode. If the cells aren't too warm, you should be fine.

Akihabara shall never die!

Don't usually need it, also too lazy.

no problemo

>> No.1216747


a relay that refuses to click in on occasion can be persuaded by some vibration. tap the relay, or the whole unit, with a pencil or back of a screwdriver.

>> No.1216763

Yes, whack the relay. Sometimes relays need that.

>> No.1216772

It stoped charging them when my meter read 4.26V or so, which I guess is fine. There was a flyback transformer calculator in there, but it required a bunch of measurements I don't yet know, so I'll wing it and see what I get. 19-23 ga on the primary, 33 on the secondary, 30:600 turns, give or take my patience. I might need twice the turns on the secondary, because I'm not sure if the ground to peak waveform of the primary will have an effective amplitude of half its maximum when the transformer is concerned.

I want to be able to adjust the frequency being generated from 400Hz to 3kHz just so I can slightly alter the colour of the EL wire (pic related), so I guess I'll be over-building the thing. Since I intend on using a sinusoidal waveform I hope I can avoid unwanted inductive spikes, provided I put a few capacitors around the place.

>> No.1216834

I bought an arduino uno from ebay a while back.
Im trying to use it for the first time now. When I plug it into the computer it doesnt show up on a serial port.
Anyone have experience with the serial ports and chinese arduinos?

>> No.1216835

So, since a swept-LO analyzer uses the LO to pick out one particular frequency, in an ideal world mixers and LOs would be perfect and you'd end up with exactly what you wanted at the IF. Unfortunately LOs have harmonics and subharmonics and maybe spurs, and mixers have distortion and spur products.

In general the band plan is engineered to avoid mixer spurs popping up where you don't want them. So as an exercise, pick some reasonable numbers for your LOs and map out where you expect your spurs. Then calculate how steep your filter skirts need to be to reject the spurs. Repeat until you have something that works.

>> No.1216853

What is the most simple timer you can make to flick a electronic switch? Time period being hours

>> No.1216857

A capacitor and a (variable) resistor for the delay, and a MOSFET's gate tied in between the two to do the switching.

>> No.1216858

Im pretty rookie with electronics so I have a followup question. The way your circuit works is that once the capacitor reaches the mosfet threshold voltage the capacitor discharges and activates the mosfet twitch?

>> No.1216859
File: 188 KB, 1362x1336, Screen Shot 2017-07-27 at 5.38.55 PM.png [View same] [iqdb] [saucenao] [google] [report]

Here. The voltage at which the FET turns on varies from FET to FET, and a probably over temperature and other things, but it should be within 10min either way. The time it takes for the capacitor to charge is proportional to R*C (its time constant), and I'd advise picking up a cheap supercapacitor for this little circuit because its pretty hard to find potentiometers with such high values. If the FET being an unknown isn't to your fancy then I'd definitely recommend using a comparator to trigger once it hits the correct voltage threshold, and then you can adjust what voltage simply by staking either terminal of the potentiometer to the high and low rails and feeding the variable tap to the comparator's input next to the RC output.

>> No.1216862
File: 179 KB, 1510x1360, Screen Shot 2017-07-27 at 5.51.29 PM.png [View same] [iqdb] [saucenao] [google] [report]

That middle 100MΩ isn't needed at all, btw. Here's the comparator design with the resistors on either side of the comparator's "-" input being the potentiometer, the 1kΩ resistor can be a normal resistor of whatever size is convenient. When it comes to comparators it's best to keep the threshold voltage within the middle 80%, because they don't respond well to the rails. The FET on the end can also be any transistor or relay in this model, provided the comparator can source enough current.

>> No.1216864
File: 104 KB, 673x971, circuit.jpg [View same] [iqdb] [saucenao] [google] [report]

Thank you. Although Im pretty stupid and would like if you'd clarify the circuit for me. From my understanding.
1. The source charges the capacitor
2. The charger capacitor discharges and activates the gate
3. Gate opens and the current flows through the FET instead of the capacitor

>> No.1216865

The capacitor doesn't discharge at all, no current can flow through the FET's gate, hence that resistor being pointless. Field Effect Transistors trigger on voltage, not current like a Bipolar Junction Transistor like the 2N2222. Once the capacitor has enough voltage across it to switch on the FET, the FET passes current from the +5V rail through the load resistor and to ground. The capacitor keeps charging, so you'll probably want to use an SPDT switch to short the cap through a resistor to ground when you turn it off, preferably with an R*C of less than 1, but not so that the resistor gets too hot. The other side of the switch will connect the voltage source to the circuit.

I really would recommend the comparator method, probably cheaper too.

>> No.1216870

I thought capacitors wont charge over the input voltage voltage?

>> No.1216872

Should I always place a 1 ohm resistor at the end of a circuit connecting to ground?
I see it a lot. Or maybe like a 1k ohm one at the end.
Say I have a part that takes 5v. It has pins but one is signal. Should I place a resistor between the ground pin and ground?

>> No.1216873

They won't, but the FET doesn't trigger at the full 5V or whatever source voltage you use, but rather at a semi-arbitrary value typically between 1V and 3V. The capacitor continues to charge up to 5V even after it has triggered the FET.

>> No.1216880
File: 38 KB, 696x442, circuit.jpg [View same] [iqdb] [saucenao] [google] [report]

How does the trigger voltage (Lets just say 2V) gets from the capacitor to the gate? Wouldn't the voltage be always 5v at the point depicted?

>> No.1216882

Initially, when completely discharged, the capacitor acts like a short circuit, and lets as much current as can possibly go through it, which in this case is equal to 5V/10MΩ. As this current flows into the capacitor, it collects this charge flowing into it instead of letting it pass through, and this separation of positive charge on the top plate and negative charge on the bottom forms a voltage. As the current keeps flowing, the charge separation grows greater causing the voltage across the capacitor to increase, and so the voltage across the resistor decreases for the total to still be 5V, and hence the current also decreases. See http://hyperphysics.phy-astr.gsu.edu/hbase/electric/capchg.html

The gate of the FET is the L shaped section on the left, so you know.

>> No.1216884

So in the beginning the voltage across the path I drew is 0V and will slowly rise along with the capacitor charge and the line I drew on the grawp is the trigger voltage for the FET gate?

>> No.1216887
File: 80 KB, 681x668, circuit.jpg [View same] [iqdb] [saucenao] [google] [report]

Sorry forgot picture.

>> No.1216888


>> No.1216890

Thank you so much. Really appreciate your time. This cleared a lot of things for me about circuitry in general!

>> No.1216891

The voltage across R1 and C1 as a pair will always equal the voltage across the voltage source, which is assumed to be stable. The voltage across C1 rises as it charges, and the voltage across R1 correspondingly reduces.

>> No.1216892

What did you draw the schematic with?

