/ohm/ - electronics general

>>1198552

>>1198554
An Rpi can't provide enough current to drive the motors for an RC car big enough to mount an Rpi on. If you try you'll only break the pin or the chip. 5AA batteries in series will provide 7.5V. The Rpi likes 5V. You'll probably break it if you try. Buy a 5V regulator on Ali or ebay, and plug your Rpi into that, they cost almost nothing. 4AA batteries would give you 6V, which is still above what the Rpi wants. It'll probably work, but you risk breaking it. Your motors will be a lot weaker, though.

The chip on the car likely has two parts: A receiver and a motor controller. The receiver outputs pwm to the motor controller, so what you'll want to do is look at how the pwm is output, and then mimic that with your Rpi.
http://rcarduino.blogspot.com/2012/05/interfacing-rc-channels-to-l293d-motor.html does something similar, but with an Arduino. Instead of plugging the receiver into an Arduino and then the Arduino into the motor controller, you should plug the receiver into the Rpi, figure out the format of signals it outputs, and then figure out a way to have your Rpi transmit the same signals to the motor controller.

how do i dissolve potting? its hard and black (heh) so i guess its epoxy. its a magnetic pickup so im pretty sure it just has copper wire around an iron core in it.

>>1198590
If mechanical removal does not work, you're pretty much fucked.

>>1198590
You could try acetone, but chances are it could dissolve any ICs, enamel insulation, ferrite binder, maybe even any resistors. I'd test it on some components, wires, and ferrites lying about. There are multiple types of ferrites, so watch out.

If the enamel dissolves you're screwed, and if the ferrite dissolves you're also screwed, and if there are any passives in there you will find it slightly inconvenient to replace them.

File: labview_logo.png (15KB, 430x520px) Image search: [Google]

15KB, 430x520px

Thoughts on LabView?

Also how do I get good?

>>1198590
>>1198594

Sometimes you can melt it with a heat gun. Just don't cook the components.

File: 1354487102912.png (4MB, 1632x1224px) Image search: [Google]

4MB, 1632x1224px

http://www.instructables.com/id/DIY-EEG-and-ECG-Circuit/

Thoughts on this? I want to write and sell lie detection software, but I don't want to pay a fortune for at least a 6 channel device to test with, so if I can get a simple scalp net and some conductive paste, what is the efficacy of using radioshack grade components for an actual EEG?

>>1198652
>7 to 31Hz passband
I don't know about your lie detection software, but that's pretty awful for both EEG and ECG. Certainly it helps to reduce noise, which is very welcome for a breadboard contraption connected to a computer, but still.

>radioshack grade components
I have no clue what Radioshack sells, but I'd expect the main problem to be their component selection than their quality. If you end up building the input amplifier from op-amps on a breadboard, at least you can expect significantly higher 50/60Hz noise.

>>1198302
>>1198340
bump for help from last thread

>>1198606
I know an engineering company that uses it but they want to move away from it for bigger projects.

>>1198697
If you can find schematics you may be able to replace the burnt components. Or find someone whose weren't burnt and can tell you the markings. You're going to have a hard time guessing, though.

>>1198723
While the exact types would be difficult to guess, it seems that there's an NPN bipolar transistor and a P-fet. There's a decent chance that a random BJT like BC847 and a logic level fet like Si2323 would do the job. Guessing is of course the last option, but still.

>>1198723
>>1198725
Thanks, maybe I can buy an identical one on Ebay real cheap, to use as reference, then sell it back after I've made the repairs.

As a side note, How do I prevent the power backflow issue in the future? Could I just cut the 5V wire in the USB cable that connects the keyboard? or do I need to use a diode?

File: imretarded.png (489KB, 1132x310px) Image search: [Google]

489KB, 1132x310px

Could i connect a vumeter directly to the PC audio output?
Every VU meter i've ever seen were connected to speakers, but could i connect it in order to make it silenced/soundless? Or connect it to headphones?
I just want to watch the lights of the vumeter at night while listening to music tru my headphones without waking up my senpai.

