/diy/ - Do It Yourself

this thread's digits brought to you in part by the LTC6405 damn fast diff amplifier/driver

>>1576359
>This series is designed to suppress spurious output pulses due to noise or disturbance signals
>The devices can distinguish data signals from noise due to differences in frequency, burst length, and envelope duty cycle
>burst length
Hell, I guess I'm going to be circuit-bending an existing remote then.

>>1576547
Well if I use a long enough delay on the schmitt trigger I won't have to worry about altering the remote either, as I'll compress all the pulses into a single long one for the CLK pulse to the toggle circuit.

I think I understand circuit loading a bit better now

why though, is it that having a signal go to a load with a lower impedance than the sources causes attenuation? Does it form a voltage divider of sorts?

>>1576559
Yes, it does form a voltage divider. Though the more pressing matter is usually that the low output-impedance output is having a significant amount of heat produced within.

>>1576406
nice

>>1576555
>>1576293
Well this is a bit of an odd circuit, but this current-limited totem pole seems to work to charge and discharge the capacitor with different RC constants and without acting like a voltage divider.

>>1576627
Ignore the backwards PNP.
Pointing a cheapy CIR remote at the new circuit, it triggers consistently and doesn't get multiple triggers per button press. I might extend that post-trigger resistor to a 470kΩ to 1MΩ instead of a 220kΩ, giving me a higher dead-time against bouncing, but I don't think it's necessary. Quiescent current is sub-1mA, so it basically fulfills all my design requirements. Not sure if actually encoding data in the IR pulse helped improve the signal:noise ratio, but it certainly simplified the project.

I have a triac with a max Vgt of 1.2v. That seems a bit low, I'm trying to understand triac and scr operation.
The gate and A1 voltage is kinda like a bjt vbe (usually the max vbe is like 5v)? In the sense that it behaves like a diode, and actually going above 1.2v means short circuiting the thing across supplies?

>>1576805
> In the sense that it behaves like a diode, and actually going above 1.2v means short circuiting the thing across supplies?

sentence makes no sense.

>(usually the max vbe is like 5v)?

no, 5V is the Veb, the max reverse voltage you can apply before it explodes and digs into your skull.

as for the Vgt, you supply a current thru a resistor, and the device determines the Vgt based on that current and its own characteristics. just like with an NPN.

>>1576805
> In the sense that it behaves like a diode, and actually going above 1.2v means short circuiting the thing across supplies?

sentence makes no sense.

>(usually the max vbe is like 5v)?

no, 5V is the Veb, the max reverse voltage you can apply before it explodes and digs into your skull.

as for the Vgt, you supply a current thru a resistor, and the device determines the Vgt based on that current and its own characteristics. just like with an NPN. to turn it on, you need to supply enough current to get a Vgt above 1.2V, or below -1.2V.

>>1576831
Which sounds pretty normal since it will be 2*0.6, or the "forward voltage" of the B-E junction.

Why are tac welders so expensive?
Isn't it literally just a bunch of caps in parallel which you charge and then press a button to release all their charge through the weld?
seems pretty easy to make

Can MOSFETs explode for shortcircuiting drain and source? I want to explore some source of failure in my circuit and not risk blowing up more MOSFETs.

>>1576859
big caps are still expensive
so are switches that won't weld shut when used, especially timer switches
so is the amount of material needed to convey all that current to the workpieces without losing most of it
but yeah timers don't need to be as expensive as they are

>>1576911
no but they might not like it very much if you have an inductive load that the short built up too much energy in, or if you happened to short drain to gate

>>1576916
>big caps are still expensive
Not quite, I could afford myself a decent sized supercap bank for less than 20 dollars.

>>1576911
Short circuiting drain and source will bypass the mosfet. It should be fine.

Chinese multimeter for up to 400V and that is safe, any recs?

Why do consumer electronics like computers have a direct path from the circuit ground to the mains earth?

Holy shit a thread on here that is actually useful and I'm versed in am too high functioning for /b/

>>1576962
It has to do with the way the AC is converted to DC in order to get the transformer to have a complete circuit the DC ground must be attached to the AC ground in order to complete the circuit

>>1576957
Get a fluke or a rated desktop meter

>>1576971
>It has to do with the way the AC is converted to DC in order to get the transformer to have a complete circuit the DC ground must be attached to the AC ground in order to complete the circuit


Are you sure? They still work without the third pin

>>1576957
Uni-T

When you're using an opamp to control a BJT and don't want the BJT to be triggered when the opamp has no input... where should you put a pulldown? In position A or B?

>>1576973
Even in the "floating ground" style of plug the ground is connected through the hot or nuetral side and the principal of connection is the same where as I previously discussed it's connected the same way

>>1576970
>>1576971
please don't be samefag please don't be samefag

>>1576962
keeps static and RFI down. it wouldn't do to have billions of exquisitely thin MOSFET gates floating at several tens or hundreds of volts above earth especially when some kid sticks his tongue into a USB port

>>1576976
the opamp will always drive strongly. it wouldn't do anything on B, put it on A

>>1576979
Holy Fucking shit how did you know

>>1576981

>>1576996
rip bob .o7

>>1576996
where is the op pic I did of him throwing a computer? I've lost it.

>>1577009
I ate it

>>1577019
The computer was completely made from electronics components he loved so much.
The irony of this eventually killed him.

>>1577004

>>1577031

>>1577049
kek
yep, was 90% sure it was sudoku

>>1577019
Sorry to hear that. So I've done another one.

>>1576859
>>1576916
I'd argue that the copper bus bars are the expensive part.

While manhandling this part trying to get a stripped screw out, I managed to knock off and lose this part.. Does anyone know what surface mounted part this is so I can attempt to replace it myself?

>>1577088
A capacitor

>>1577088
gj
>C5
capacitors for dc blocking on the SATA data lanes, guessing 0.1µF
most likely they just copied JMicron's reference design and put it on their own board, if you want you could look up the datasheet for that chip number and see what they suggested

Absolute brainlet coming through. Why would the pmos be in cut off position when vi = +2.5 V? I keep seeing everyone on the internet saying Vg > Vt therefore the pmos is off. Why are they ignoring the source voltage? Isn't the equation actually VGS > Vt is cut off? Am I just not seeing why Vs is 0V?

>>1577140
Suppose QN is a closed switch because vi is 2.5V.
What is the VS voltage, then, for QP? It'll be 2.5V because QN is a closed, conducting switch.
VG for QP is of course 2.5V
VGS then for QP will be zero volts. Thus QP will be an open, nonconducting switch.

>>1577155
So maybe here is where the brainlet part comes out. So I've been doing assumption likes that and it doesn't feel particularly great to me.

Like I can assume QN is on but why wouldn't I want to worry about Vgs again but for the QN transistor?

Why can't I assume QP is on instead and therefore the source voltage would be Vo? Where Vo is just idp *10Kohm*-1?

Do I assume that QN will be on instead because the numbers are easier, where I can directly calculate idn and find Vo?

>>1577155
>>1577161
Nevermind I am just retarded I guess. I wouldn't have to worry then because VD= VS in that case.

Idk, I don't like doing all these "assume x" or "assume saturation region" and then having to check afterwards to see if I made the correct assumption. I feel like there should be a more intuitive understanding out there?

>>1577161
>>1577165
The circuit doesn't work exactly as I described, the voltage at VS is not actually zero. The assumption I gave you was to help your understanding.
Simulate this in falstad and you'll see what I mean.

>>1577169
I know that Vs would actually be approximately Vo = Idn * 10Kohm if only QN is on. Since Vo would be a positive voltage in that case then for QP, Vgs > Vt and therefore cut off. When solving these kinds of circuits, do I just make educated assumptions and slowly verify them till I have solved the circuit and proven my assumptions?

I would imagine that there are some circuits where that just becomes very tedious...

>>1577184
Yeah you can learn by testing educated assumptions, and correcting them if they turn out to be wrong. Also you might as well learn by building and simulating circuits. Intuition in electronics comes from a lot of study and practice. Like if you know a mosfet is cut off, you treat it like an open switch. Or if you have an amplifier, you can study it with the small signal model. Or if you see a forward biased diode, there will be about 0.7V across its terminals for small currents.
"Advanced" books like for switching power supplies will also show simplified versions of circuits (closed and open switches) to help teach how they work. And you'll see the small signal model used for amplifiers.

>>1576957
don't they have pretty cheap harbor freight etc meters that can do this. really depends on budget

>>1577031
>made from electronics components he loved so much
yeah kinda true. was that an all-in-one CRT pc? hard to tell from vid

>>1577216
Thanks for the help anon. After taking a hulu break, eating dinner and taking a shower, it hit me that Vs in that circuit for both transistors is dependent on the gate voltage. So if Vg > Vt only Qn is on , if Vg < Vt only Qp is on, and if Vg = 0 both Qn and Qp are off. Not sure why it didn't click until after all that.

>>1577278
Here's a better and simpler way to put it.
N-FET normally off device. P-FET normally on. When you exceed Vgs(th) with an N-FET the device turns on. When you exceed Vgs(th) with a P-FET it turns off.

Same principle applies to BJTs as well but instead of voltages it's currents.

>>1577281
>N-FET normally off device. P-FET normally on.
Unless it's depletion mode.
And I'd personally use the active-low and active-high digital logic terminology, not that it matters too much. I find that when you use a transistor in common-emitter/source mode, it's easy to intuitively understand that pulling the gate/base to that closest voltage rail will turn it off, but I'll admit totem poles and other topologies can confuse me.

>>1577293
Almost all MOSFETs are Enhancement. If you run into any depletion mode MOSFETs they will be N-channel and you should simply treat them the same as a JFET. You will never run into P-channel depletion FETs. They are literally never made.

>>1577299
What are depletion modes used for anyhow? Just the same as JFETs?

>>1577302
I don't know off hand. Typically a JFET would be better than a depletion MOSFET in most applications since JFETs have a lower gate capacitance due to their structure. It's possible there might be high power applications for depletion mode devices in which case the MOSFET is your only choice because JFETs are only designed for small signal low power applications like amplifiers. I can't think of any examples though.

>>1577281
>P-FET normally on.
This isn't true. "Regular" enhancement mode mosfets, both P and N, are normally off when VGS=0V
If you think this is wrong, go look at a few P-MOSFET datasheets and report back.

>>1577305
In normal orientation (common-source), off = VGS = 0V which happens when VGG = V+, and from a digital viewpoint that means active low. Thinking in terms of VGS all the time can be confusing when what you'll be measuring will be voltage to ground, but semantically speaking you're correct.