>> No.1216893
File: 5 KB, 640x400, timer circuit.png [View same] [iqdb] [saucenao] [google] [report]

Here's the timer circuit and a few rules to choosing values:
>trigger_voltage = V*R_3/(R_2 + R_3)
>maximum_delay(seconds) = P*C*ln((R_2 + R_3)/R_3)
>maximum_load_current = B_T*I_c
(B_T is the beta value of the transistor, I_c is the maximum current the comparator can source, don't worry about this if you use a FET instead, which I suggest you do if you plan on running something high-power)
>waste_current = V^2/(R_2 + R_3)
>discharge_time(seconds) = R_1*C*5
>maximum_discharge_power = V/R_1

And a few guidelines:
> V/L < maximum_load_current
> maximum_discharge_power < R_1_power_limit
> 0.2*V < trigger_voltage < 0.8*V
> waste_current < 1mA
> discharge_time(seconds) < 1s
> P >> comparator_input_impedance
> time(seconds) = 3600*time(hours)

Because of the soft limit on the discharging of the capacitor as set by the resistor's power limit, the capacitor has to have a fairly small value, so a supercap probably won't do, but something that requires a resistance value higher than a few hundred kΩ will interfere with the capacitor charging by virtue of the comparator's input impedance. So if you want the maximum delay to be around 5 hours, then you'll have to either use a power resistor R_1 significantly higher than 0.25W, or settle in for a discharge time significantly longer than 1s.

LTspice, it's a free circuit simulator by chip manufacturer/designer Linear Technologies. All you need to know to get started is to always put a ground reference somewhere, and start the simulation with the spice directive ".tran (start time) (total time recorded)", like the ".tran 0 3k" I've written to denote start at t=0s and simulate for 3000s. It supports all common SI prefixes but not case sensitive, so write "meg" instead of "M" because "M" is reserved for "m" as in "mA". Take some shitty GIMP circuit now.

>> No.1216894

Oh, and the beta of a transistor is specific to an individual transistor, and its range is typically specified on its datasheet (google them) like "100-200", so plan for the worst. When it comes to picking components I suggest you just go with whatever is popular and cheap, like an LM397 for the comparator, and a 2N2222 for the BJT. P is a variable resistor / rheostat / potentiometer.

>> No.1216912

>LM397 for the comparator
As is typical of comparators, "the LM397 features an open-collector output stage." You probably should add a pullup, or flip the transistor to PNP on the high side of the load and add a base resistor.

>> No.1216914
File: 85 KB, 770x526, timer-relay-4541.jpg [View same] [iqdb] [saucenao] [google] [report]

>using straight R-C to produce several hours long delays
>requiring a special comparator with internal hysteresis to avoid output chatter
>relying on comparator's output current limit to drive the BJT
>worrying about comparator's input current, but ignoring capacitor's leakage
>no limiting resistor in series with P
>math suggests no overdriving of BJT
>dragging logarithms everywhere, when you could simply choose R2/R3 to give t=P*C or some other convenient value

Since large resistors and capacitors seriously suck, it is more practical to use a higher frequency oscillator and divide the frequency down. Since this is a common problem, there are ICs like 4541. It even requires (slightly) less components. Pic related has all kinds of extra shit, like extra ranges, though.

>> No.1216916

Thanks again! My local electronics store doesnt have the same fet and comparator as your schematic. What would be the characteristics to look at to pick an alternative?

>> No.1216917

Well my local store does have the 4541. What are the minimum component requirements to your suggestion?

>> No.1216923

Valid crits, all, but you missed the problem statement...
>What is the most simple timer you can make to flick a electronic switch? Time period being hours
Also pin 8 is only briefly on if I read the datasheet correctly, and anon wants a latched action.

Alternate part numbers for transistor: 2N3904, 2N4124
Hours-long delays are hard, anon.

>> No.1216926

He said simplest, that's what I did. Digital doesn't sound simple, and I thought it would be good to give something understandable. But I agree that the circuit I gave is not ideal. I chose to vary P instead of R_2 and R_3 because changing R_2 and R_3 would not be linear, and you'd get very fast changes at the top of the waveform, likely voltages that couldn't be measured without a rail-to-rail comparator.

I don't actually know what an LM397 is, I just googled "popular comparator and saw it first, which might say something about the gap between my theory and my practice. But I don't think there'll be any chatter, will there? Anyway, after looking at leakage current values, yes they will completely destroy this design. DO NOT BUILD THE CHARGING CAPACITOR CIRCUIT. Actually I'll test it with an electrolytic I've got lying about.

What do you intend on soldering it to? I would recommend perfboard if you never want to look at it again, but it is some ugly garbage. Etching a board with a marker for DIPs is also a PITA, so good luck with that.

>> No.1216927
File: 39 KB, 987x818, 4541.png [View same] [iqdb] [saucenao] [google] [report]

Something like this. The nominal supply voltage range for the IC is 3...15V.
Choose the mosfet according to your output current/load requirements (RL). 2N7000 is ok for small (50mA or so) loads and when using it, you could even omit RG. For larger loads something like IRF540 could be suitable. In that case RG is nice to have, 1k is a suitable value. 2N7000 is ok with 5V supply voltage, but IRF540 requires more.
RS isn't strictly necessary (you can short it), but it makes the timer somewhat more accurate at short delays. D is required only if your load is inductive, like a relay or a motor.
This circuit uses 4541's power-on reset circuit, but you can add an external reset button if you want to. In that case you would pull pin 5 low with a 10k resistor and use a switch to connect it momentarily to V+.
C should be non-polarized, preferably a plastic capacitor.

>> No.1216930

The simplest option was presented many posts ago, using a mosfet, a resistor and a capacitor.
The output of the 4541 latches when pin 10 is low.

>> No.1216937

This looks like the winner design. I'm testing the a 470µF cap in series with my multimeter's 1MΩ voltage input impedance, and at the moment the parallel resistance is at least 2.3MΩ, but apparently it isn't fixed (the leakage current is?), so I'll have to look at a few different resistors in series/parallel with the multimeter. Now it's at least 3MΩ.

>> No.1216940

Does anybody know a good way of measuring voltage with effectively infinite input impedance? I know 1M is much worse than the common 10M, but a 10GΩ input impedance would be really useful for this sort of low current electronics. Hitting 16MΩ now.

>> No.1216955

130MΩ. Not sure if that means anything though.

>> No.1216961

You could use an opamp buffer. Even antiquated LF356 and CA3140 in particular have pretty high input impedance and if that's not enough, there are much better amplifiers.
And since you're measuring the leakage of a capacitor, you can use the capacitor itself to integrate the leakage current - assuming you trust the capacitance value. In that case you could charge the capacitor to, say, 1V and then wait an hour or more before measuring the remaining voltage. Assuming a normal 10M multimeter and that 470u capacitor, couple of seconds long measurement would not discharge the capacitor much.

>But I don't think there'll be any chatter, will there
Assuming 5h time constant, 10V supply voltage and a trip point around t=RC, the voltage across the capacitor is rising around 10mV per minute.
- Any ripple in the supply voltage is seen by the comparator and can change the output state.
- The suggested comparator has a typical gain of 120000. The full output swing would thus require around 40uV overdrive, resulting in a rise time of couple of hundred milliseconds.
- Even slightest accidental feedback from output to input can make comparators oscillate when they're close their linear region. They simply aren't designed to work with negative feedback.

>> No.1216967
File: 6 KB, 225x225, NOMA Indoor Grounded Timer.jpg [View same] [iqdb] [saucenao] [google] [report]

> What is the most simple timer you can make to flick a electronic switch? Time period being hours

$2 at the thrift store.

>> No.1216969

How do I limit voltage? I have a pot, which if a few things go horribly wrong can send a voltage higher than some of the other parts specs.
In case something happens and the pot sends out a voltage higher than the rest can take, how can I stop the lower the voltage then. I dont mind losing the pots output until its back down.