File: common emitter amplifier.gif (8KB, 377x296px) Image search: [Google]

8KB, 377x296px

>>1199139
Depends on the voltage its meant to run on. if it's meant to go inside a speaker housing and take the already amplified audio signal, then probably not. I don't know what voltage amplified audio usually goes to, but you'll probably need to use an amplifier to get there. A single transistor amp like pic related should be fine.

If there is some sort of calibration potentiometer on the board then you might be able to get away with the amp being powered by the VU meter's PSU, or even just run it without an amp, so I'd mess around with the existing hardware a bit first. It might even have a built-in amp, though I wouldn't get your hopes up.

>>1199139

an LM3915 type device can be driven from as little as 1.2V full scale. the audio output jack is capable of putting out that kind of voltage into earphones, so you'll be fine as long as the input circuitry doesnt attenuate the audio signal. usually there's a pot for adjustment.

File: 38.png (155KB, 996x367px) Image search: [Google]

155KB, 996x367px

Am I going to receive 1200 pieces of chinese industrial slag on a metal stick with known approximate resistance?

>>1199293

There's a good chance that they will be fine. It amazes me how they can make resistors, tiny capacitors, crystals, etc., and sell them for pennies and still meet the spec.

Especially crystals. Clock crystals are made in the millions, but how the hell do they manage to be so precise for dirt cheap.

Anyway, my opinion of "made in China" is that they are in the transition period that Japan went through in the 60s. "Made in Japan" in the 60s was synonymous with pure garbage. By the 80s or so, Japanese products were high quality and affordable even after tariffs.

You can still find blatant fraud in Chinese products, but some of their tools at Harbor Freight are excellent and cost 1/4 of what you find in Lowe's or an auto parts store. I recently bought a YFA carburetor. Autozone, NAPA, etc. want 200+ for a rebuilt carb that can easily be crap. I bought a brand new YFA off Ebay for $97 and it seems fine so far.

>>1199293
I've never gotten a bad batch of electronic parts from China, but if you're that worried, just test the resistor before you install it, or test random ones when you get the package, and return if you find too many variances

>>1199352
well not really worried given the price, just wondering.

>>1199352
>just test the resistor before you install it
That doesn't solve the issue. The tolerances on resistors aren't just how much they're allowed to stray after the production, but also the maximum change in resistance after soldering and usage over time in nominal conditions.

>>1199374
>change in resistance
You should be able to find that out on your own. If your product is that sensitive to tolerances then simply buy elsewhere.

Two questions
One is working with electronics what electric engineers do? From my understanding it's designing's elections, but I want to get a second option

Secondly, If I wanted to build a sex bot what would be my best starting point?
Would it be desgin>build>test>?

>>1199446
>what would be my best starting point?
gaining experience with real women for reference

>>1199449
But anon I have a gf, I just want to satisfy my root fetish.

>>1199446
>From my understanding it's designing's elections
that may be true only if you are a russian computer engineer

File: these-13-sexy-cosplays-totally-shouldn-t-exist-but-the-twisted-part-of-me-is-glad-they-do.jpg (25KB, 500x750px) Image search: [Google]

25KB, 500x750px

>>1199463
>my groot fetish.

>>1198346
I know it's lazy, but does anyone know of a 5v step up module that'll exceed 2a from 3v?

>>1199497
>russian computer engineer
or a neolibconman

>>1199497
So what's the common thing that engineer to do? Just program?
>>1199498
That's a nice groot, but I meant robot

>>1199520
You have to understand that there's a big overlap between programming and electrical engineering. Any computer program can be made using only wire and relays, and a significant part of the engineering is making a circuit do essentially the same thing as a small program. For instance a transistor will open if the input is higher than a certain voltage. So in a 5V circuit, a 3.5V transistor is essentially the same thing as if (int i >=3.5) { }, or simplified, since the base voltage is going to be 5V, if (bool i) {}.

>>1199522
I understood half of that, going to order that book in the resources as soon as possible

>>1199446
> is working with electronics what electric engineers do?
No, it's what electronic engineers do. Electrical engineer = industrial electrician.