>>1577308
Alright I understand where you're coming from

>>1577304
I'm sorry, what was that you were saying?

>>1577312
Replied to wrong guy, meant for this one >>1577305

>>1577316
This is what I mean:
P-MOSFET, VGS=0, no current flowing (well, 50nA leakage)

>>1577312
Your transistor is fucking backwards, that current will be flowing through the body diode.

>>1577320
>>1577321
Jesus christ, how can everyone be this retarded?

>>1577325
That transistor is a PFET (see its datasheet), hence the body diode goes from drain to source, pic related. The voltage across it is also ~0.6V, which it wouldn't be if it was conducting normally since the on resistance of the FET should be peanuts compared to that of R1.

>>1577312
You don't do this stuff very much do you. 652mV is from the body diode, you have this shit backwards you nutsack.

hey /ohm/

what is a redpilled multimeter for under $30?

>>1577363
>r for under $30?
god damn pajeets.
quality and poo tier pricing are mutually exclusive
you will need to drop down $50 as a BARE minimum

>>1577365

doubt.jpg

so do you have a recommendation or what?

>>1577363
Aneng AN8009 I think goes for under that price. It's a chinkmeter but it's pretty well regarded on a few online communities. /g/csg/ at least likes it, as does EEVblog.
https://www.aliexpress.com/item/actually-both/32842548196.html
This listing has both the AN8008 and AN8009, the 8009 is more recent and has NCV sensing and thermocouple capability. Though it looks like the AN8008 has a function generator capability that the AN8009 lacks. Both models have no moulder rubber on the outside, so don't drop them I guess.

>>1577236
HP2621A dumb terminal

>>1577302
mostly higher-voltage stuff where JFETs can't reach. HV current sources, SMPS startup power, stuff like that

>>1577383
>auto-ranging
hard pass

>>1577363
DT-830

>>1577390
>not liking autoranging
Are you an absolute shitlord or not? It's got a manual range button at the very least. Having a bunch of settings on your front dial is clutter that makes the signal traces within snaky and suboptimal, handling ranging in firmware is far better.

Big clive grabs live 240V AC mains wires for our viewing pleasure
https://www.youtube.com/watch?v=-5R-KBa18ME
And it's not some faggy finger to finger zap, it is a proper hand to hand straight through the heart zap
this shit is why mains should be limited to 50V tops and just use thicker wires

>>1577402
>Are you an absolute shitlord
yes, to 5.5 digit accuracy
>a manual range button
>a
needs two at least: one to move the decimal point to the left, one to move it to the right
otoh with my trusty rusty B+K 2704A I can just twist that dial to what I want, and move a slide switch for dc vs. ac (if I need to). within a second, I determine with certainty what decimal place and unit are going to be displayed
t. started out on an analog meter and fuck you

>>1577419
fine b8, I r8 5.8/8

>>1577402
Using analog and non-auto ranging stuff makes you much more aware of imprecision and errors. After some time. Autoranging is just something that saves half a second.

Can I make a dash cam for my hatchback with zero experience in electronics (except learning "V = IR" in high-school physics class)?

Stuff to be bought:
- Raspberry Pi 3B or 3B+: $35 (https://www.adafruit.com/product/3055 https://www.adafruit.com/product/3775))
- - Raspbian: Free
- - ffmpeg or raspivid: Free
- - Clear case base (hats don't fit under a cover; maybe add a cardboard box): $5 (https://www.adafruit.com/product/2253))
- - 64-GB uSD card: $0
- Raspberry Pi Camera 2: $30 (https://www.adafruit.com/product/3099))
- - Mount: $5 (https://www.adafruit.com/product/1434))
- - 3-ft cable: $4 (https://www.adafruit.com/product/2143))
- Uninterruptible power supply (includes on/off button): $20 (https://www.adafruit.com/product/4114))
- - Socket riser header (with extra-long pins): $2 (https://www.adafruit.com/product/4079))
- - 2.5-ft USB-to-uUSB cord: $0
- - - Car-to-USB adapter (5 V, 2.4 A): $9 (https://www.newegg.com/Product/Product.aspx?Item=N82E16875922049))
- - Li-ion JST battery pack (3.7 V, 2.2 Ah; smallest capable of 2.4 A): $10 (https://www.adafruit.com/product/1781))
- - - Pins-to-JST cord: $1 (https://www.adafruit.com/product/3814))
- - - JST-to-uUSB charger: $7 (https://www.adafruit.com/product/1904))
- PiRTC clock (includes riser): $6 (https://www.adafruit.com/product/3386))
- - CR1220 battery: $1 (https://www.adafruit.com/product/380))
TOTAL COST: $135

Not included:
- Apparatus for securing all this stuff on that little box in the middle of my car (some kind of wire rack to bring the camera to eye level?)
- - A literal dashboard camera is an illegal windshield obstruction in my state.
- Quarter-inch Allen wrench (for camera mount; not super-necessary?)

>>1577420
>needs two at least: one to move the decimal point to the left, one to move it to the right
You're right, that would be great.

>>1577433
It's usually fairly easy to tell what range the DMM is in, at least on my high-quality superior japanese multimeter. It has both a bar graph at the bottom to show you what fraction of the maximum 6000 count it's at, and a prefix right next to the value to tell you what range you're in. The autoranging is also really speedy (600.0Ω to 60.00MΩ in less than a second) and the display updates 4 or 5 times each second.

>>1577441
Yes, but I'd compare the price to existing solutions on ali or ebay. Also buying those components off ali in the first place will save you probably half that price.

>>1577098
Thank you Anon. You're amazing.

>>1577456
Sounds like a normal case of "if in doubt, check the datasheet" to me, but knowledge of anons who actually know what they're looking at can be a very valuable thing. This is a good board.

>>1577441
depends, how's your software skillz?
>Raspberry Pi 3B or 3B+
>Raspberry Pi Camera 2: $30
if you don't mind fucking around with some software, there are Orange Pi versions of the same that are half the price, or, use a USB web cam
>$50 worth of power stuff
sounds like you're blowing a lot of money on power supply. save yourself a shitload of money and time https://www.kickstarter.com/projects/1312527055/raspberry-pi-car-power-supply-ignition-switch

>>1577456
cheers. tbf it was just a couple of google searches

>>1577393
Yup DT-830B €3.31 Free Shipping

>>1577503
aren't voltage dividers built on a principle where whatever you draw from them has much higher resistance than the resistors used in the divider, and thus only affects the voltage drop marginally?
it wouldn't make sense to have 3 identical resistors

>>1577506
correct
but that meter's input resistance is on the order of 1Mohm/volt so probing it with that meter will perturb the divider and the reading

>>1577506
The idea is to calculate what the meter shows and then compare to a better meter that has 10MΩ input resistance instead of 1MΩ. You need to know what to expect when using a cheap multimeter.

>>1577509
>1Mohm/volt
A typical DMM has a fixed input resistance. Ω/V belongs to moving coil meters and the actual input resistance depends on the chosen range. Example 500V DC range: 40kΩ/1V = 20MΩ/500V. At the 1kV range the input R is 40MΩ. All this points to the full scale input current (the sensitivity) of the meter: R/V = 1/I, I = V/R and 1V/40kΩ = 25µA.

How do I do my own cpu and test it?

>>1577557
if you have to ask, you can't
go read over other people's CPU projects and decide what the hell you even want to accomplish

>>1577559
Nothing can stop me, gimme these articles. I could not find decent ones

>>1577566

https://boingboing.net/2016/08/19/meet-the-megaprocessor-a-16-b.html

>>1577566
most CPU homebrewers are happy to provide justifications for their design decisions and motivations. for example:
https://gigatron.io/
http://www.homebrewcpu.com/
also if you can find a pdf of Pic related, it is good reading, almost mandatory
it's important to have a goal for the project. what are you trying to do? implement a particular instruction set? explore a wrinkle in instruction encoding? experiment with microcode? restrict yourself to a certain number of ICs/LUTs or board size? achieve some minimum clock frequency?
what medium do you expect to use? FPGA? 74xx discrete logic? 40xx discrete logic? discrete transistors? what limits? will it be modular? will there be an external bus of some sort?
what kind of peripherals do you want to drive? video? audio? keyboard? UART? RAM? flash? SD card? how closely will they be coupled with your CPU? how do you propose to get data onto and off of the machine and what equipment does that require?
you have a great lot to think about, grasshopper

how does a capacitor work in a dc circuit? its my understanding that it only lets as much current through as can be stored in its plates. after that, all current stops since the plates are charged, right? so what role do they play in a dc circuit besides just preventing it from working? every text I read just shows the same pld graph and says 'they store voltage' but I want a more intuitive understanding. What function do they serve in a DC circuit if DC is what they dont let through?

Recommended topology for a VFO with differential outputs? Being used as a local oscillator for a mixer. Needs to be able to oscillate at up to 10MHz but only needs to be swept over a 20kHz range. The "up to 10MHz" requirement is to allow me to use crystal ladder filters to bandpass filter the first IF.

Ok, so after doing a repair job on an old cellphone phone I had lying around two ports of pic related (antenna/wifi port)got snapped off the board. If I buy conductive glue/solder glue and use it to reattach the antennas to the side where its broken off will that work ?

>>1577614
It kinda depends. The "they store voltage" (or more precicely charge) isn't wrong. That's also why sometimes disconnected capacitors are still problematic. You can use them to delay DC signals (if you want them to charge until they reach a certain threshold) and many more things. There's a wikipedia page that's pretty to the point albeit not limited to DC:
https://en.wikipedia.org/wiki/Applications_of_capacitors#Pulsed_power_and_weapons

(1/2 due to character limit)
So I want to build a watch with certain features. I do want it to be analog (Don't like digital watches)

Fig1 shows the basic idea. By turning the yellow knob you can move the yellow points rotation. When the hour hand matches the point something happens.
So basically like a timer.

Here's how I thought I could implement it.
First is using a quartz movement and using physical contacts between the hour hand and yellow point.
That being said physical contacts are a pain in the ass and will wear out far too fast for no real reason.


Fig 2 is a quick sketch of the basic idea, you can press a push button to choose between on and off. If it's set to "on" it sends a signal every time the time matches the set time. (in that case a clk signal is sent).