>> No.1216989
File: 23 KB, 360x540, 4541.jpg [View same] [iqdb] [saucenao] [google] [report]

>my local store does have the 4541
Go get one, no wait, get five or ten. I definitely will because this chip will disappear from the market. My local dealer has it for 0.40 each and 3.20 for ten. Same for the SMD SO14 version.

First thanks to >>1216914 for mentioning the 4541. I didn't know about it and it seems to be the most viable solution for long time delay circuits.

Suggestions to use a huge resistor and a huge capacitor usually rest on the 'threshold' myth. In reality transistors have no threshold, that's just a popular metaphor for what's really going on, i.e. the transfer characteristic. Bipolar transistors have an exponential transfer function and MODFETs have both an exponential (for small currents) and a quadratic one. The transition region between both is what is called the MOSFET 'threshold'. They do not suddenly switch on, especially not when the gate-to-source voltage rises slowly, which is the case for the RC non-solution so many people seem to favor. For a BJT no such region exists and the 'threshold' is totally arbitrary.

Now when you come across a new (for me) component, the Golden Rule is RTFD, read the fucking datasheet. I got mine from https://assets.nexperia.com/documents/data-sheet/HEF4541B.pdf

All you need to know is in there. Basic properties, functional diagram, pin assignment as well as valuable information e.g. about how to calculate the clock frequency/time and (surprise) that the power-on reset function is unreliable for supply voltages below 8.5V. There doesn't seem to be a 5V 74HC version of the 4541 (despite the many bait links google offers).

Diagrams have been posted enough but I'll still draw my own (including pcb), build and test it and file it in my repository of useful circuits.

Today is a good day.

>> No.1217031
File: 68 KB, 1246x828, 40.png [View same] [iqdb] [saucenao] [google] [report]

Not sure this is the correct thread to ask, but I have a problem with sigrok's pulseview.
I have a VCD file containing random bytes (8 wires with random values), when I try to import it in pulseview I get long periodic "idle periods" with increasing frequency like in pic, on the bottom GTKwave shows the correct signals.
I guess something wrong is happening in sigrok when reading the timestamps but going to the lines corresponding with the start of the idle periods I don't see anything out of place (and I don't expect to, considering that on the same file gtkwave is working correctly).
What could it be?

>> No.1217059

Normally a MM has an input divider chain that defines the input impedance. There are some models that have no internal termination at the lowest DC range. One example is the UT61E which has about 100MΩ but >3GΩ in its 220mV range, which is essentially the internal ESD protection circuit at low voltage.

I have a Metex autorange pen-type MM that has the same 'feature', normally 10MΩ but 'open' at the 200mV range. This is totally annoying, it beeps all the time because the surrounding electric field causes it to change range and polarity if not shorted or locked to the 2V range by pushing the range button.

Also, electrolytic capacitors have internal chemical processes that heavily depend on temperature, charge state and also the past charging history. They are not reliable for timing applications of any kind. The capacity tolerance is huge, +100% -20% is far from impossible.

Real electrometers do exist but they are either expensive lab equipment or .. diy equipment :)

>> No.1217080

a button fell off my car remote and i had to fix it, nothing major. i noticed there are quite few parts on the circuit board and it doesnt look very special at all. it's supposed to run at 433 MHz.

now i'm curious: what kind of transmitter is generally used in those applications? i assume its some LC oscillator

also since its running at 433 MHz, don't you need a certain design of the PCB do make it work at those frequency? it looks so generic

>> No.1217082
File: 21 KB, 169x258, remote.jpg [View same] [iqdb] [saucenao] [google] [report]

pic of the PCB

on the other side of the board are battery contacts and an IC+crystal. thats it

>> No.1217085
File: 810 KB, 1920x1080, Untitled 181.jpg [View same] [iqdb] [saucenao] [google] [report]

Thanks guys! Went with 4541. The slower 15s-10h tested with potentiometer at 15s and works like a charm!

>> No.1217090

Depends, but diode clamps are popular.

>> No.1217094
File: 7 KB, 486x256, transmitter.png [View same] [iqdb] [saucenao] [google] [report]

Look at the image at >>1215928 There's something called SAW on the component side. That's a surface acoustic wave resonator (Y1 in the pic). Pure LC is much too unstable and would be frowned upon by the FCC.

>> No.1217097

very interesting, thank you!

>> No.1217115

Congrats! You saved yourself a lot of disappointment. Btw: my dealer told me that the DIP14 version is end of life but the SMD variants are still available.

>> No.1217167
File: 3 KB, 119x100, 1500952289418.gif [View same] [iqdb] [saucenao] [google] [report]

Can you use a soldering guy to do some light welds? Might have to do some for some college art club

>> No.1217168


plastic welds only.

>> No.1217173

Cant wait to try this

>> No.1217221

at some point i learned that rechargeable batteries are charged in a certain way. for example constant current at first and then at some point switch to constant voltage

in case of a phone or laptop, is the circuit realizing this function in the external charger or in the phone/laptop itself?

>> No.1217231

Tthe charging process depends on accurate readings. Therefore, charge control happens within the device. The external charger is merely a power supply, of varying quality.

>> No.1217239

okay thanks, follow up question:

those charge adapters often have different output characteristics like lets say A="20V, 1A" or B="20V, 2A"

now when you have a laptop that comes with charger A, people (and the manual) say you shouldnt use a different charger than A. but if the charge control happens in the device, this shouldnt be a problem right? only if there is a huge difference in voltage like C="10V, 1A"

>> No.1217244

>laptop charger
No, they're laptop power supplies. In the case of laptops, the power connector is often treated by designers as a proprietary interface because why not. Designers aren't bound to any standards beyond broad safety codes and can design their charging and other power circuitry as they like.
One of the few good things the EU has done is standardizing mobile phone power inputs so that they don't have this problem.
>A="20V, 1A" or B="20V, 2A"
Assuming the pinouts match, B should be usable in place of A because, unless they rely on current limiting behavior of the external charger for some ungodly reason, only the voltage needs to match and any excess current capacity is of no consequence.

>> No.1217245

okay, thanks!

>> No.1217263

dell is so evil they added a data pin to their charging port that connects to a chip in the charger just to accept only dell chargers. I got in a unsolvable problem with an old dell laptop I found
>battery drained
>need charger and full battery to update bios
>need bios update to fuck with the pin thing and charge (the battery) with chinesium
>chinesium powers the laptop but does not charge the battery
>have no dell charger

>> No.1217271

Any got recommendations for cheap op-amps with rail-to-rail input and output capabilities? Preferably LT chips.

>> No.1217303

Well fuck

>> No.1217305

buying wire for a compressor motor. its 5 hp 14 amp at 240v and spec sheet says use a 30 amp breaker. its going to likely be 25 feet or so from the outlet. should i use 12 gauge or 10 gauge?

>> No.1217316

>manufacturer and price are more important than supply voltage, supply current, GBW, input offset current,...
Why not use the manufacturer's parametric selector to find something that fits your application? Also, not that anything's wrong with them, but why LT in particular?

>> No.1217348

10AWG would be better.

>> No.1217358

Lots of HP laptops use the same connector with centre pin, but instead of the 1-wire chip for identification it's just a series resistor. So pretty easy to convert a Dell charger into an HP one.

>> No.1217387

>i assume its some LC oscillator

>> No.1217391

Is their a alternative to the TOSHIBA TC40H245P?