>>1199591
What it means is that if the transistor receives a high logic signal (input is sent to the middle pin), it'll open and enable the circuit behind it to work. So basically, the transistor is the same as the if () {} block in programming. If the statement inside () is true, whatever is in the {} block will happen. If the input to the transistor is high, the transistor will open and the circuit behind it will work.

An example is if you have a push button connected to a transistor, and a camera behind the transistor. In programming this might look like if (button) {takePicture}. Press the button and the camera does its job.

I want to send digital information (a bitcoin transaction, ~ 300 bytes) through radio, just as a learning experiment. where do I start learnin about sending and receiving radio?

>>1199780
>Electrical engineer = industrial electrician.
Could you elaborate? Isn't that more related with a trade ?

>>1199906
>where do I start learnin about sending and receiving radio?

to what? the internet? another radio?

you can attach wifi, bluetooth, xbee, and perhaps other radio modules to an AVR or anything with a UART, but this is all digital and is child's play when it comes to the radio part.

The confusing part of your post is the bitcoin part. what exactly are you wanting to do that you can't do with a computer or smart phone?

A smart phone is a radio, btw.

>>1199906
Usually people wanting to send digital shit wirelessly use some ready-made radio. In which case you usually start by reading the radio datasheet etc.
If you happen to choose some particularly shitty radio due to its low cost, learning about DC-less encodings, like Manchester code, could help a bit. Not that it's usually needed. Error detection and correction methods would be worth knowing as well, so that you can verify whether the received data was ok.

>>1199907

What he posted is nonsense. Electricians install things according to a code, whether it's at my house or in an industrial setting.

Electrical Engineers tend to design things. It might be something "simple" like designing a substation to meet a code, or it might be designing an electrical device like a motor, or they might sign off on the wiring plan for a house that was drawn up by an architect or a contractor.

>>1199927
What about electronic engineer?

>>1200023

what about it? they don't exist.

>>1200029
Aren't Electronic Engineers a real thing in your country?

>>1200023
Generally speaking electronic engineers deal with the transfer and control of data; for example microcontrollers, FPGAs, etc. Electrical engineers deal with the transfer and control of power; like motors, generators, grids etc.

>>1198346
Requesting help for >>>/sci/8997638

File: 847.png (29KB, 1431x716px) Image search: [Google]

29KB, 1431x716px

What am I doing wrong? it clamps correctly for higher amplitudes but doesn't work for ~1V.

File: panel.jpg (69KB, 1032x774px) Image search: [Google]

69KB, 1032x774px

(1/2)

Pic related, I have this small indicator panel. it was taken from an old Cisco router. I'd like to control the indicator lamps with a microcontroller so that I could use it in my projects. The panel also has three pushbuttons switches that I probably do not need.

File: tr.png (26KB, 875x323px) Image search: [Google]

26KB, 875x323px

>>1200053
oh wait, I think I got it, it's because I am working exactly at the diode forward voltage.

File: 1.png (136KB, 750x2019px) Image search: [Google]

136KB, 750x2019px

>>1200064
(2/2)

The problem is that I'm not sure how I should control the data pins, for example to scan the two 4-led columns sequentially.
I used a 5 VDC power supply and tried every twenty pins in an attempt to see which pin controls which LED. It took some time and produced this huge truth table, but clearly it can be seen that there are a lot of repeated patterns. Also, I think I may have made a mistake in LED1 as the table doesn't show 2 & 3 LEDs lighting up with any combination of inputs.

Currently I'm feeling kind of unsure how to transfer this huge truth table into a C program. Should I use some kind of array of bool variables to store it?

>>1200073

you've wasted a huge amount of time and effort unnecessarily trying to figure out this puzzle. someone with a closer relationship with reality would have spent 20 minutes removing the LEDs from the PCB and attaching them to a piece of perfboard, wiring all the cathodes together in a common ground, and running 16 wires from the anodes to the micro-controller, after putting a 330-ohm series resistor in series with each one. this would make any subsequent programing child's play. if 16 I/O pins is too many, shift registers can reduce that to 3 or 4, while multiplexing can reduce it to 8.