(2/2 of post before)

All of that is significantly easier (and more economical) with a microcontroller I guess.
Any ideas on a microcontroller that either uses the signal from the quartz movement or has an integrated RTC (or a cheap small RTC chip to use instead).
The requirements to the microcontroller are as follows:
A analog input to set the wanted trigger time (potentiometer comes to mind so a way to read the voltage)
A digital input for the push button
Low power consumption
A digital output to trigger the other circuit
If it uses its own RTC a way to control the watch faces.

Thanks for some ideas for ICs or implementation.

>>1577614
In a very pragmatic point of view, capacitors just slow down current changes.
Ideally on DC circuits they just act as open circuits, but in practice you never get a perfectly DC source because you always get unintentional voltage fluctuations due EMG noise, DC converter ripples, electromechanical components, your circuit non-linearities, etc; so you'll always have some AC component there, be it intended or non intended.
You often use 0.1uF capacitors close to MCUs in microcontroller circuits to diminish the risk of unintended behavior, as some unintended voltage peaks may reset the MCU, so they are put between Vcc and GND to keep the voltage as close as DC as possible. (These are known as decoupling capacitors)
In higher performance MCUs or CPUs (such as the Raspberry Pi schematic) you often can see series of capacitors in parallel. You can learn by experience that they are actually non linear and you need them to attenuate these peaks in different frequencies, and they don't just sum each other.

More info: http://hsi.web.cern.ch/HSI/s-link/devices/g-ldc/decouple.pdf

>>1577641
>slow down current changes.
*voltage changes, sorry

>>1577641
almost everything you just wrote is incorrect
try learning some stuff before attempting to teach others, kthxbai

>>1577643
Teach me senpai

>>1577619
probably digitally synthesized, these days. anything else is likely to require continuous adjustments

>>1577640
RTC functionality could be implemented in software. capacitive sensing may be your best bet for sensing proximity of plates to some fixed position. you could monitor the positions of mechanics to middling accuracy. WLCSP and flex circuits might be up to the size requirements. you probably had best talk to a watchmaker

>>1577640
Didn't think of capacitive sensing, makes sense though, thanks.

>>1576925
you wont like the esr, most of the pulse energy will be turned into heat in the caps if using $20 of cheap caps.

>>1577643
Is that picture satire? I refuse to accept these niggers are this retarded.

>>1577661
google image search says it came via 9gag

>>1577419
>mains should be limited to 50V tops and just use thicker wires
The measured voltage between his hands was ~35vac when he was getting the 10ma current.
You solved nothing and just made every mains wiring job more expensive (thicker wires).

Pretty easy question, I found a circuit for those "shocker pens" from childhood, where you press the pen and get a small shock (pic related)
I found two version though, one where the orange wire in the picture is there and one where it isn't and I don't understand the advantages of having it. Should it even be there?

>>1577643
(you)

>>1577603
thanks for response

> Pic related
It isnt last edition, right? Berkeley probably switched to risc-v arch recently

>>1576962
the other responses are wrong. you can not sell mains powered machines that are not double insulated to the public because a wire inside or wtf could come loose and touch the conductive steeel case or leakage current and some moron would touch the energized conductive case at the same time as something grounded and then they anhero from the shock. a completely plastic exterior like a power tool with at least one plastic gear doesnt need a ground.

>>1577679
oh you should certainly get the latest edition. I've got the 3rd ed. in hard copy in a box somewhere and it was quite educational

>>1577681
not even if they're in metal chasses and earthed?

>>1577233
op specified safe. theres a ton of expensive safety stuff in a Fluke such that if you accidentally throw it across a MV circuit it wont grenade your hand off or electrocute you or flash burn your eyes. basically if you f up using a fluke the meter will die but you wont; cheaper manufacturer meters will result in some higher level of personal injury. sparkies take this to stupid levels, anything is safe enough to see if an AA battery is dead or not. also the odds of a shitty meter having an intermittent switch are higher leading to measure a wall socket incorrectly as 0.0 then hurr durr stick dick or whatever in actually energized wall outlet and die. Flukes generally dont break so its harder to die of a false reading.

>>1577674
ok orange provides a ground reference between two sides of the transformer.
i mean its not going to make any difference really, you are getting the output voltage between the terminals either way.
i would leave it out because it can only cause trouble if you manage to short the high side of the output to e.g. the button through the hand it might damage something on the low side. impossibly unlikely but no use in having it so why bother.

>>1577689
I phrased it bad, cant sell 2prong metal devices but obviously grounded steel boxes are safe assuming the ground is actually hooked up

>>1577703
Okay, that sounds easier. I would have left it out just because I didn't understand the use. Also a ground reference in this use case (to shock) shouldn't matter right? Only the voltage matters? So essentially if I did connect it I would have the potential downsides but none of the upsides.
Thanks for the explanation

>>1577703
>>1577709
the orange wire is probably a common connection between the battery spring and the butt end of the pen to ease manufacturing. insulating them would only complicate the process. so you don't have to connect them but you could if it makes your life easier

>>1577713
>probably a common connection between the battery spring
thanks this makes sense

>>1577652
suggested MCU: https://www.youtube.com/watch?v=q3aorW4KhCY

>>1577693
>stick dick or whatever in
Typical fantasies of overprotected children devoid of common sense.

Hi /diy/
I'm trying to fix gf's headphones by re-soldering the wires to one of the speakers but I don't remember much of the class I had on soldering many years ago.
My problem is that the wires refuse to get into the melted tin... It looks like the tin is repelled away.
What do?

>>1577738
use flux
or use conductive glue instead

>>1577738
Remove the enamel.

>>1577741
>>1577743
Thank you for the hint guys

>>1577623
>conductive glue
Fuck no, you'll need to reflow it with solder paste and hot air. It's just a single part so you may be able to do it with a paint stripping heat gun. First I'd wick existing solder and debris away with a normal iron, and check that you haven't delaminated any of the pads. If you have delaminated pads, then the only option I see is to physically solder on the antenna cable, as trying to scrape out a place for that port to go would be nigh impossible.

Can someone in plain terms describe the difference between an fpga and a cpld.
I used an fpga in uni years back. And want to buy one to try misc stuff. But the really cheap things popping up are cplds.

>>1577737
This is legit common in China, it weirded me out when I was there to see kids on leashes like pets.

>>1577855
CPLD logic is built from a few large OR-AND arrays with fixed routing, also relatively small and simple, a bit easier to predict in their timing, mostly useful as e.g. glue logic or level translators or other bus-type stuff
FPGA logic is built from many small blocks containing a 4-bit LUT and a DFF with flexible routing fabric between them, relatively large, you specify the timing and the toolchain works to satisfy your spec, often have specialized hard cores e.g. high-speed serial communication, hard CPU cores, RAM, hardware multipliers, flash
there are hybrids and fence-sitters between the two. e.g. Altera MAX II is has the typical LUT/FF architecture but low-density and no extras
>really cheap
Lattice has an iCE range of true low-density FPGAs starting at like a buck each. you can get the iCEstick eval board for like $25

>>1577868
Not him, but are CPLD programming environments as bad as FPGA ones are rumoured to be?

>>1577870
I see many vendors have unified their design suites, so probably yes? I don't find them particularly awful but I generally only use them to compile source into a bitstream, instead using vim and iverilog to work out my design

>>1577855
oh, another difference is that CPLDs are almost always nonvolatile and boot right into the user logic, whereas FPGAs usually but not always rely on external configuration, whether as masters reading external flash or as slaves to a processor

>>1577876
Do CPLDs use EPROM or something? From what I learnt about PLAs and/or PALs those guys use polyfuses.

>>1577878
>EPROM
meant to say PROM

>>1577878
they seem to be EEPROM-based, programmed thru the JTAG port. there were also some EEPROM-based PALs (GALs?) too though most were indeed OTP and a rare few were windowed EPROM
just to be funny, the Lattice iCE chips have an OTP memory that's programmed via its SPI slave, but can also be volatilely configured via the same port and can also load from a 25xx series SPI flash. like some other FPGAs, they can self-reconfigure, one absolute madlad created a USB bootloader for it https://github.com/tinyfpga/TinyFPGA-Bootloader

This may border on /qtddtot/ however it applies specifically to electronic design.

Anyways, I'm looking for a category of switches or I suppose more accurately a switch bank, wherein one switch out of many (lets say 4 for example) is always on: when you activate one switch, it disables all other switches e.g. only one switch can be on at a time.

I'm looking for something like in pic related, or sort of like a cassette tape deck when you're playing a tape then press the stop button. One button should cancel out another and only one button should be active at a time.

My question is, what exactly are these called?
Googling "selector switch" is the closest I've gotten, but I know there's a specific name for them. I heard ages ago that they were called something like radio switches, though of course "radio switches" is too broad for successful googling.

Note that I'm not looking for a rotary switch: I need something where the order of positions isn't predetermined.

>>1577957
https://www.surplussales.com/Switches/SWPushB-2.html

>>1577957
I'd personally do it with digital logic or an MCU. With analog switches if need be.

>>1576957
Check out commercial electric, extech, klein, or bk precision. You might have to live with a manual ranging meter but at least you will be half way safe if you know what you're doing.

>>1577976
What's an MCU?
Also, found some interlocking push button switch banks. What I don't get is how do I wire them? How do the other buttons "know" when to mechanically turn off when one is turned on?

>>1577363
Fluke 87v bare minimum or you are a scrub

>>1577957
These are called "interlocked switches" -- my fluke 8020B has them, and that's what fluke called them. Digital versions are bullshit.

>>1577985
Microcontroller unit. An arduino but not for babs. I imagine the "knowing" is either done via a mechanical connection between them (the button assemblies will need to be right next to each other for this to be a possibility), or by latching solenoids (for which there will need to be a few more pins than usual).

>>1577999
>Digital versions are bullshit
Sure you don't get the tactile feedback, but you're also not paying $100 for fucking buttons.

Could probably make a 3D-printed or laser-cut version of the latching mechanism if you really wanted to also, provided the parts can take the spring load and 200lb gorillas.

>>1577985
MCU = Microcontroller. Can do a row of momentary switches and do the one active in software, preferably with some kind of LED or other indicator near the switches.

Other names to search for are ganged, banked, n position push button. They work by having some kind of bar / rod that when you start pushing one button in releases the latch on all buttons, then when you push a bit further it latches that one in. Move it slowly and you can usually get it to a state where all buttons are out, or sometimes two buttons are pushed in.

Should I leave comparator input pins grounded, pulled to Vcc, or floating?

>>1578034
Pull to ground.

>>1578036
if I pull them both to ground, wouldn't there be some ambiguity as to which input is larger, hence the output might mess about?