>> No.1217403

What's it used in? If not particularly high speed Ti SN74AC245 might work. Or any other 40*245/74*245 chip, they seem to all use same pinout if in DIP package.

>> No.1217404

Its used as part of a keyboard matrix and buttons on a synthesizer.

>> No.1217405
File: 164 KB, 466x350, file.png [View same] [iqdb] [saucenao] [google] [report]

Here's some more info on it, IC 34 and 45 use the Mitsubishi M74LS138P

>> No.1217416

Sounds like any CMOS 74*245 variant ought to do it. Check the Vcc on the board and replacement IC first before selecting the logic family.

>> No.1217418

Looks like it should be suitable then, doesn't look like something that'd be particularly sensitive to timing differences.

>> No.1217476
File: 88 KB, 1051x781, 1479916279056.png [View same] [iqdb] [saucenao] [google] [report]

so i found myself in need of a wideband, frequency driven, but not too accurate sine generator. this circuit works from dc to 1mhz. as noted in the picture the multiplier's output is a differential current sink. i turned that into a single ended current source the dumbest way possible.

can anyone think of a more elegant solution to convert the differential output to single ended?

>> No.1217488
File: 4 KB, 420x367, 1478848533295.png [View same] [iqdb] [saucenao] [google] [report]

I haven't simulated this, but maybe just lose two of your current mirrors?

>> No.1217490
File: 46 KB, 1150x703, 1489791266287.jpg [View same] [iqdb] [saucenao] [google] [report]

yeah i stumbled on that by accident but now i'm suspicious because i don't get how the B side has a voltage source to draw current from.

>> No.1217496

From the output, of course. You may want to change the load resistor to a voltage divider between the supply rails.

>> No.1217510

What tools do you have access to?
A data sheet may be unnecessary, as You could try to make it count until it overflows.

>> No.1217548

It's a security chip. You're unlikely to find public data. It's probably implemented as a microcontroller running a secret handshake with the plotter, is standard printer industry practice holds here. Better just get the new chip and recondition the maintenance unit.
Also lose the name, newfriend.

>> No.1217549

was for

>> No.1217605

On second thought, there's a possibility it could be an I2C EEPROM of the 24LCxxx variety. That might be worth checking into. Then you'd just have to figure out what to write to it to reset the count.

>> No.1217619
File: 3.26 MB, 2336x1498, back_pcb.jpg [View same] [iqdb] [saucenao] [google] [report]

I've got access to some tools. Got an Arduino laying around that I can program. Soldering tools and parts.

Looking at the back of the PCB I see four pins. I assume these are for Vcc, GND, Data and the Latch. Now I only need to figure out which is what and then what to write to it

Was thinking of soldering some wires to the pcb so I can read some values when it's in use. I'll keep you updated

On a side note: ordered a new chip, so we can play around with this one

>> No.1217621

was for you

Assumptions for now:
Pin 1 Vcc
Pin 2 GND
Both have a larger copper area for better heat distribution
Pin 3 Data or latch
Pin 4 Data or latch
Finer lines for less intensive use

>> No.1217624

So you plug the power source straight to pins 7 and 14?

>> No.1217627

Provided its between 3V and 15V, yes. It's pretty common for DIP14s, in fact they're probably all like that.

>> No.1217633

>common for DIP14s
unless 74HC or HCT

>> No.1217640
File: 192 KB, 922x790, il.jpg [View same] [iqdb] [saucenao] [google] [report]

Couldnt get your version to work. Even tried this and still the test led didnt light. Faulty 4541 or am I just an idiot?

>> No.1217643
File: 2.48 MB, 2336x4160, IMG_20170728_133914.jpg [View same] [iqdb] [saucenao] [google] [report]

Long story short; I broke it!

The chip needs to be mounted onto the maintenance cartridge which gets inserted into the plotter where I cannot see what happens.
It probably got stuck at something and now some legs have been broken off.
My assumptions about Vcc and GND were just the other way around and it ran on 750mV.

That's all folks, the funeral will be held this afternoon at your nearest garbage bin!

>inb4 shitty soldering, I know

>> No.1217655

>unless 74HC or HCT
My 74HC04s are Vcc at 14 and GND at 7, are there others that aren't like this?

The LED won't be very bright if you're running it off 3V, and is there any current flowing through the chip?

>> No.1217661

Running at 9v and yes.

>> No.1217662

When you stripped a length of wire and soldered its strands end to end to make 2.15m of thin wire to use as a power resistor, but it's only 3.2Ω...

I guess I'll use a jar of salt water, but I hope I can get resistances from 2Ω to 60Ω.

>> No.1217674

Whats a good brand of hand tools?
i'm lookin for side cutters that will actually last a while and maybe some new pliers

>> No.1217676

There's 9v charge on the transistor gate leg right away once i plug in the battery? Why

>> No.1217678

Assuming you're using >>1216927 as your circuit, that would mean your load is constantly active. Try moving pin 9's voltage from GND to Vcc instead. Changing A and B might also have an effect. I'm not the guy who made that circuit, btw.

>> No.1217680
File: 81 KB, 1222x268, knipex.jpg [View same] [iqdb] [saucenao] [google] [report]

> side cutters that will actually last a while and maybe some new pliers

unless you live in germany, this is probably your best bet.

>> No.1217684

In addition to what >>1217678 said, the internal power-on reset circuit requires that the supply voltage starts from nearly zero and rises quite rapidly.
You can try resetting the chip manually by connecting pin 6 momentarily to supply voltage.

>> No.1217688

So I don't have to screw around adding some other guy's model to LTSpice and trying to remember which port of the rectangle is which op-amp pin.

Also because I'm an LT fanboy.

And yeah, I care more about price than any of those other specs. It's for a hobby project, not a million units/year consumer device.

>> No.1217693
File: 10 KB, 281x131, dip-sop.png [View same] [iqdb] [saucenao] [google] [report]

Sorry, bad quotation. I meant 3..15V is not common (HC 2..6V, HCT 5V±10%) while (14)=Vdd, (7)=Vss for DIP14 and (16)=Vdd, (8)=Vss for DIP16 is indeed the common pinout (the 'diagonal' pins), even for the SMD versions. NXP and ONsemi no longer offer the DIP14 version of the 4541.

>> No.1217705

>LT fanboy
Too bad it's Analog Devices now. And both Bob Widlar and Jim Williams are dead.
An LT fanboi should know that the parametric search the other anon mentioned tells you the price. Not for single pieces, but it's still a reasonable starting point.
> I care more about price than any of those other specs
Maybe you pick some other manufacturer.

>> No.1217706



>> No.1217715

Come on all you need to do is click "Select Column" and then "Price 1k*". 1k is a lot but will tell you which are generally chepare than the others at least.

>> No.1217719
File: 77 KB, 948x258, 4541-settings.png [View same] [iqdb] [saucenao] [google] [report]

You need at least a 9V supply for the auto-reset function to work. Earlier in this thread someone confused 470nF (=0.47µF) and 470µF. I'd suggest you try manual reset, output initially high, recycle mode and n=13. LED should blink or flicker. Calculate expected frequency and compare result.

This is not critical but I think RS is way too small, it should be at least 2*RTC. The reason is that the peak voltage at pin (3) exceeds Vdd and Vss and activates the protection circuitry which speeds up the clock generator.

>> No.1217724

Hi. Audio device question on class D amplifier.

Is PAM8403 Parallel Bridge Tied Load compatible?

No direct answer on datasheet.