>>1200031
>Electronic Engineers

Electrical Engineer

People who use "electronic" are generally technicians, not real engineers.

As in Electronic Engineering Technology (EET certificate).

Not like BSEE (B.s..Electrical Engineering)

>>1200073
You're making this way more complicated than it needs to be. Use your multimeter or eyes, trace which pin on the input goes to the anode and cathode each on all the LEDs, they're most likely all facing the same direction.

From your table it's pretty clear P3,7,10,15 are all connected together and to all the anodes (assuming your "1" input is your +5v), then you just have to trace physically the line from each LED to the connector.

Once you've got that, you'll have 12 known cathode inputs, one for each LED, hook those up to the micro's GPIO pins, hook the common anode to +5, when you want to turn one on, turn the corresponding output to drive low.

Anyone here have any experience with making bike lights? I'm just wondering what sort of mcd or wattage values I should be going for. The local store stocks Cree LEDs with 45000mcd (which I think matters more than lumens) for 5.20 USD each, any chance I can get away with using a few of these without spending more than $40? This is not so I can see the road, just so other drivers can see me. I've got a bunch of extra low-wattage red LEDs that I figure I can strap on the back and they'll be fine for a rear light.

>>1200136
BSEE is a Bachelor of Science in EE, not an accredited engineering degree like BEng or MEng. Electronic/Electrical engineers have taken accredited degrees and may have become charted engineers by appropriate organisations (often the ones which accredit the degree). BSc/BSEE are sometimes given when a BEng/MEng student hasn't managed to acquire the number of credits for their accredited degree.

>>1200327
So I plan to go into college for this I should go to a school that has BEng or meng accreditation? Most schools in my area only have BSEE I think.

>>1200361
look for ABET accreditation ( most major schools will have this )

>>1200361
I couldn't tell you for sure, I know in the UK it's the IET that does accreditation, and ABET does at least a lot of it. Tbh, I think you'd be safe with just finding an 'accredited' degree, which will be advertised by the uni (means it's a worthwhile degree). Not 100% sure for you, as I honestly have very little clue what it's like outside the UK. Best of luck.

>>1199924
>>1199926

Thanks.

for the bitcoin part, just for fun, as in using bitcoin when internet is down in a locality.

Whats a good fpga to learn the basics on?

Anyone got experience with PICPgm or JDM Programmers? I can't get the former to detect the latter. I don't know what the detection protocol is so I don't know how to troubleshoot it.

>>1200464
I used a board from eBay with a cyclone IV on it, has ZRTech and designed by WXEDA written on it, but you'll probably find it sold by some other vendor. Best bang for buck imo, has loads of breakout points too, which is most important as no one cares about the peripherals on the board once you are ready to try using it for a 'real' project.

>>1200521
I remember getting a low cost cyclone IV fpga board on ebay too, but it was kind of shitty (waveshare I believe). The biggest problem was with the usb blaster jtag sold with it, it would fail programming the fpga very often and altera's integrated logic analyzer (singaltap) was impossible to use.

I have a standalone 10 key pad that I would like to transform into a calculator with a display. Any ideas on how I can start doing this? I kind of want to recreate this:

https://www.amazon.com/Ducky-Pocket-Mechanical-Keyboard-Number/dp/B071XX3KWP/ref=sr_1_1?s=electronics&ie=UTF8&qid=1498601632&sr=1-1&keywords=Ducky+Pocket+RGB+Mechanical+Keyboard+Number+Pad+%28Brown+Cherry+MX%29
but maybe with a larger display. Any ideas on where I can start?

File: trap.jpg (34KB, 501x471px) Image search: [Google]

34KB, 501x471px

anyone else want /ohm/ to be its own board? id love it.

>>1200597
Well this thread is relatively slow, it doesn't need to have an entire board dedicated to it.

>>1200535
Yeah the usb blaster is terrible, but for £20 include a pcb with tons of breakout points I could put up with it. The more 'official' stuff charge 10x that for something 10x smaller.