>>1577510
>a better meter that has 10MΩ
surely chinks won't skimp on that, 10mohm resistors costs like .001 cents more than 1mohm

I am trying to bypass the power of an old samsung phone from battery into direct power. So basically the positive wire of usb cable to positive terminal of the phone, and - to -. Is there such thing that i can use to reduce the 5V being given by the USB port down to 3.7V that's only required by the phone?

>>1578042
I don't believe most meters achieve their 10M input resistances with resistors. I'd think 10M resistors would be too noisy since Johnson-Nyquist noise increases with resistance and having a lot of noise on the input of precision equipment like scopes and DMMs would be bad. They typical approach would be to use high impedance buffer amplifiers or INAs and they'd achieve a precise input impedance by applying the right amount of negative feedback.

>>1578053
old ones did (the point-to point hand soldered ones), nowadays meters are just a black blob of epoxy.

>>1578057
Well yeah of course because they didn't have active electronics... not counting vacuum tubes and VTVMs obviously.

>>1578058
How does 'active electronics' handle a few hundred volts without a divider chain?

>>1578062
Okay so *some* inputs use T and pi attenuators as well... minor oversight. My bad.

>>1577649
>digitally synthesized
Sound like a pain. I'm gonna try generating the sweep at low frequency with either a a ramp generator and OTA based VCO or ramp generator and a V->F converter like the LM331. Then I'll use a PLL like the 74HC4046 or CD4046 and try to multiply that up to the frequency range I want.

Not sure this will work well in practice but in my head it seems like a good idea.

>>1578070
Where do you learn about PLLs? To me they're as alien as Dirac notation or microwave PCB design.

So, I want to make a circuit to measure the current of my DC motor so I can reduce the pwm as the motor starts to stall/torque. Is this a brainlet idea or is it feasable?

>>1578066
Neither T nor PI. No impedance matching.

>>1578088
>all that dielectric grease and physical contacts
>this is what rangelets think is superior

>>1578087
Sure it is feasable, allthough most people would want to increase power when a torque is applied. But I am assuming this is some sort of savety feature. Just be carefull that you don't make a system when the motor shuts down directly after you start it from the startup peak current. You could do this by either making it slow acting or making it take into acount the current you are applying (probably using some sort of microcontroller.

>>1578086
Try the wikipedia page? It's just an oscillator controlled by a servo loop which adjusts the frequency to sync with some other signal.

PLL frequency synthesis just divides the oscillator output by an integer factor and syncs the result, so the oscillator frequency is an integer multiple of the reference frequency.

>>1578088
that image makes me physically ill.
this is why hobbyists are considered a joke
this is the same faggotry as "audiophiles" using 4inch thick cables to play their flac files

>>1578038
there is usually a slight imbalance that will keep them from doing so, such that they eventually settle into one state or the other. each type's datasheet will usually contain advice on what to do with unused circuits in a package. I vaguely remember one type suggesting one input to V- and one input to V+ but I don't remember which

>>1578046
yes, a step-down converter. available for pennies on alibay
or just fix the USB jack on the phone

>>1578070
surprisingly, no. you literally just buy a chip and send it I2C commands. the Si5351 is a common choice
but if you wanna sweep, your design is a better choice

I know I can't have batteries in parallel because one will drain the other, but can I use diodes in parallel + a capacitor so that if I disconnect one pack, the other will take over?

>>1578327
sort of, both will drain evenly into the load. use low-voltage-drop diodes for best efficiency. the capacitor doesn't do anything for you

>>1578331
The capacitor is there to provide current while either battery is disconnected and the other isn't fully engaged yet

>>1578327
You can have batteries in parallel so long as they're the same type and charged and discharged in unison. But once one starts to fail, it will increase the load on the others which will cause them to age faster. So the life expectancy of the set is basically the life expectancy of the worst battery in the set.

>>1578335
the diodes take only microseconds at most to recover from reverse bias (tens of nanoseconds for Schottky or fast-recovery diodes), and if they're already forward biased (as they would be if the battery voltages are within ~0.6V of one another) they won't need any time at all. you would only need the cap if both batteries were disconnected at once. most likely, whatever load you're supplying will have plenty of capacitance to tide it over for that infinitesimal sliver of time

>>1578339
they're separately removable, not bound into a pack. stfu Donny

>>1578327
Diodes+resistor of low value. This is one of the few modern uses of diode logic, an OR gate.

I need to discharge large ammounts of current from at mains voltage from a capacitor. I'll be using IGBTs I scavanged. But there is an issue, I need to clamp the gates at less than +-20V, probably with a 18v zener. But how do I feed the zener with it's current from mains without using
>transformer
>dc-dc converter
>50W resistor
?

>>1577419
>this shit is why mains should be limited to 50V tops and just use thicker wires
Amerilards regularly set their shitty wooden houses on fire by overloading their anemic extension cords. Limiting the mains voltage to 50V would essentially turn the whole country into a raging inferno because the average tard would get a extension cord rated for 500-800 Watts since it's the cheapest and plug their entire fucking kitchen into it.

>>1578362
what are you trying to avoid by ruling those out? why would it be a 50W resistor and not 3W? you’re not leaving yourself a lot of options.

>>1578384
Money and heat.

>>1577419
>what are transmission losses

>>1578372

we need a name for this new religion of making up shit in order to excuse hatred of some other group. Trumpterians is too awkward.

>>1578391
>denial
https://www.youtube.com/watch?v=63vzHhCOku8

>>1578362
the zener doesn't need any current if it's not conducting i.e. clamping. it'll only conduct at 18V or whatever in the reverse direction (and 0.6V in the forward direction, as usual) and stay out of the way the rest of the time
otoh if you are confusing voltage and current maybe you shouldn't be doing this project

boards for my binary to decimal converter came in today. Couldn't help myself and had to put one together tonight. Completely useless, but I think it's neat. Probably gonna machine an acrylic housing for it. there's also a small software bug with it that I gotta track down

>>1578430
>software
I thought you used only logic IC's

>>1578430
nobbad :D
'twould have been a fine project for babby's first SMD imo
now use that ICSP port to chain two together for 32 bit/10 digit conversions

>>1578430
Lemme guess, you used a MCU because binary-to-7-seg ICs don't exist, only BCD-to-7-seg? I hate it too, but I'd have gone full retard and busted out the 74HC00s and what. But the plus side is that you can just write a bit of code to do a binary-to-hexadecimal conversion, right?

>>1578434
the chip on the left is a PIC16F677. The others are encoders and a multiplexer. Just reads in the inputs, compares that with a lookup array of values, adds it all together and spits it out. Right now when all switches are off it's displaying a zero in the tens place and not the ones place

>>1578441
I've only had a couple experiences with SMD shit and it did not work well for me. Think I'm done with this, really just wanted a knick-knack. maybe one day i'll make a number systems calculator in binary decimal and hex

>>1578450
>lookup values
>not just a huge boolean expression
make a 16-input karnaugh map, I dare you

So I've been trying to rig up a capacitive touch switch as the power switch for my PC. I figured it should be simple enough, the existing switch just connects one wire to another wire, I can do that. So I do, using the left circuit, and it doesn't work. So then I put in an LED across the non-live wire from the switch, just as an easier way to verify that the switch is on when I touch it. And the circuit as a whole works fine, turns on the computer. But why?

>>1578458

>>1578465
Switch might need an internal pulldown? I'd plug in a normal switch, and measure its voltage before and after its depressed.

>>1578466
Sorry, that's *35* 16-input karnaugh maps. I'm actually the anon who might have posted here a year or so ago about writing a shitty code to turn truth tables into digital karnaugh maps and then into logic espressions, but I haven't touched the code in ages and I doubt I ever will again since it's a right pain to get working. I'm also the anon for whom logisim doesn't work for, which might or might not be obvious.

>>1578465
transistors ≠ solid-state switches. for one thing, they're polarity-sensitive
if either of those switch wires connects directly to one of the power rails (either ground or the 3.3V standby supply) just ignore it and use the touch module to drive the other one directly. quality touch modules can be configured for active-high or -low output (and anything with a TTP223 can be hacked to change the active state by changing the input on pin 4)

I want a controllable high voltage DC-DC converter that regulates the voltage on a leaky capacitor from 0-15 KV that takes in about 7.4 volts from a small lipo or something. I've heard I need something like a flyback circuit to do this?

>>1578562
>7.4V to 15kV
lol good luck with that buddy. It's doable but if your capacitor is large and its leakage current is high then you're gonna have a problem. If your flyback has to supply a constant 10uA at 15kV to keep your cap charged up for example that's a constant 83mA on your primary which is gonna kill your battery pretty fast. Good forbid it has to draw more current not only will battery life get worse but let's say the cap is REALLY leaky and it's leaking 1mA then you need to supply over 8A peak primary current which means larger wires/traces and larger transformer core and windings plus your battery life is probably under one hour now.

Also you need to consider the leakage inductance spike on your primary which will reach several thousand volts and that will destroy your transistor. Getting FETs that are rated for 15kV Vds(max) is pretty much impossible or will cost you an arm and a leg. Also simple solutions like flyback diodes will most likely melt in this scenario as well so you'll need to look into more complex snubber design. It is extremely challenging to design switched mode power supplies with such large boost ratios.

Anyway, I do have one question, if you have a leaky capacitor why don't you just replace it?

>>1578577
the capacitor is an actuator. The leakage isn't that bad. I will calculate it later. It's easier to let a capacitive actuator just discharge through a resistor rather than making a complicated regenerative circuit.
>>battery life
5 minutes is all I need. I really just want something small. Right now I have a tiny 5V-10 KV DC-DC converter that can only provide half a watt, but I want higher voltage. This DC-DC converter is really nice, but crazy expensive.

>>1578590
Oh and I need like four of these things, which is why I don't want to use expensive DC-DC converters

>>1577619
not sure what any of this meant but boy does it sound badass

>>1578562
A flyback with one of those 20kV silicon stack diodes should do it. Can be pretty simple too, but you'll need to buy a few very large value resistors for the feedback divider. Might I ask what the capacitor is? Thinking of rolling my own HV capacitor.

are most you guys uni or trade fags?

>>1578612
~18 year hobby, took 1 uni course online

>>1578612
Uni, physics.

>>1578612
EE student here, not indian; god bless

Went to get my quarterly chinesium dose and saw this. Sad!

>>1578663
Makes me feel a bit better that the US isn't the only country with ridiculous import tariffs anymore.

>>1578665

>>1578612
Backend programmer salaryman who makes electronic projects to not let my EE abilities get stale.