>> No.1217726

what are you trying to say? the crystal could be an external clock for the µC

>> No.1217742

Will the beginner books teach me how to actually solder and not fuck up? I want to get over my fear and learn to do it but I just fuck up every time and I have some electronics to repair

>> No.1217757

How about a chopper drive like a stepper motor driver? Just an oscillator, S-R flip flop, comparator, inductor and a MOSFET.

>> No.1217758
File: 2 KB, 303x166, image.jpg [View same] [iqdb] [saucenao] [google] [report]

Soldering itself (especially the timing) is better demonstrated than described. The main problem is that you only have two hands, one for the iron and one for the solder wire. This means you have to somehow fix the components in place so they don't move while you solder. Take a deep breath then slowly exhale while you melt it. All else is practise. The easiest way to start is to get a kit with a printed circuit board. Ali is your friend in such matters. Repair work often envolves desoldering which a bit more demanding, skill-wise.

>> No.1217759

The two outputs of that IC already use BTL. You can't BTL it further. Dunno if you can parallel the outputs.

If it's for 433MHz, you can bet the crystal sets the frequency.

You could try Youtube soldering videos.
After that, practice with wires and junk electronics.

>> No.1217774
File: 20 KB, 405x259, saw.jpg [View same] [iqdb] [saucenao] [google] [report]

>crystal sets the frequency
There are no 433 MHz crystals, that's physically impossible. The standard is the SAW resonator. They come in different shapes and sizes. Remotes need to be dirt cheap.

>> No.1217775

Repair means de soldering which is messier.
Soldering is easy. Heat up one side. Bring in solder from the opposite side. Tickle the solder on the tip of the iron to get it started, drag the solder where you want it to go while holding the iron steady. Remove solder then iron.
Get a solder sucker for removal, heat up then suck off. Repeat add nausium until clean enough.

>> No.1217784

Please give me a source that elaborates on it. I have watched people do it so many times and they do it like its nothing and extremely fast but actually its quite difficult for someone who has no idea about it especially when I can permanently damage the components.

>> No.1217786

There are phase locks. If there's no 433MHz saw filter, you can expect that the output frequency is produced with a PLL, using the crystal as a reference. The actual oscillator the 433MHz PLL uses is normally integrated on the same chip.

>> No.1217788

"The best way to get involved in electronics is just to make stuff. Don't be afraid to get your hands dirty." -- OP

>> No.1217791

Correct, but not for cheap remotes that run on ISM frequencies.

>> No.1217816

i found an old analog volt gauge display which appears to be in working condition (placing a voltage across its leads yields an accurate displacement on the needle), and want to use it in a variable power supply project.

my question is, there's no form of max current rating marked anywhere on the device, and i assume at some point it could burn up if i didn't use any kind of resistor- is this assumption correct? and if so, whats a good current to have running through this?

>> No.1217822

If you have a multimeter you can find that out I think.

>> No.1217827

>analog volt gauge

analog volt meters like any voltmeter are high impedance. the current is negligible.

>> No.1217830


I should have pointed out that you can damage it if you use it improperly, like applying 100 volts when the full scale voltage is 1 volt. As long as the needle does not peg it will not draw much current.

>> No.1217850

>Too bad it's Analog Devices now. And both Bob Widlar and Jim Williams are dead.
I want to collect all my old booze bottles and set up a shrine to Widlar.

>> No.1217857
File: 1.76 MB, 2475x3500, 8e680c82d7cbabeded2b7d37fd2ea11c.jpg [View same] [iqdb] [saucenao] [google] [report]

Is there something like the 7400 series, but can switch at 10MHz and handle an input of up to 10v?

>> No.1217864

>I want to collect all my old booze bottles and set up a shrine to Widlar.

If I could live my life as anyone it would probably be Widlar first and then Carmack second, except Carmack would have died painlessly after the release of Quake 2 or maybe even DOOM 2 so that Romero was still more or less cool.

On the opposite end of the spectrum, if i were forced to live my life again as Bill Gates I would die no matter how painful it could be to avoid raping the world with dos and windows garbage all the while becoming the richest man in the world and thinking you were anything but a ruthless evil business shit.

The funny thing about Microsoft is that from day one to this very day they never know that something is a hot steaming turd until they do it; no amount of in-house testing or evaluation ever works. I suppose if the guys at the top are like Bill and Steve you tend to be stocked with so many idiots and assholes that the few competent people just sigh and go cry in their cubicle.

>> No.1217870
File: 72 KB, 602x664, HEF4013B.png [View same] [iqdb] [saucenao] [google] [report]

Yes, the CMOS HEF4000 series. Pic shows HEF4013B Dual D-type flip-flop data.

>> No.1217874
File: 106 KB, 880x900, AA-to-5V.jpg [View same] [iqdb] [saucenao] [google] [report]

>the current is negligible
But not compared to modern digital voltmeters. I have such a historic instrument. Image shows the sensitivity.

>> No.1217880

Oh, nice. Thank you.

>> No.1217884

>But not compared to modern digital voltmeters. I have such a historic instrument. Image shows the sensitivity.

so you're telling that guy that yes indeed his analog voltmeter will draw excessive amperage or are you just being an autistic twit?

voltmeters properly used, whether manufactured in 1940 or 2017 do not draw a lot of current when used properly. right?

>> No.1217885

I have an audio signal coming into my circuit and I need to reference my rails to it so that the audio ground is right in the middle of my single-sided supply. If I run V+ through a voltage divider through V-, buffer it, and connect the output to audio ground, will that relocate the supply voltage around it?

Analog ground current will be low enough that the op-amp can sink or source it without issue.

>> No.1217892
File: 6 KB, 471x553, 1479524716162.png [View same] [iqdb] [saucenao] [google] [report]

I tend to think of it in terms of shifting the input ground level rather than the rails, but sure. Take care that the input Vvgnd is floating with respect to the power supply of your circuit. I've toasted equipment when it wasn't. If you can't guarantee that, add a cap on the ring of the jack too.

>> No.1217897
File: 72 KB, 533x800, a525teen06.jpg [View same] [iqdb] [saucenao] [google] [report]


I love questions like this.

>> No.1217899

Thanks. I'd think in terms of shifting AGND but I don't know how well the signal source (tablet) would play along. I'm also considering running them off the same 5V USB supply which would further complicate things, but I'll probably run this circuit off a different 5V adapter for safety.

What are C1 and R6 doing in your schematic?

>> No.1217900

1 - get some practice boards from the junk.
2 - use more flux.

>> No.1217903
File: 25 KB, 1492x599, file.png [View same] [iqdb] [saucenao] [google] [report]

Circuit looks like this, by the way. I haven't figured out how to change the symbol for custom models in LTSpice so those rectangles are op-amps with 1 = Vin+, 2= Vin-, 40 = Vout

>> No.1217921
File: 7 KB, 590x492, 1481946950874.png [View same] [iqdb] [saucenao] [google] [report]

>I'm also considering running them off the same 5V USB supply which would further complicate things
Redrawing and reannotating.
You should definitely add that cap C3 on the input ground or you could be sad.
The tablet probably assumes the output has no other dc paths to any of its own components. C1-C3 block any dc path between it and your common USB supply. These should be non-polarized, decent tolerance and fairly large, 1uF 20% is probably good. R2 and R3 keep the inner side of C1 and C2 near the virtual ground potential. Values in the 1k-50k range are probably good, adjust for desired frequency response. If your tablet expects headphones rather than line levels on the output, consider low-value resistors (~64 ohms, not shown) on the jack between pins 1 and 3, and pins 2 and 3, so that the amplifier in the tablet has the dc path it expects.