File: 1200px-BNC_connector_50_ohm_male.jpg (58KB, 1200x973px) Image search: [Google]

58KB, 1200x973px

Just getting into transmission line basics at university, however my course doesnt have a lab and I wanted to test out coaxial load matching and play around with reflections.

What I have been looking for is some sort of termination load that has a variable impedance, but I cant find what this would be called or any schematics. I'm assuming this would just be a variable resistor in series with a variable cap and inductor in parallel to adjust the imaginary impedance. Am I mistaken with this assumption?

Tested a surge protector with my multimeter.

>>1200602
that's true. it would be a very slow board. personally i only come on /diy/ for the electronic threads really.

>>1200650
That'd work. But they won't be ideal, so the variable resistor will only go to to 10% of max R, for example.

>>1200602
On this note, what is the slowest board on 4chan?

>>1200656

You mean the tolerance of the resistor at its worst would limit it to 10% under the desired value?

I mean you can't have a variable resistor go to 0Ω because of how they physically work. There's still some amount of resistive material that mechanically can't be bypassed. There's also the parasitics.

>>1200687

Maybe a switch to short it out?

>>1200692
That would help. You'll still have parasitics. You'll never have a perfect component. You can get close enough for meaningful experiments despite this.

>>1200694

Yeah that's fine I just wanted to play around with different loads and try out the theory for myself, thanks.

>>1200602
i would love for /ohm/ to be an active board that attracted more users, but most technical discussion websites need to be forums so you can look at relevant questions from 3 years ago.

Should /diy/'s own superhero be "Kapton Tape"?

Here's my dumb led question.

I have about half a dozen leds I want to light up just with a watch battery or whatever. I want uniform brightness, so I should connect them in parallel? And what is the best way of connecting >6 leads, I can't get solder to take to it. Just twist 'em up and tape it?

>>1200725

putting LEDs in series ensures they all get the same current, so they have equal brightness. but that means you need a pretty high voltage, like 14-18V (depends on the color of the LEDs)

for parallel, you can use a lower voltage, like 3-5V but you need individual resistors on each LED to equalize the current thru each one. google for an online LED calculator. but note that these tend to be completely retarded when the LED voltage and the supply voltage are close. remember to use worst-case conditions, not typical values, with battery voltage at its highest point possible, and LED forward voltage at its lowest rated value.

>>1200734
Why can't I just use one resistor? Wouldn't that limit the current for the whole parallel circuit? Or would that only be the case in a perfect world where every led is identical?
>>1200735
I'm worried about burning out the cheap little fuckers with too much heat, I already have to a couple. Though I guess if it isn't hot enough to melt the solder there should be no problem

File: daisy chain.jpg (22KB, 400x400px) Image search: [Google]

22KB, 400x400px

>>1200734
>I can't get solder to take to it. Just twist 'em up and tape it?

no nigger, you just need more heat. use a match or candle or whatever.

or daisy chain your connections

>>1200738
> Or would that only be the case in a perfect world where every led is identical?

correct. variances are always present, even within the same batch of components.

>>1200738
There's some variation in forward voltage between LEDs, single resistor with ones that aren't well matched will result in much more noticeable differences in brightness than if they all had their own resistor.

Can also do all one series chain, but then you'd need a higher voltage source.

>>1200725
Better be using plenty of heat, and ~60/40 tin/lead flux core solder.

So you'll probably want to hook them up in parallel, but if your maximum and minimum source voltage is greater than a multiple of the LED's forward voltage, then you can hook up three parallel lots of two series LEDs, or even the other way around. The lower the voltage you drop on the series resistor the better.

>>1200742
But to tune the resistors finely enough to correct any difference in LED brightness you'd need either very exact component values, or some tiny pots, which I assume will take up too much room. Unless you want to try your hand at different lengths of resistive wire, which sounds like more trouble than it's worth.