>>1578602
a dielectric elastomer actuator. I already got some big boy voltage dividers.

>>1578612
basically the same as
>>1578700
but hourly and full-stack

>>1578665
the US import tariffs almost never apply to small personal shipments. I've ordered hundreds of dollars worth of chinesium every year and never once had to pay a duty

>>1578663
kek, you dun goofd
you can still make larger orders from one store that ar elike $49 so the import fee is small in comparison

I've had one of these asus vs247h-p monitors since about 2013 and the backlight recently went out in it. I took it apart and replaced the power supply board and the backlight is still bad. I know in the older monitors with CCFL backlights they had a small inverter board which was normally the problem if the backlight went out, however I think this monitor has LED backlights. Is there a similar inverter board I can replace for these newer LED monitors?

>>1578742
open it up and see
also go find a service manual

How the fuck do I design a board with plated sloted holes in a way that JLCPCB will recognize them as slots? When I send the gerbers the previewer shows just tiny holes

Also:
I tested this >>1576976 in a breadboard with the pulldown on the + input of the lm358 and it worked, the transistor being driven by it can turn a led on a off, meaning a small toy motor won't turn on in the "off" state (I hope)

>>1578738
I routinely make orders that are over 200$ and don't pay any import taxes because they're small boxes and the chinks write like 2$ on them and brazilians don't bother to check anything. It's a numbers game, some times people will get charged 100$ import taxes on a cellphone case and some times they'll let industrial motors pass through without any dues

>>1578758
Perhaps consider moving into a 1st world country.

>>1578760
One it's too expensive
Two, first world countries are being flooded by third worlders so there's no point

>>1577980
Any opinions on Southwire?

>>1578769
> flooded by third worlders
then maybe get of your butt and do something about it?
you can afford a white sharpie can't you?

Stupid question here.
Can I use blow torch to remove SMD ICs?

>>1578386
Okay but why 50W. You didn’t calculate that number you just pulled it out of your ass.

>>1578784
Hm, I tried, seems to work.

>>1578784
I'd say yes because you can control how much heat you put by increasing the distance but don't do this if you can't destroy the board/ic's

>>1578781
Do what about it? I'm not moving to another country especially if I'll be scammed by brazilians on my stay there

>>1578731
Don't know where you're buying from but on Digikey/Mouser I'd say at least 75% of components are marked tariff applied.

>>1578742
An LED TV shouldn't need an inverter, the cathode lamps need x-x-xtra high voltage, LEDs don't.

>>1578753
try putting a cutout on the edge layer in the middle of your pad and make sure the copper goes right up to it... if they even support plated slots. otherwise just drill a big hole

>>1578760
I hope you're not
>implying the anglosphere hasn't devolved to Third World status as a matter of fact

>>1578831
NRND stuff on alibay, fren
>tariff applied
which doesn't affect me if the supplier elects to absorb the tariffs. I usually expect Digi-Key single prices for ICs to be kinda high anyway

Which is the better way of wiring a three-phase transformer for single-phase AC, left or right? (Just look at the field lines I drew and ignore all the symbols). Currently it's wired like the one on the left and it works ok but the right would seem more logical to me because more of the field lines would stay inside the iron thingy.

I just realized that a chain of 74HC74 would make a nice "stackable" LED indicator beacon. Each LED module would have a single 74HC74 with D-flop #1 working as a data register and D-flop #2 as an output register. That way data bits could be shifted through the chain of flip-flops without affecting the state of LED outputs, avoiding any flickering.
Then any required number of LED modules could be stacked on top of each other and each of them could be controlled individually via a simple 4-bit serial interface: data, clock, store, reset.
A PLC (or microcontroller) with 4 digital outputs could be used to control this device.

>>1578863
Example LED beacon

>>1578817
I needed do de-solder an audio jack from trash board, and couple SMD MOSFETs.
Worked surprisingly good, much better than construction heating gun.
I guess visible flame helps.
One board is a kill tho, FR2 is shit.

>>1578715
>dielectric elastomer actuator
well you learn something new every day.

>>1578863
A 74HC595 is a shift register with latched outputs that basically does exactly what you're describing. It's already used for this kind of demuxing by many including the tarduino crowd, which I think is why it's one of the most purchased ICs on AliExpress.

>>1578665
US is really cheap compared to most of the world.

>>1578753
Lazy solution is just make xboxhueg round holes. Smart solution is use a different format, standard Gerbers + single NC drill file doesn't really do it in a standardised way. Altium can do a separate NC file just for the slots, but cheapo china manufacturers may not support it.

>>1577503
You can get it like for 1$ including shipping, and it wont explode at 220V

>>1578790
hyperbole.

>>1578935
I have such a yellow meter and I'm on a standard 3~/400/N/PE distribution. It doesn't splode on 400VAC either. For electronics I mostly use a 4000 count 10MΩ meter though. If you do the suggested calculation you know why that is.

I'm reading The Art of Electronics. They keep on referring to __x chapters, but they're nowhere in the book.
What are they referring to? The supplementary lab book?

>>1578992
>a yellow meter

I googled "yellow meter" and no success.

I want to learn and this seems like the place, but then shit like this.

>>1579003
yellow probably means a Fluke meter

>>1578926
I did do the huge holes but they're really terrible for soldering

I present to my fellow /ohm/budsmen the hysteretic constant-current converter. advantages:
>easy to design
>no loop compensation required
>very fast response
disadvantages:
>output current ripple required for circuit to oscillate
>switching frequency varies with load

>>1579032
y u no SMT power jacks

>>1576976
I added a 100k resistor between the output and inverting input to see what would happen and the result was that I got, besides the clearly on and clearly off states, some kind of gradient between on and off. Without the negative feedback resistor there is pretty much no input voltage that will put the transistor into linear mode, it's either saturated or it's not. Now... what would happen if I placed the resistor between the output and the non-inverting input? Would it make the opamp change states abruptly with an even smaller difference?

>>1579035
try it and see
*cough*hysteresis*cough*

>>1579002
I think it's some kind of follow-up book that expands on select concepts in more detail. No it's not out yet and there is no release date set as far as I'm aware. We'll get it when they finish it I guess.

Fuck, just shocked myself with 240V mains for the first time by touching a driverless LED. Was significantly less painful than an electric fence so I suspect something limited the current significantly, but it left a burnt spot on my finger and it smells like burnt hair. It also felt like more of a gripping sensation, which I guess is due to the frequency. Since I'm on the second floor I guess my capacitive path to ground was less than it could have been? I wonder how you'd measure that capacitance.

>>1579090
Got RCD protection? If not, get it, don't die.

>>1579092
No, but only one hand contacted so I wasn't exactly in danger of passing a current through my heart. Might consider getting one anyway though. An RCD would not protect me from putting both hands across mains as the current would flow through me to neutral, so perhaps an antistatic band on each wrist tied directly to ground through an RCD would be a smart choice.

>>1578866
see thse on top of self service checkouts

>>1579090
>I guess my capacitive path to ground was less than it could have been?

i dont believe this for a second. you wont get burned through a parasitic capacitor unless you're dealing in transmission line voltages (10,000+ V say). you just touched 2 hot points.

>an antistatic band on each wrist

grounding yourself to protect from shocks has the opposite effect. instead of needing to contact 2 hot points, you just have to touch 1 and you're zapped. (however the 1M resistor inside these bands will mean you get tickled but not hurt)

>>1578911
Yes of course, but 7 outputs would be kind of wasted in this application.

>>1579135
No I was thinking using two metal contacts, one on each arm, and grounding them through a GFCI, such that when you accidentally touch a live pin, the breaker trips and the current will take the path through your hand and therefore not through your heart. Even if you grab neutral with one hand before touching live with the other, the thing still trips without current going through your heart since the path down the live-side wire to ground should be much lower resistance. I think you can still get a shock even if you touch live without touching neutral at mains voltages, so now I want to plug a large resistor into live and touch it just to see what happens.

Ok with a 100kΩ resistor and dry hands, I'm at 200VAC from me to ground, so that's somewhere between 40VAC and 120VAC depending on how reactive the load is. That means somewhere between 400µA and 1.2mA through the resistor and I. My current impedance is somewhere between 160kΩ and 500kΩ, meaning the maximum current that could go through me is somewhere between 480µA and 1.5mA. So you're probably right.

>>1579142
At only 2 D-FFs per 74HC74, I'd say that's much more of a waste of silicon.

>>1579144
RCD usually trip at 10 or 30mA, adding a resistive path is just going to make it hurt more.

>>1579159
>adding a resistive path
No it would be a dead short from my wrist to ground, so the resistance would be as low as possible to trip the RCD as soon as possible. And I think the bathroom outlet-style RCDs trip at much lower currents than DIN RCD breakers.

>>1579003
They have different names and the colors are mostly yellow or black. If you search for '830B multimeter' you'll find something.
Image and instruction at http://jbryant.eu/pages/DMM.htm

Does the 10000 µF capacitor in this shot look like it's gone bad to you?

I have this old Yamaha receiver that stopped working. (it keeps turning itself off a couple of seconds after you've turned it on) I'm not all that handy with electronics, only soldered a handful of trivial things, but I figured I could try replacing the cap before throwing it out and getting a new one.

>>1579211
middle one is bulging a tad

>>1579212
That's what I thought as well. There also seems to be a tiny amount of residue on the top. Could also be some crumbs that fell through the grating though, not sure.

Guess this is more of an SQT thing to ask, but how careful do I have to be when taking this amp apart? Do I still need to manually discharge all caps when it's been unplugged for a day or two? The two largest caps are 6800uf@63V. I made sure not to touch the electronics so far.

>>1579224
Just stick a screwdriver near the capacitors you think you might touch/handle. That smidge is just dirt, the top of those caps is made to break (it's hermetic-ish otherwise).

>>1579250
in the capacitors near where you think*

>>1579224
For those caps to still hold a significant charge after 24 hours, you'd need a leakage current below 5uA. The usual estimate for leakage current for aluminium electrolytics is I=0.3*sqrt(C*V) where C is the capacitance in uF, V is the rated voltage in volts, giving a leakage current in uA. 6300uF@63V gives 200uA, which would discharge 63V in about half an hour.

Will I be happy with this soldering iron?
>https://www.amazon.com/X-Tronic-3020-XTS-Digital-Display-Soldering/dp/B01DGZFSNE

I would like to get into synthdiy mostly, but also have some electronics around the house that could repair and would like to become an /ohm/ monk.