>> No.1217934

Ah, gotcha. The tablet is typically used to drive an audio amp with a high input impedance so I should be ok without the resistors between audio lines. I'm buying a bunch of 1uF caps anyway so I'll use them for C1, 2, and 3. R2 and R3 12k for the same reason and to not filter out any bass (13Hz Fc).

Thanks a lot for the help.

>> No.1217942

You might want to parallel two of those as C3 for symmetry.
No prob m8. Hope your project works out well.

>> No.1218007

>I'm an LT fanboy
I'm a TI guy myself, but I'm pretty ignorant to what their founding fathers are famous for or what innovations got them established in the business. Anyone care to enlighten me?

>> No.1218017

How much power are you talking about?

>> No.1218029

10W, I've got it all calculated. I'm trying to work out at what power/current draw my USB power bank will automatically turn on, and at what, higher power draw will it stay on. It might have something to do with the data lines, so if there's a good resource about USB data lines and their use in charging, I'm all ears.

>> No.1218036

many thanks. i looked into analog voltmeter design and gleaned similar info. neat stuff.

>> No.1218040

I'm wondering if you could create an analogue voltmeter that doesn't pull current through a solenoid to measure the deflection of a magnet's magnetic field, but rather something that puts voltage across two plates and measures the deflection of an electret's electric field?

>> No.1218046

>I'm a TI guy myself, but I'm pretty ignorant to what their founding fathers are famous for or what innovations got them established in the business. Anyone care to enlighten me?
I believe it was calculators, m8.

D+ to D- with a 200 ohm resistor.

>> No.1218048

Resistors are sometimes needed on the gate of a FET. FETs have a gate capacitance. If you give it a lot of current by switching from low to high voltage with little path resistance, you can blow it up.

It's not needed in this case because of R1.



>> No.1218050

Sure, physics is sound. But it may take some to charge.

>> No.1218053

>ignorant to what their founding fathers are famous for or what innovations got them established in the business.

>Texas Instruments emerged in 1951 after a reorganization of Geophysical Service Incorporated, a company founded in 1930 that manufactured equipment for use in the seismic industry, as well as defense electronics. TI produced the world's first commercial silicon transistor in 1950, and designed and manufactured the first transistor radio in 1954. Jack Kilby invented the integrated circuit in 1958 while working at TI's Central Research Labs. TI also invented the hand-held calculator in 1967, and introduced the first single-chip microcontroller (MCU) in 1970, which combined all the elements of computing onto one piece of silicon.

t. wikipedia

>> No.1218060

Anyone know where I can find an audio spectrum analyzer program that lets you change the number of bands and levels?

>> No.1218068

>D+ to D- with a 200 ohm resistor.
I'll be testing a range of power outputs, and would like to be able to know exactly what resistors where cause what result. On inspection, both the data pins on the 2A and 1A socket on my USB power bank are shorted together and not connected to the µC at all, so should I assume that they're only used for the device being charged? Or is this just a particularly shitty power bank?

Are you recommending that I put up to 10W of power through a standard resistor? Or that I go out of my way to buy a selection of power resistors? I already have a bunch of 1/4W resistors that I plan on using for the values over 50Ω, but for the ones smaller than that I need something that can handle the current, and overseas shipping costs too much where I live, so I'm going with the water jar for now.

>> No.1218071

Sorry, should have been a link to power resistors. Didn't know you needed it adjustable, too.

>> No.1218075

My mistake. That was probably on the host end, not the device end.
How many 100 ohm resistors do you have? Those would be good for about 50mA each (a small fan would be advised). You would add omre in parallel until you get juice.

>> No.1218089

I think I've got 10 of each flavour from 10R to 1M, all 1/4W. I'd be able to get 10Ω if I could find them all. The salt water didn't work for shit, and now I've got saline all over my desk. I'm cutting open 4H and 6H pencils now, I hope this works.

>> No.1218096

USB 1.0's maximum current limit is 500mA. Ten 100-ohm resistors should be plenty.

>> No.1218110

Been here for a couple of years now, and I just have to say this is the shittiest ohm thread pic yet. I shudder a little every time I scroll by. So sorry. Carry on.

>> No.1218112

>USB 1.0
What decade is this?
I was able to pull 2.7A from the 2A port, so I guess the power bank's over-current shutoff device (which does kick in if you accidentally short + and -) must be higher than 2A. I've got readings from 0.17A to 1.85A, and they all activated when plugged in and remained active, which is good. But the 2-wire USB power plug I'm using is different to my Apple USB lead because it doesn't activate the power bank by plugging it in with no load, which the apple lead does. I'm guessing it's resistors from ground to the data leads, but I'm not having luck measuring them.

>> No.1218128

Ok its somewhere between 60mA and 90mA that above which the power bank remains activated, and below which the power bank activates for 5-9s and turns off again when the load is attached. It's important that I land my load in the first area, because in the second the switched-mode power supply turns off after those 5-9s and the USB puts 3.7V out instead. If this causes my LEDs and EL wire to be too dim to see, then I could be draining the battery down past danger level without even realising it.

Also I think I blew my shitty multimeter's fuse, so I'll solder a short length of thin (36AWG or so) wire in as a replacement.

>> No.1218135


Consider it a worst case. I didn't catch that you wanted to read overcurrent protection point too, sorry.
You may have to reverse-engineer the charger to see just what it's expecting. Try 1.5k resistors from D+ or D- to VBUS or GND. Also see if you can make any sense of the USB battery charging spec at http://composter.c om.ua/documents/BC1.2_FINAL.pdf

>> No.1218146

>LT fanboy
>care about price

Pick one.

>> No.1218149

>LT fanboy
>care about price
I get the impression our LT fanboy anon really cares about SPICE.

>> No.1218153

>you wanted to read overcurrent protection point
I was actually getting that by accident, I don't want to fry the thing. That document makes some degree of sense, but it really doesn't explain what different resistances where do. But there's nothing like trial and error.

>> No.1218156

The only sense I could make of it was that a portable device should figure out what it's connected to and how hard to pull before it just starts sucking current.
I'd be surprised if the power bank didn't have its own overdischarge cutoff. Maybe you should test that next with your new dummy load?

>> No.1218161

I don't want to push it, and I know I shouldn't be drawing much more than 1A anyways, so as long as my lower current limit is above 90mA, I shouldn't have to worry.

There's a strange memory effect within the power bank where if a load less than 60mA is plugged in, it activates the boost converter and sinks 5V for 5-9s, after it turns off I can't unplug it and plug it back in again to get it to activate for another 5-9s. Rather I have to push the "activate" button when it's plugged in, or unplug it, push the button, wait for it to deactivate 5-9s later, and plug it in again, where it activates for another 5-9s. Plugging in a 5MΩ load did not activate it.

I plan on wiring the lighting load directly up to the USB-output on the inside, and using a toggle switch to disconnect the load from the power supply when its not in use. This looks like it will work because the power bank will remain activated when drawing more than 90mA, but if my load was less, which it would be if I was only using 4 LEDs, I'd have to push the button every 5-9s.