>>1200742
>>1200744
>>1200757
Surely the difference in current won't be that great that it will be a largely noticeable difference though? This is going to be a set piece for tabletop games so if I can simplify things by skipping the resistors that's fine. I just don't want a sequence of lights that get progressively dimmer, a mixed bag is ok if they're all fairly close

>>1200759
You can't skip all the resistors or your LEDs will blow up. If you don't care that much about perfectly equal currents you can use one resistor for a bank of parallel LEDs. Rule of thumb is 20% of the voltage drop should be across the resistor. So pick a current for the LEDs from the datasheet and crunch the numbers.

>>1200759
If you're doing low power and all the same model LED parallel pack with single resistor is generally okay.

>>1200759
It shouldn't be too noticeable, if at all. I'd just try it with a USB lead cut in half and a handful of resistors. If you don't have a handful of resistors lying about then you should.

>>1200287
I'll probably cover the bike in EL wire too, so people can see me from all angles. Also their perplexed stares will be directed at the bike, and not at the autism of the guy hunched over on its seat with his coat billowing behind him like a sandy hook shooter.

Can someone recommend a starter kit not to expensive for an 11 year old? Say max 20 euro/ 20 dollar. I'm don't know electronics myself so don't know what is important or not.

I have an EE degree but I've been a neet for over a year. How do I get a job?

>>1200876
You're probably more likely to get better value for money by buying value packs off AliExpress for resistors, capacitors, etc. but it depends on what sort of starter kit you want. Do you plan on giving your 11 year-old a soldering iron? Because that's pretty much $20 baseline if you want something that will last. If you just want the kid to go for breadboards, then I can't imagine him having too much fun. If you do get a decent iron it will last him well, so think of it as an investment if it makes you feel better, or buy it for yourself so the kid's nervous about breaking it.

The problem with electronics is that you need to buy things like perfboard (which looks like shit and is a bit of a pain to solder with) if you're going down the non-reusable route, or a breadboard (which can be a bit delicate) if you're going down the reusable route. Just using free-floating wires is also an option, though it requires a soldering iron. On the other hand you could go for one of those plug-together EZ electronics kits, which will probably be 30x the component price and 1/10th the versatility, but it will certainly be a hassle free way to get the kid started. I'd personally go the soldering-iron route, though it would require constant supervision, and this being the supervision of someone who knows what they're doing helps greatly.

If you were willing you could make some breadboard modules by soldering LEDs to their own dropper resistors and then to robust header pins, then make similar things with switches and other components, but that really is going the extra mile.

Note that even I, a (mostly) grown man have trouble pulling some components out of breadboards.

What made you pick electronics?

>>1200890
And if you end up going into the deeper realm of electronics, there is a bit of physics theory that's more suited to ages 13+, not to mention the need for a multimeter (actually like $5 from AliExpress will probably be fine for his first one).

>>1200890
8DIP!! LMAO 8DIP!!!1!!

>>1200890
I saw some kits with breadboard. I like that option especially when someone has to learn everything about electronics. You can do a lot of projects that way. Also not in favor of the education kits which don't use regular components. Maybe I should just copy a kit, buy those components and buy a beginner book? I would buy that kit but the booklet is only in English and shipping is expensive.

This is the list of components of such a beginner kit:

400 point Solderless Breadboard
11 colours of Hookup wire (0.45m of each)
4x 1n4001 diodes
9v Battery Clip
80 resistors (16 values, 5 of each)
50 Ceramic Capacitors (10 values, 5 of each)
10 Electrolytic Capacitors (7 values)
4 Potentiometer Trimmers (Variable Resistors)
3 RED LED's
3 GREEN LED's
3 YELLOW LED's
4 NPN Transistors
1 555 Timer IC
Instruction booklet with your first 5 circuits

File: tmp_31022-14760449909671078812225.jpg (61KB, 800x634px) Image search: [Google]

61KB, 800x634px

>graduate with an ee degree last year
>apply to 500+ jobs (design, test, qa, etc)
>10 failed interviews (no offers), ~50 straight rejections (no interview), rest just ignore
>networking has been useless for me. Didn't do coop in school
Getting really fucking desperate here. No one wants to give you a chance with 0 work experience, and I've even applied outside of my country.