>>1579295
BTW, I like that this model (compared to the Hakko FX888D) has knobs to turn rather than a menu to hop through, and that it's a bit cheaper.

Why multimeter gives resistance of scheme: 3542 Ohms, but when i calculate it on paper its clearly 3500 Ohms? Maybe i am retarded. but:
Resistance of Paralel
1/R23 = 1/R2+1/R3 = 2/660=1/330
330 + 230 = 560
1/R678 = 3/330 = 1/110
560 + 110 = 670
670 + 550 = 1220

and now down:
230 + 210 = 440
1/R101112 = 2/440 = 1/220
220 + 220 + 780 =1220

Parallel of:
1/1220 + 1/1220 = 2/1220 = 1/610

2890 + 610 = 3500

3542!=3500
Is there some resistance of multimeter alone?

>>1579300
i believe most resistors have a tolerance of +/- 5%

>>1579295
sure it ain't bad
but for $30 more you can get one with an air tool, if you don't mind taking up a bit of bench space

>>1579300
measure each resistor individually and plug them into your network analysis

>>1579300

maybe electronics is not in your future.

>>1579354
>ain't bad
it's hard to tell what i want when I've never even soldered anything before, but I know I'm going to be using it frequently (which, for me, means at least once a week). what do you recommend? the models I hear about the most are the Weller WP25 or WP35, the Hakko FX888D and then the one I linked.

I would like something that is as dummy proof as possible, so the digital models seem nice with the temp control and all, but I don't want to do menu hopping and would rather turn knobs. What is this model with the air tool you speak of?

>>1579300
Your calculation is correct but your diagram is flawed (no current loop). Where does the value of 3542Ω come from? Do you have the real circuit diagram of this (yet unnamed) multimeter?

>>1579295
Do you never read the customer reviews?

>>1579372
I've had good luck with Aoyue soldering kit, but the one I usually recommend (888A 2-in-1 model) is out of stock. lots of chink manufacturers use the same model numbers for a particular feature set, so if you search for 852D and 952D stations, you might find something good. usually there aren't any menus to hop outside of supervisor settings, most of which you're probably not interested in (idle timeout, temperature setting lockout, etc.) but I totally feel you on preferring physical knobs

>>1579387
awesome just what I needed to solder some glonuts

>>1579387
Of course I do. And I read product specifications and manuals. But I don't know a single thing about electronics or soldering, so it's a lot of hub-ub and getting personal advice is great.

I finally finished and polished the control circuit of a resistance spot welder with a supercap bank which charges up to 5.5V.
It happened to me once that all four MOSFETs that control the on/off circuit blew up and I want to make sure that doesn't happen to me again.
As for now, I implemented in the microcontroller of the control circuit a time lapse of 400ms and a cooldown of 3 seconds. When I popped my MOSFETs this cooldown wasn't implemented and the pedal button was tied with a bad 10k resistor to ground, so it tended to activate with my fingers touching pretty much any contact of the circuit. Now I fixed it and put even a parallel 0.1uF capacitor in parallel just to play safe.
Now, the question is: How do I need to tune the variables such as distance between welding terminals and on/off switching time lapse for my MOSFETs not to get destroyed again?
As for additional information, they use a MCP1407 driver, are driven by a Vgs of 12V and have a 1MOhm resistor between gate and ground.

>>1579394
>And I read product specifications and manuals
you're all right, anon

>>1579417
>parallel 0.1uF capacitor
don't worry about this, debounce in software. just loop a bunch of times (at least 20ms) making sure the trigger input is always high
>tune the variables
by making them not so variable
you might want to consider that the leads to the welding rods have inductance which you will need to take into account, if you don't instead just outright swamp them. at the very least you will want to add a power diode to shunt any kickback energy around the FETs
>As for additional information
this is an image board. post image

>>1579423
Okay, here it is. I had to do some fixing to the circuit so it looks like a bunch of parts stitched together.
As for diodes, can reverse polarity 1N4001 work?

Does a mosfet trigger when the gate to ground voltage reaches a certain threshold or the gate to source voltage?

>>1579447
Gate to source. A device doesn't know or care what "ground" is, it only knows the voltages at its pins.

Does a mosfet trigger when the gate to ground voltage reaches a certain threshold or the gate to collector voltage?

>>1579442
>1n4001
no, too wimpy and too slow
>long wire between gate and driver
>why are my fets blowing
go read the datasheets of every part you are using, in full, especially layout guidelines. you are not dealing with the most forgiving technology

>>1579469
>layout guidelines
It's the first time I had to deal with these issues, so I guess it's time to fabricate a PCB then. I ignored such things as parasitic inductance thinking that this would just work, but at least that stiched-together circuit served me to test some other concepts.
Thanks for the feedback, anon.

>>1579450
It's the voltage from the gate to that source that matters, hence it being written as V_gs (threshold). If the FET is an NMOS and in a common-source arrangement, then yes that's the voltage between the gate and ground.

>>1579486
You are fine using your perf board. Just avoid all those boards, all those wires etc. Or buy a sheet of copper clad and use an knife or something to carve out different sections.

how/why does placing a resistor between stages of a circuit prevent attenuation from circuit loading, if the resistor is just going to dissipate the signal anyway?

>>1579450
Sort of. Think of it like a resistor between Rds controlled by the voltage between Gate and Source (Vgs). You make it twice or three times the Vgs (voltage threshold) and you get a near open circuit between D and S (Rds-on). You make it below Vgs and you get a closed circuit. You make it anything in between and you get a very non-linear variable resistor.
Now, depending on your application you might want to consider other variables as well. Any projects in mind?

>>1579450
Sort of. Think of it like a resistor between Drain and Source (Rds) controlled by the voltage between Gate and Source (Vgs). You make it twice or three times the Vgs (voltage threshold) and you get a near open circuit between D and S (Rds-on). You make it below Vgs and you get a closed circuit. You make it anything in between and you get a very non-linear variable resistor.
Now, depending on your application you might want to consider other variables as well. Any projects in mind?

Is there a single-component version of this?

>>1579662
Use a multiplexer and use the selector as the inverting input

>>1579662
Just use a resistor-transistor inverter if you don't want a whole 74HC14, assuming it's high-speed enough. Also you should be using NANDs or NORs anyhow, which you can easily turn into an inverter by tying their inputs together.

>>1579670
>>1579665

Well turns out there are transistor-sized single-gate versions of 74hc08 and 74hc04, so I went with them! There are even smaller versions but I won't be able to solder them by hand and the pick and place machine we'll be using is probably going to have trouble with them. I never cared for logic gates but now that I need to make a board for newer motor drivers + a microcontroller running an old firmware I love them.

Apologies in advance for having to make this thread, but do any of you know how to turn one of these damned things off? At a relatives, and the pool filter won't stop running. The metal On/Off bar won't budge- but afraid of really putting muscle into it to move. Are the timer pieces supposed to be hard to move as well? I can't get them to move. Really afraid of breaking the damn thing.

If I'm majoring in computer/electronics engineering what would be the best programming language to learn if I know next to none

>>1579682
C

>>1579682
Haskell

>>1579679
It's already in the off position, looks like you need to replace the timer

>>1579691
Oh, I googled that. My bad, didn't specify.

Update though, if anyone manages to google this in the future: If the pool water filter metal piece is stuck or doesn't move, it has something to do with a piece blocking it from shutting down. Place your hand around the entire timer dial (yellow in above picture). Pull the dial toward you and try turning clockwise or counter clockwise. Spin it like 4 hours on the dial markings, then set it back into place. Try pulling the metal piece to OFF now, it should work.

Example I found helpful.
https://www.youtube.com/watch?v=P7iOh_Plh2g

>>1579651
nice usage of open vs. closed there

>>1579675
I gotchu senpai, see Pic related figure 3

>>1579450
>Does a mosfet trigger..
>>1579651

Beware of memetronics.

A MOSFET has no threshold voltage in the sense that at a certain Vgs (e.g. 2..4 volt) a magic contact suddenly closes turning it from OFF to ON. Data sheets explicitely specify what 'threshold voltage' really means in this context and how to measure it: Vgs=Vds (gate connected to drain) and Id=250µA (mostly). The operational conditions can be found in the diagrams elsewhere in the data sheet.

Pic: In reality a MOSFET has no threshold at all. Vt (2.1V in this example) only marks the voltage where the exponential characteristic (at very low currents, right diagram) transitions to the quadratic one for higher currents (left diagram).

>>1579726
>2N7000
What year is this again?

>>1579746
The really old ones can be useful if you want atto-amp gate currents.

>>1579765
Oh do they have a particularly low capacitance?

I'm trying to control the speed of bldc fan (this one: https://www.digikey.com/product-detail/en/sunon-fans/EE92251B1-000U-A99/259-1647-ND/2037307).).
I tried doing this by connecting fans positive to +12V and negative to n-mosfet drain (at first BS170, then IRLZ34N) and controlling the transistor gate with pwm output from uC. This approach worked for me when controlling brushed dc motors and simple loads like leds.
For whatever reason the fan stalls and whines loudly as soon as I go bellow ~90% duty cycle. I tested the same circuit with a led and it works (the led dims as expected). I then tried messing around with pwm frequency (I went from 62.5kHz to 500Hz), but it only changed the pitch of the whine. I also tried connecting the fan to l298n breakout board and controlling the enable pin with pwm, same result.
At first I thought that the fan is unable to run at anything below 12V, but if I supply voltage from power supply, the fan runs at voltages as low as 3V without any problems.
Where the fuck am I messing up? Do I need to smooth the output voltage from the transistor?

>>1579795
It could be too fast for the fan's internal electronics to boot, or perhaps you need to pick a frequency significantly below that of the oscillation frequency. I know I dimmed a similar fan at like 1Hz and it worked fine, but an anon here told me it should work at high frequencies too. Nothing wrong with throwing an inductor and capacitor (and protection diode or two) to smooth out the waveform though, because it's basically guaranteed to work if the fan is fine with lower voltages.

>>1579795

the fan contains an electronic circuit to generate the necessary waveforms. when you supply it with a PWM voltage, the circuit is losing power on every cycle, so it just cant do anything right. presumably, above 90% duty cycle, a capacitor is holding enough charge to keep it constantly powered. so try adding a big old cap across the fan.

The heating element of my soldering iron died, so I dug out an old one that I remembered I had. Anyone know what kind of a tip this is or what to google to find replacement ones?