The power bank has an anaemic, single LED flashlight function, caused by double-clicking the button, which I'd hoped to power the LEDs and EL wire from, but since its a constant-current supply rated at under 2mA, I wouldn't be able to run anything off it. The advantage of running lights off that instead of the USB is that the power bank's charging LEDs flash when activated, but not when the flashlight is in use. Since I'd have to run the 2mA into a BJT (or a voltage divider) and power the lights off the USB regardless, I'd rather go with the toggle switch method (because toggle switches are cool). I'll probably wire the flashlight to point at my gears so I can give them an inspection when I want to.

My last task when it comes to this power bank is figuring out whether the charging indicators are constant voltage or constant current, because they're too fucking bright. Hopefully they're CV.

>> No.1218177

Electronic based ignition system capable of lighting up gasoline? Looking for something simple

>> No.1218178

That's called an electroscope.

Crude ones require kilovolts for reasonable indication, but it is possible to make electroscopes for, say, 50V full deflection.

>> No.1218181

More relevant links:

>> No.1218190

I found this datasheet for an integrated power bank controller IC that may seem familiar in certain aspects: http://www.richtek.com/assets/product_file/RT9480/DS9480-05.pdf
On this particular chip, the LEDs are internally current-regulated. I don't think the chip will mind at all if they're unhooked or externally enabled by a toggle switch or pushbutton on the common leads.
The 3.7V you saw on VBUS when inactive was probably just a sense voltage. You can test that by drawing the battery bank down to shutoff point over several hours and monitoring Vbat to ensure that it does so at a sane voltage.

Cars do that famously well.

>> No.1218192
File: 50 KB, 600x375, Lg-Hantek-DSO-3062AL-Setup.jpg [View same] [iqdb] [saucenao] [google] [report]

Hey /ohm/, I'm at the point where an oscilloscope is now a necessity. I'd really like a digital bench top one, but I'm really interested in those USB o-scopes like pic related. Would latency be an issue with the USB o-scope? I mostly repair receivers and amplifiers but I also want to design and build synthesizers and other electronic instruments. The USB o-scope would be a huge plus for portability and working on things in the field, and I travel a lot for work and it's nice being able to work on hobbies while on the road.

>> No.1218200

its good as a beginner scope. you wont truly appreciate the beauty of rigol scope if you didnt start with usb scopes

>> No.1218201

Will I still appreciate the beauty of the Rigol if I had one of those chink dual 20MHz CRT scopes from $back_in_the_day?

>> No.1218204
File: 4 KB, 640x400, eugh.png [View same] [iqdb] [saucenao] [google] [report]

I don't think I can pull much current out of the 3.7V USB anyways, I'm getting 350µA through a 100Ω resistor. I guess this means I don't have to worry about it draining my battery down past safe levels if it's inactive while my load is still turned on. But if that happens I might have to worry about my filter capacitors feeding back into the battery bank. Hopefully that will only happen in the event of a microcontroller bug.

The PCB's control chip doesn't have any writing on it at all, so that link helps, but its not the one I've got, since it only supports 1.2A charging and doesn't have a solar charging feature. How did you find that particular IC?

Pic related is how the buttons and LEDs connect to the IC. I can only assume that the two pins are multiplexed get the right output signals because all LEDs appear to be able to blink independently. I don't like it. I will desolder one now to see what sort of power supply it has.

>> No.1218206

>Rigol after USB thingy
Oooh, awesome! Why did I waste my time with that fucking obnoxious toy?
>Rigol after 20MHz CRT scope
Um, okay. Technology marches on etc. etc.

>> No.1218208

>buy superior Nippon Goot soldering iron
>lasts 2 months before the element goes open-circuit
I guess I won't be desoldering those LEDs or shorting my broken fuse after all. What the fuck went wrong?

>> No.1218209

for audio stuff I'd rather have a 20MHz CRT over a cheap USB digital.

>> No.1218210

>filter caps
Just don't have very large ones. The USB spec prescribes a certain maximum, something like 10uF iirc.
>How did you find that particular IC?
Web search for "usb power bank controller".
>pic related
Ew. Masking tape it is.
>soldering iron deth note
So how do we like Youyue soldering gear in here?

>Um, okay. Technology marches on etc. etc.
20MHz guy here. I've never had a DSO before. That'll be new.

Absolutely. Lag is a kicker.

>> No.1218211

Yea getting minimum of 10kV to spark a spark plug from a 9v batter?

>> No.1218216

>search in google
>find no official website
>only AliExpress and eBay links

I was planning on having somewhat large caps to stop my flyback transformer's oscillating current draw from fucking with the potentially similar frequency boost converter of the power bank itself. As long as the caps are on the other side of the switch I should be fine, and I thought if I put a supercap there I could get a cool fade-out when I flip the switch. But that's just an excuse to waste money, which I really don't need now that I've blown a soldering iron. Or maybe a 28USD 10W temperature controlled soldering station.

What are soldering iron tips commonly made of anyway? Is it copper or brass?

>> No.1218217
File: 73 KB, 800x480, mmm.png [View same] [iqdb] [saucenao] [google] [report]


When I was looking for a new scope I read that you can damage your laptop with a usb scope if you are not careful. Maybe you have to do something incredibly stupid, but that warning was enough for me.

My $400 Rigol is wonderful. DS1054Z has four channels, which I used to think was overkill, but quite often I use all four. My old scope is a 20 Mhz analog and that's way better than nothing, but being able to capture traces before and after the trigger is awesome when you have a circuit where all hell breaks loose when a relay closes.

>> No.1218219

>potentially similar frequency boost converter
The boost converter will probably be running in the few hundreds of kilohertz if not higher. To the power bank, it's just a fluctuating load and it will just follow along.
>As long as the caps are on the other side of the switch I should be fine
The cap size spec is to limit inrush current. Confirmed 10uF.
>What are soldering iron tips made of
iirc, copper core, iron cladding, maybe tin plating.

Is the 1054Z still """"user-upgradeable"""" to unlock options and 100MHz operation?

>> No.1218220

>Is the 1054Z still """"user-upgradeable"""" to unlock options and 100MHz operation?

I haven't tried on mine. I'm happy with it as it is, and don't want to have to restore it if it no longer works.

The upgrade is awesome though. I can understand bundling software and making you pay to unlock extras because it costs them nothing to put the code in, but I cannot understand the logic behind installing the ram and not letting me use it. They paid for it and now I have it. What on earth makes sense there?

>> No.1218223

IIRC it's an "artificial" way to have different product of different price range on the market, which allow them to make more money.
Making a low end version of a product by just locking some features is sometimes cheaper and have better returns than actually designing a new version. Of course they can't just lower the price of the "unlocked" full version because clients who want the "pro" features are generally happy to pay the higher price.

>> No.1218227

>maybe tin plating
I think it has to be something like tin on the outside because it wets fairly well to solder, but tin itself melts at too low a temperature, and the outside isn't copper coloured.

After looking a bit, it seems the tip is almost always iron cladding over a copper core, because it apparently wets the solder, but isn't dissolved into it. I feel like there's a better solution somewhere. Is it lasers? I think it's lasers.

>> No.1218228

>by just locking some features is sometimes cheaper

How can it be cheaper to disable some of the ram? I understand the rest of your points, and as I said above, locking software or firmware makes sense because they don't lose any money. But they bought the ram and gave it to me.

It would be like putting a 5 speed transmission in your car but using software to lock out 5th unless you paid for the upgrade. If they can manufacture and ship 5th gear for X dollars why not let me use it for X dollars? Won't you possibly sell more cars? I'm stuck on the hardware vs. software aspect of it.