I started applying to electronic technician positions and I'm getting interviews left and right now. Is it possible starting out as a technician with an ee background, then move up into an engineering role? Or am I just shooting myself in the foot and locking myself in a dead end technician role with no chance of moving up?

>>1201084
if it's been a year it might be a push, but consider asking your uni's careers department. Some unis spend a load of effort helping students find employment because that makes their degrees look like they are worth getting. Otherwise try and do something that sets you apart. What have you been doing during this year? Anything related to electronics? Being able to bring along a bit of hardware or something you can talk about is often a boon at interviews. If you haven't done stuff, why not? Get programming apps, games, design some pcbs etc etc to show that you've got your head in the game and can slot into the company with minimal effort.
This advice comes, of course, on top of general interview techniques - be clean, shake hands confidently, maintain a noncreepy level of eye contact and wear a well fitting suit - even if no one else is. Far better to stand out as someone who dressed smartly than someone who came as a slob.

>>1201104
My uni has been absolutely useless in terms of finding a job. All they do is stroke your ego and tell you you're good enough only to tell you what job boards to look at when push comes to shove.

As far as projects have gone, I have tried doing some personal projects at home and have brought things in to interviews. Its kinda the reason I've been lurking /ohm/ for the past while now. Interviewers seem interested when I bring in physical hardware but they look completely disinterested with any personal projects I mention I'm working on at home but don't bring in. Either way, I'm losing hope on how effective personal projects are at getting a job since everyone turns me down with "we're looking for someone with more experience", yet every employer tells me personal projects boost your chances of getting hired.

Interview techniques I pretty much got down. Clean cut, suit and tie, confident composure, etc. At least, what I think I'm doing is ok and others that know me say the same.

Either way I look at it, I'm kinda reassessing here because whatever I'm doing doesn't seem to be working.

>>1201131
Mine had a "job finding programme" in cooperation with a local work agency. They promised to find jobs appropriate to your field of study, so of course I paid for an appointment. My degree in mechanical engineering was only good enough to get me a part-time job on an assembly line with them.

What you should do is go to your interviews as you are, and if they turn you down, ask them why. Showing that little extra bit of attention will make them think you were serious about working with them in the future, and it might make them actually put your email in the list of candidates for the next time they need someone. And if it turns out you were doing something wrong, well, now you know. And you might have to simply acknowledge that you won't be getting your dream job right out of school. Your CV, even with personal projects, is going to look poor compared to someone who's been in the business for two decades. Find a business that does have your dream job, and that you want to work for. Get a job lower in the chain, and prove your abilities. Many companies recruit exclusively internally for these sort of positions.

Would it be worth modifying your LED lights so that half of them have inductive droppers instead of capacitive droppers to keep your power factor good? Would it be worth attaching capacitors in series with your washing machine motor to make the power factor better? What is the power factor of an ideal or a non-ideal transformer?

>>1201137
Ah, see I know my uncle who coincidentally is also a mech engineer and had a similar program in his school, he managed to get a job straight out of school as a mech engineer. But my school, as far as I know, doesn't have such a program, best they offer is a coop program and they tell you to piss off if you ask for help after you graduated.

I have been talking to employers once they turn me down, but the answer I get is that I lack experience almost always. Otherwise they don't bother replying to me at all and just give me the cold shoulder. A few have mentioned personal projects help, but other than that they've been relatively useless. Give me false hope that they'll keep my resume on file, and despite trying to stay in contact with a select few, they never reciprocate or dodge my questions. It's like having no work experience is equivalent to having leprosy to these people. I get I won't get my dream job straight out of school, and I'm willing to work lower, which is why I'm asking if its worth it for me to work as a technician. At the same time I've heard working as a technician doesn't help you get an engineering job and is more just useful if you are just desperate for money (and I can survive on daddy's money for only so long). I just don't really know what else I can do at this point.

File: inspiring image.jpg (1MB, 2048x1536px) Image search: [Google]

1MB, 2048x1536px

Hey /ohm/ I don't come here often, looking to rebuild a section of my motorcycles electrical system, I bought all new connectors [the good ones with the rubber gaskets on the insides to keep water and dirt out.]