>>1579809
I've so far managed to find out that the soldering iron is an "Express Tempo 49". Made by a French company. They do list replacement tips on their website, but the only contact info on there are two phone numbers: one for a distributor in France, and one for a distributor in Poland.
Are the tips really proprietary and there are no 3rd party replacements? This is ridiculous.

>>1579811
Must be a free energy model no one dares to replicate.

>>1579720
I remember someone posting about this IC a few days ago but I forgot about it. With it I can use 4 IC's instead of 6. How does this motherfucker even work internally?

>>1579783
No, they generally have no gate protection (some manufacturers sneak one in though, not always possible to tell from the datasheet). Without gate protection there's just a relatively thick slab of silicon oxide in between the gate and the channel, close to no DC current can pass.

On the other hand that's why you have to be very careful using them, they are extremely ESD sensitive. Unless you really need an unprotected gate, don't use them.

https://www.edn.com/electronics-blogs/tales-from-the-cube/4392140/The-shocking-truth-about-FETs

>>1579802
>>1579808
I tried connecting various capacitor values I had laying around (from 100nF bypass caps up to fat 100uF electrolitic cap). At lower values nothing changed (motor stalled and whined), at higher values the motor simply ran at full speed (even at 10% duty cycle).
I connected the cap positive to +12V and to motor positive, cap negative to mosfet drain and motor negative. Please tell me I'm not retarded and that's how you're supposed to connect caps.
Anyways, do I need a full RLC filter for it to work properly? I tried adding a resistor (10k and 1k) across capacitor, but nothing changed.

>>1579861
>full speed (even at 10% duty cycle)
That has to be expected. Next try a resistor of ~10% of the nominal fan resistance from drain to motor||capacitor. Adjust value of capacitor and see what happens.

>>1579875
ok, I'm completely fucking lost now.
am I supposed to connect the fan to capacitor or to +12V and mosfet drain?
I also tried adding a low-pass filter, but without any luck. All the low pass filters I find are filtering the high side, I need to control the low side. does that matter?
This is how my circuit looks like right now: tinyurl (dot) c/o/m (slash) y2xyodnz. No wonder it doesnt work, the output is jumping between 6V (expected value) and 0V.
The fan draws ~0.2 amps @ 12V, so I guess its "nominal resistance" is ~60 Ohms?

>>1579919
Why don't you start the fan at 100% and then use its own inertia to keep it spinning at lower speeds, with a pretty low pwm frequency like 200hz?

>>1579929
that doesnt work, I've tried it. like some anons already noted, its most likely that the motor has some brains inside it that refuse to work if not supplied with constant voltage. its stalls not due to lack of power.
what i need to do is smooth out the stepped voltage in the circuit here >>1579919. unfortunately im too much of a brainlet to actually figure out how to do that effectively.

>>1579919

>>1579938
Also put a diode across the motor, from it's "negative" to the "positive" rail.

>>1579849
iz just logic

>>1579953
>diode across the motor
Voodoo diode would always be in reverse mode.

Trying to make another bench supply, this time using a buck converter to improve size & efficiency.
My question is, how can I make a regulator IC track the SMPS stage? All I could thing was pic related, it doesn't seem to use a switching output as reference a voltage.
>>1579966
>switching inductive loads

>>1579967
>switching inductive loads
Meme doesn't apply here, capacitor across load.
>All I could thing
Nice Failstad painting. Want to heat something?

>>1579974
Edgy

>>1579967
depends on how you're setting the output voltage. three-terminal regulators don't have separate feedback inputs so it's hard to play games with them. you might think about inflating the buck's voltage instead of deflating the linear's voltage. assuming you're using a dual-ganged pot for voltage adjustment, just add a couple of diodes (or a TL431 or whatever) into the voltage divider for the buck
>output on the pass transistor's emitter
you should probably reconsider this for stability's sake
>a brushless dc fan is a simple inductive load
hmm

>>1579967
Do it the other way around, make the buck converter follow the linear regulator output plus a couple volts. Just use a PNP transistor with 2 resistors as a common emitter amplifier to reference the voltage difference between buck voltage and linear regulator output to ground, to feed back to the buck converter.

That way you can use a proper steady voltage reference for the linear regulator.

>>1579995
>>1579983
Thanks friends.

>>1580013
>>1579995
Could you help a brainlet? Do you mean to put the pnp output where? In the place of R2?

How would I decide what controller to get?
Project is using pic related, forward and reverse. Plugged into 110 mains, with as small of a footprint as possible.

I see a ton of different ones that could work but im unsure.

>>1580038
Nevermind, the circuit can't be that simple, you wouldn't be able to regulate down to zero that way.

It still makes more sense to simply drive the buck regulator output a couple of volt above the linear regulator output, but the feedback circuit is going to need to be a little more complex.

>>1580055
first, get any mains-to-12V converter of at least 18W/1.5A output current rating, that suits whatever other requirements you have
then, assuming you're controlling all this with an MCU or lolduino, any full-bridge controller that can handle 12V at no less than 1.5A can serve. L298 modules are cheap, easy, and popular

>>1579861
>>1579919
Just a capacitor won't do anything productive, you need to make a low-pass filter. I'd advise making an LC filter so it's comparatively lossless, but if the fan is low enough power that you won't burn a resistor then an RC will work fine. Not sure how necessary the diodes are, but they can't hurt. The circuit can be rearranged to work with an NFET fairly easily.

are capacitors like batteries?

>>1580072
Its going to turn a leadscrew on a manual lathe, using a pulley. Other people have used this motor for this already and its apparently worked out fine.
Really basic, turn on and off with a switch, no arduinos or anything.

But I want to get as much power out of it as possible at its low RPMs (itll be used mostly at 30-40% of its rated RPM). I know thats what the geared head is for, im not sure if the controller will matter any.

Do I need headroom in my power supply?
If I buy a 3A, 36W power supply and then get a 6-30V, 10A max controller itll be good?
Would there be a tangible difference by getting a 10A power supply?

Okay faggots, bear with me. I have a sensitive 3.3v piece of equipment that draws 120mA and apparently cannot deal with switching DC-DC supplies, because of ripoff. Are my calculations for using an LDO instead correct?

So I am trying to settle with a single cell 1Ah li-po battery to feed the 120mA*3.3v into it.

Current draw that LDO will do is the same 120mA, power consumption will be 4.2v * 120mA = 500mW when fully charged, 460mW on 80% and 450mW on 50%
The LDO will give out 3.3v * 120mA = 400mW, wasting 100mW in heat (1/5th of the all power) when fully charged and avout 50mW (1/10th) when charged 80% and below.
Assuming discharge rate for a 1Ah battery is 0.1C, the voltage will drop below 3.3 only when the battery has less than 5% of capacity left.

So LDO will waste ~1/10th of the power and ~5% of the capacity.

Am I correct?

>>1580078
To a small approximation, yes? They store charge proportional to the voltage across them, and since current is charge per unit time, a constant current into a capacitor will mean the voltage increases linearly. Practically speaking, capacitors have a much lower energy density than batteries, and their output voltage will vary much more than batteries. But they can take far higher currents, hence why they're used to buffer noisy power supplies' output voltage, and why they're used as the power supply for coilguns and the like.

>>1580084
I'd first look up the datasheet for your LDO to see how it behaves with only a ~0.5V drop or less across it. At those levels your ripple rejection is also likely to be awful, but since you're drawing from a lithium ion it shouldn't matter much.

>>1580087
> At those levels your ripple rejection is also likely to be awful
The key word being instead, no DC-DC, just li-po and LDO

>>1580092
I know that, that's why after the comma I wrote "but since you're drawing from a lithium ion it shouldn't matter much".

>>1580118
Sorry for that.

Most LDOs i've googled so far have dropout of 0.4v, that means I need 3.7 and above, wasting 30% of the capacity. That is disappointing. I've also found ones like LD39015M33R that do insane 100mV, leaving me with 3.4, so pretty much still have 90-95% of the capacity with my load of 0.1C. That's very cool. That one works, right?

>>1580129
Might want a boost converter and then a regulator, or perhaps two cells in series. I'm no expert on those regulators but if the datasheet says so it's probably a good one to go for.

>>1580082
>manual
then you don't need a controller at all. just reverse power to it with a DPDT switch. there are even switches with center-off that are ideal for that
>headroom
I'm no motor expert but yeah headroom is a good idea. can't see it making good use of much more than 3A. 12V power supplies are cheap enough

test

>circuit is working fine off 5V breadboard supply
>oscope showing perfect behaviour
>unplug breadboard's supply
>swap it to use a CR2032 instead
>the waveform on the oscope is all over the place
>can't figure out what's wrong
>realise I forgot to put the scope's ground clip onto my negative rail
Wow that's another thing that spice simulates well!

>>1580076
you see, the problem is that im too much of a fucking idiot to understand how to filter the low side (with n mosfet). all filters i find online are filtering the high side. and i cant seem to find relevant info googling. heres my pants-on-head attempt: tinyurl (dot) [com] (slash) y3qmptmz

>>1580323

you're using the cap to smooth the waveform across the resistor, which is kinda silly. you wanna do it across the motor, or both.

this seems to work a bit better. the average voltage drops or rises with duty cycle.

>>1580323
Yeah that's basically it, though you forgot to model the inductance of the load (which is the entire reason that we put that diode there) and forgot to include the diode for the inductor itself. I'd also put the diode from the inductor-motor junction to Vcc instead of to ground. The cap can be either to ground or to Vcc.

>>1580335
I think the motor is represented by the resistor, the inductor is part of the low-pass filter.

Does anyone know what kind of part that is?

>>1580129

>>1580340

3 coils and 2 caps, so it looks like a custom hybrid filter. its function could be anything.

>>1580335
>>1580337
yeah, sorry for not being clear enough. the 60 ohm resistor is supposed to be my fan (it draws ~0.2 amps @ 12v).
i finally managed to find some relevant information by searching for buck converters (afaiu, this is basically one). pic related is the circuit i have atm, seems to be working alright (tinyurl (dot) [com] (slash) y2nyt47f).
I dont really get why, but if I connect the floating diode (from inductor-motor junction to vcc, as suggested), the voltage doesnt seem to follow duty cycle as well as without it.
also, am i risking frying my little n-mosfet? its rated at 350mW, but at the very start (if I reset simulation) it shoots way up to ~5W for several cycles before settling.

>>1580345
That advice is useless, please read the original post. I cannot use switching regulators, their ripoff just drives my device nuts.

>>1580068
I asked stack exchange a dude gave me this, seems to be a decent idea. I'll built one when I have the time and report back, if some other anons want a smps with a regulator stage.