I absolutely understand putting multiple games on one DVD and you having to pay for each game.

>> No.1218265

> How can it be cheaper to disable some of the ram?
It may be cheaper to make a single model and selectively disable some of the RAM than to make 2 physically-distinct models.

> If they can manufacture and ship 5th gear for X dollars why not let me use it for X dollars?
Because then you won't pay X+Y dollars for the more expensive version.

Cost isn't the only factor in determining the selling price, or even the primary factor. The primary factor is what the customer is willing to pay for it.

>> No.1218269

It's not about how much it costs them, it's more about how much they can make from a customer. If you can bulk buy x amount of ram, and charge y dollars for 1/2 of it, and 2y dollars for it all you're squeezing more money out of the people prepared to spend 2y on it.

There's a train of thought that Rigol were deliberately pretty lax and intended for hobbyists to hack this model. They would tend to have lower budgets, but seeing a 'hack' to get full features would mean everyone would rush to them, buy their scopes and then 'upgrade'. Rigol weren't losing money because those people were mostly never going to spend the extra money on the upgraded features. However, companies, labs etc will 99% of the time, just spend the extra $$$ on the full version so they dont waste time hacking their stuff, voiding warranties etc so for the same bit of hardware Rigol has multiple streams of revenue and free advertising.

>> No.1218270


This actually makes sense and is pretty clever on their part if it's what they did.

>> No.1218275

>Electronic based ignition system capable of lighting up gasoline? Looking for something simple

all the components are available at hobby stores, in the gas-powered RC planes section.

>> No.1218360
File: 4 KB, 391x241, vol_reg.png [View same] [iqdb] [saucenao] [google] [report]

I want to have a simple 5v regulator for my circuit. Looking at various examples, I see different capacitor values used. What determines the capacitance values?

>> No.1218370

>What determines the capacitance values?
Mostly superstition. If the input is clean DC and the regulator is close to the source you need no capacitor at all. Output C depends on type of load (pulse or continous?). Data sheet for the 78xx series recommends 100nF for input and output if the regulator is _not_ close to source or load. If your input comes from rectified AC it's the acceptable ripple voltage.

>> No.1218379

>I see different capacitor values used.

if you replaced the cap with a socket allowing you to change the value easily, you'd likely find that any value between 1pF and 1F would work equally well.

>> No.1218383

>Mostly superstition. If the input is clean DC

well then it's not superstition then, is it?

the input cap cleans up the input so that the device can do its job. the output cap is to keep the load voltage constant so that the load is happy. Nothing in electronics is superstition or has ever been, except when you are incompetent.

>> No.1218399

Calculate the required capacitance.

>> No.1218401


is this a trick question?

based on what? If you know how bad the input is you can specify the input cap. if you know the nature of the load you can specify the output cap.

otherwise do bench tests with realistic inputs and outputs. find capacitor values that work. but never rely upon superstitions.

>> No.1218408

Not all three terminal regulators are alike. It turns out that 7805's actually do not need input or output caps for stability. LDO regulators are fussy about their output caps, and for some the ESR of the output capacitance, including all those bypass caps sprinkled around a logic board, is more important than the capacitance value and must be chosen within a specific range for the regulator loop to be stable.
Also, any load that has current demand spikes too fast for the regulator to respond, such as high speed CMOS logic, will require capacitance to supply power at those instants. The same for loads with large step changes in current, where the output cap needs to supply power until the regulator can respond to the new load current by passing more from the input and the input cap.

>> No.1218409

So i've seen that it is possible to use some microcontrollers to make programmable custom keyboards (hhkb mods and those using teensy boards), i have this cortex m uC laying around and and old keyboard only with it's matrix plastic sheets, (i managed to fuck up the controller circuit, an old zilog uC) now i wouldn't be that hard to program the cortex to use the matrix as input and implement a ps/2 protocol so i can use it on my laptop with a ps/2 -> usb adapter i already have, right ? I don't know a lot about uCs so i think this could be a good oportunity to do a project that actually interests me, just want a second opinion confirming that this is indeed possible.
tl;dr using a uC to make a ps/2 keyboard, yay or nay ?

>> No.1218419

There are many microcontrollers that can just do USB, I know popular stm32f103c8t6 can do it and atmega32u4 (found on arduino micro) can too. PS/2 is doable too.

>> No.1218421

Sure, why not, that's exactly what they did.

Also true. HID is kind of convoluted, though, and doesn't help with what anon has on hand.

>> No.1218430
File: 95 KB, 800x361, 800px-STM32_Blue_Pill_top.jpg [View same] [iqdb] [saucenao] [google] [report]

> stm32f103c8t6
that is the one i currently have
>HID is kind of convoluted
i guess i'll try a ps/2 implementation first and then USB if i manage to understand how to use it in this uC, thanks guys.

>> No.1218432

Oh, you got one that speaks USB. Maybe that does help solve anon's problem. You might be able to find some sample code in app notes to work from, and iirc STM32s with USB have some nice USB support in ROM.

>> No.1218433

Trying for worst picture again, new bread

>> No.1218550

But we're not even at bump limit, are we?

>> No.1218567

>new bread
did you mean 'breed'?

>> No.1218619

ebin gabatsitor :DDD

I found it like that

no, bread as in thread

Beware you may have to change out a resistor to make native USB work.
>Probably won't work as a USB device if it has a 10k or 4k7 pull-up on PA12. You have to check the pull up on PA12 (USB D+). If it has a 10k or 4k7 pull-up resistor, you will need to replace it with a 1k5 one to use the native USB.

>> No.1218646

After having used zeners

Zeners are ass. Their voltage regulation is only accurate at like 20mA+, and their voltage varies dramatically (talking like 4.8-5.3 on a 5V zener)

>> No.1218651

Many times soft-locking (and in rare cases) hard locking stuff can be required because it doesn't function properly at "maximum" capacity. Very common in GPUs because they may have like a 5% yield to their target - but a 50% yield for 1/2 the way. CPUs that are multicore are often the same way, and along the same veins, nearly all CPUs that have integrated GPU lines (Intel chips mostly) have the internal GPU locked off for the higher chips. Locking the GPU prevents additional heat being generated, which can permit higher clocks. Going further, if a CPU was out of spec thermally with both the CPU and GPU at full throttle, the GPU could be disabled and permit selling a gpu-less chip at the clock frequency of the combination - not an issue for most complete PC makers as they'll be willing to spend an additional $10-20 if it means they are bundling a higher end GPU, or buy chips for lower cost and greater available quantity.

>> No.1219415

Can someone tell me why I suddenly suck at soldering. Havent soldering in ages but right now I cannot get the solder to melt. I can desolder things but I cannot get the solder I have to melt. The rosin starts to bubble of evaporate out of the solder while the solder doesnt melt.
This hasnt happened to me since I first started soldering and I blamed it on solder from china then.

>> No.1219421

Maybe you have lead-free solder? And yes, cheap solder from China can be the reason.

>> No.1219427

I went and bought some solder from radio shack a long time ago and have been using that. This solder has been working for me since. Unless storing conditions could effect it I dont understand what Im doing wrong.

>> No.1219428

Have you tried additional flux, maybe the rosin in the solder went bad?

>> No.1219745


>> No.1219778



>> No.1219868


now >>1218431

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