>it converts the drop across the regulator into a ground referenced voltage on the feedback pin.

>>1578612
34y/o EE
Uni straight out of high school
Been tearing shit up and building stuff all my life

I work as a reverse engineer for the DoD

>>1578849
Assuming all of the windings are going in the same direction, left is correct.

>>1580397
Why my rocker switch has three terminals?It has only 2 states. They are labeled 1b,1 and 1a.

>>1580406
No idea why I quoted the cia agent. Sorry.

>>1580406
common
normally open
normally closed
com and no are connected in pos.1, com and nc are connected in pos.2.

>>1580398
Why? I know that both are sub-optimal since it was designed for three-phase AC, but it makes way less noise when wired like the right and so far it seems to work.

spent 15 quid on my componenets because there's a minimum order for most of the bits
I'll have enough to make at least 10 of what I want

Any recommendations for soldering/basic electronic kits for someone just starting out?

I'm struggling to find suitable switches
this is the closest I've found that meets my requirements, ideally I want a cluster of 9 or 10 using as little space as possible
my original design uses 9 of the dip buttons which is workable but very cumbersome, this would be better but still not quite there

>>1580435
there are 1.27mm pitch SMD "DIP" switches

>>1580425
congratulations, you just leveled up your parts inventory

>>1580443
SMD scares me
I'm not the most dexterous with a soldering iron

>>1580443
use pins and jumpers

>>1580450
there are also 2.54mm pitch SMD "DIP" switches, if your concerns involve vertical space
1.27mm pitch is pretty easy, only twice the density of a DIP. extra paste/liquid flux and a small screwdriver tip <=1.2mm makes it easy
>apply flux to pads
>melt small drop of solder onto a corner pad
>add flux to pad
>apply heat to pad and solder drop while introducing and positioning component with vacuum tool or tweezers, allow to cool
starting at opposite corner:
>heat pad and lead
>feed very small amount of solder into joint
or,
>apply small drop of solder to clean iron tip
>heat pad and lead, allow solder to flow into joint
or
>say fuck it and use pin headers and jumpers

>>1580356
That is a good idea, reduces the minimum margin of the buck converter significantly over the single transistor I suggested.

Simplest would be to have R1=R2, for buck being 1.8V over the output

>>1580450
Dude. Just add flux, you literally can not put too much flux.

>>1580493
Would a gallon be enough?

>>1580493
cool
thanks

what's flux

>>1580501
barely

>>1580502
frickin' magic, that's what

I disassembled my projector to hopefully swap the fan for something more silent but to my surprise it's a blower like on GPUs.
What do? Adapt the design to acomodadte a fan or find a silent blower?

>>1580511
The kind of fan shouldn't matter, as long as its blowing the air out of the device (to keep dust out) and providing enough airflow (keep low temp).

That said I'd be careful to meet the second requirement.

>>1580406
Another posibility is that the third terminal is for a light indicator. I'd check continuity to rule out the NO/NC

>>1580349
Try actually grounding the capacitor, as now it simply goes in parallel with the motor. I'd think that you'd get better performance. I think pic related would work well. And 100uF is easily possible at 12V if you think you need it, but higher than 1mH might be pushing it.

>>1580600
pic related

>>1580523
It's actually turbining the air towards a funnel that hits the LCD panel, it's got an opening on the underside

>>1580600
power rating of a MOSFET is generally an average. check the datasheet to see that instantaneous current through it is below the pulsed absolute maximum and average current is below the average absolute maximum
>floating diode
steals current from the inductor that would have been forced thru the motor or cap and freewheel diode

>>1580610
congratulations, you just invented a pulse-with modulator with a shitty LC filter on the input

>>1580593
>>1580408
looking closely there is a neon indicator inside. Now I'll figure the correct configuration

>>1580622
>pulse-with modulator with a shitty LC filter on the input
Wasn't that the entire point? Oh and I guess putting a single diode across both the inductor and motor would be better, but not by that much. Don't see how the FET's power rating has anything to do with it, I'd be surprised to see it dissipating more than 50mW.

>>1580623
With this configuration, the light is on when the load is on. If you switch mid and bottom wires then the lamp is always on. Hope it helps

>>1580631
thanks my man

>>1580625
no, see
>>1579795

>>1580636
He's trying to decrease the speed of a motor with PWM, but it doesn't like the chopped-up voltage, and runs on voltage as low as 3V. So an LC filter will turn that PWMing into an approximately DC signal with lower voltage, making it work with the motor. Unless I'm missing something?

>>1580659
>doesn't like chopped-up voltage
>totally not at all what >>1580610 does as far as the load sees
do you even electronics?

>>1580662
I revise my design.

>>1580610

the cap is not doing anything useful. its purpose is to feed the motor when the FET is off, thus smoothing the change between full-on and full-off. but when the FET is off, how is current gonna flow out of the cap and to ground? it doesnt have a path.

>>1580689
I revised my design.

>>1580692

still useless. the load gets a square wave. the cap does nothing.

>>1580694

>>1580696
the inductor resists changes in current and supplies the load through D1 whether the cap is there or not. from the load's perspective, it doesn't really do anything. fine when using a pure resistive load, but something more complicated than that will have problems

>>1580700
Even with a load that also has 1mH ESL I'm still getting a perfectly flat current and voltage through the load. Not sure how much switching a fan does though, you'd have to look at its current with respect to time with a scope, but I think the capacitor would smooth that out, no?

>>1580704
>Not sure how much switching a fan does though
a brushless dc fan? plenty
>I think the capacitor would smooth that out, no?
hmm, perhaps, but it's in series with V1 and its impedance, and the antennas aka wires to and from V1, which is not friendly from an EMI perspective. may as well put it across the load alone

>>1580709
In that case, would the 2 seperate diodes be better? So if the load goes open-circuit, the inductor itself won't have any issues with voltage spiking, just the internal inductance of the motor, and I assume the motor's circuit is already built to handle these switching spikes since it would experience them in ordinary non-PWM operation.

How do I filter? Is there something more in making output filters for smps? Because if I go by the basic 2nd order system thing, I'm getting mF to 10mH passives for a 100hz cutoff.

>>1580717
Also, should I put the filter inside or outside the feedback loops?

>>1580717
PSU output filters are typically just a few big capacitors in parallel, inductors aren't good to introduce into a general purpose circuit like that where the load can change the current its drawing.

>>1580714
any motor inductance would be in series with L1
wouldn't even bother with the second diode. if the load goes open, the worst that happens is the stored energy in L1 and C1 gets dumped back into the power supply through D1. far from fatal

>>1580737
sometimes, perhaps more often than you might think, it's a pi filter

>>1580433

pls respond

>>1580754
Kits? Just order everything you think you'll need, kits usually come with a bunch of shit you'll never need. Temperature controlled soldering irons/stations are superior, get soldering flux, solid core wire is nice for breadboarding, get a couple breadboards and some perfboard too.

>>1580766
I think he means one of these

>>1580767
Oh practice kits? I'd go with something somewhat useful, like a VU-meter or remote garage door opener. But I would rather design my own circuits as it actually teaches you how the circuit works, so you learn electronics and soldering at the same time. Failing that, getting a kit with a circuit diagram and documentation on how the thing works could be good, though you're not likely to get non-chinglish documentation. For this reason I'd go for a discrete-element but elegant circuit like a AM or FM receiver, H-bridge, amplifier, etc.

>>1580770
>But I would rather design my own circuits as it actually teaches you how the circuit works, so you learn electronics and soldering at the same time
easy to say in retrospect. but there's a bit of a steep learning curve in design vs. simply executing someone else's. maybe that's an argument in favor of breadboards vs. practice kits
admittedly the noobs should be reading theory, but better not overwhelm them too early with too many failure points

>>1580773
>easy to say in retrospect
You're probably right about that. Especially since working with a nice PCB with plated thru-holes and solder mask isn't something you get to do until you've really spent a long time on a design, and while I personally enjoy that process more than the finished product, if you want to learn quickly that's a bit of a pain. Not to mention troubleshooting of someone else's design is usually not very necessary or fairly simple, I know troubleshooting my current (shitty perfboard) project has required heavy use of an oscilloscope.

But I stay by my recommendations, learning about amplifiers and H-bridges and demodulators like that is learning about all the different lego bricks that fit together into one more sophisticated circuit.

>>1579582
I read some datasheets and I found out about TVS diodes. As these MOSFETS don't need to generate too many fluctuations too quickly, are these a good option to clamp the voltage between gate and source to, say, 15V?

>>1580433
It depends, what do you want to do with electronics? Robotics? Audio? Microcontrollers? Interfaces? It all depends on that.
Anyways, in any case you would need one of these assorted Resistor and Capacitor kits and a couple of LEDs. If you have money, afford yourself one of these USB oscilloscopes.

>>1580782
TVS aren't rated very precisely. one rated for 15V stand-off might start conducting anywhere from 16V to 30V. a plain zener would work better if you actually care about the voltage. for best results place it close to the first FET
to save power and prevent overheating of your FETs, you do want a good crisp rise and fall on the gate signal. I suggest you place the drivers close to the FETs too. the control signal is somewhat less sensitive to edge speed so can be located remotely

>people use ESP8266s as wifi modules for their arduinos
holy shit I didn't realise humans could be this garbage

If I wanted to read data (~256 bits / 32 bytes) out of an EEPROM without an MCU, are there EEPROM topologies that would easily let me do this? I'd be displaying the values on 7-segment displays, for which purpose I'd probably run them through latching shift registers and send those to decoders to display the 64 hex digits in series, probably lots of 8.

>>1580476
big ass knife tip > tiny screwdriver tip for SOIC

>>1580670
Better yes. This is a buck converter.

>>1580886
An unregulated buck converter.

Is it acceptable to make a 5V resistive dropper PSU like this? Resistor value isn't final. I won't be needing more than ~2mA of current and I need it to have this kind of reference to GND (hence not using a capacitive dropper). I don't think ordinary linear regulators will sink current to stop the voltage getting too high like a zener does, so I don't think there's a simple but better way of getting my 5V. I'll probably use a 1W resistor (or 2 0.5W resistors) just to ensure I'm not going near the power threshold since it will be in an enclosed case, and as-is the power consumption is averaging at around 0.5W. To get a little more leeway with the allowable current drain I may have to drop that to a 82k or even 68k, which is really pushing near that 1W, so I may want to hit up 3 1/2W resistors in series instead.

>>1580902
maybe