Where the hell do I buy color coded electrical wiring!? Like black dotted striped, orange with white stripe, green and black?..

I just don't know what to google really.

>picture unrelated

>>1201228
Buying small quantities of different striped wire colours gets real expensive, especially for less common things like dotted stripes, the cheaper option is solid colours + coloured/numbered/labelled heatshrink for identification at each end.

>>1201201
inductive dropper LEDs bulbs practically don't exist on the market, since they're more expensive than traditional capacitive dropper and the absolute power factor still wouldn't be much better. Having a mix of capacitive + inductive droppers still wouldn't be very good in terms of power factors since the spikey current of the capacitive + the squareish of the inductive still doesn't add to a nice sine, there's lots of harmonics in there.
For power factor correction of a typical induction motor you'd put the caps in parallel, not series. Fancy washing machines now are often brushless motors, in which case the power factor correction would be an inductor or controlled boost converter on the input.

An ideal transformer itself has an undefined power factor, there's no R L or C elements anywhere. Ideal transformer with load reflects the load's power factor.

A non ideal transformer will be mostly inductive (magnetising current) with a bit of resistive (core/copper losses). Under full rated real load it'll look mostly restive since the reflected real load is much larger than the magnetising current.

>>1200650
Generally labs have a variable length line rather than a variable load. I think that would be difficult to build.

>>1200759
Try it and see. In my experience the differences can be substantial.

>>1201084
I've started being test maintenance technician (lots of manual work), after a year got promoted to team leader, another year later found new job designing products.

During years at first company I've done some personal projects, showed during interview in another company, got a job. Networking might help but you need work experience as well. Technician seems like good place to get since they usually require no prior work experience (depending on field though)

I took apart an old microwave and got out the display.
How do I light up part of the numbers? I tired 12v DC but it won't light up in any simple thing between two pins.
The display has the wiring diagram on the back bit I can't interprete it. Also it is vacuum sealed in glass kinda like a nixie tube
I just want to apply voltage and get display for fun, help please!

>>1200876
11 year old? I'm just getting started with this and I'm 17. Did I start too late?

File: Seagate-HDD.jpg (70KB, 640x503px) Image search: [Google]

70KB, 640x503px

>>1201286
Not that anon but what sort of personal projects would help get an employers attention? I've been working on a harddrive clock and put it on my resume but employers don't seem to care, even if I bring it up in an interview.

https://www.youtube.com/watch?v=XK9en2H7rBA

Should I be spending my time on something else? I thought about maybe making an automated fish tank regulator or something but I can't really think of much else.

>>1201333
Maybe call it a "POV clock" instead. Until I watched your video I thought you meant the etsy thing of sticking clock mechanisms in random objects.

>>1201336
That's my bad. Verbatim quote from my resume:

Hard Drive POV Clock - Personal Project
2017 - Present
Constructing an analog clock by cutting a single slit into a hard drive disk and flashing LEDs underneath the platter while it spins to create the illusion of moving clock hands. (Work in progress)

Should I take out the work in progress thing maybe? Explain that I'm building the motor controller from scratch using n-fets?

>>1201296
Sounds like a VFD. You need to power the filament.

>>1201333
>>1201338
Make it sound as sophisticated as you can. Mention exact technologies. Is it AVR? PIC? Closed-loop motor control?

>>1201333

At first I was very impressed until the video ended and then it seems that there are lots of similar projects. Did you just follow a recipe or did you do substantial development work on your own? If you are looking for a job as a line technician then the ability to follow instructions is good, but if you want to design things it helps to show that you can create new circuits on your own.

The video was well made, but I couldn't stand the music. Maybe something more vanilla that doesn't distract from the product might be more marketable; maybe some voiceover that shows your ability to convey technical information.

Overall, that's a very impressive design. If you largely created the design in terms of motor control and programming the LEDs then you are a pretty damn good designer.

>>1198652
If you use time-division multiplexing, you can use a single audio channel. If you assume the audio chip can do as good as 10khz for microphone, you could split that into 500hz segments, and assuming a full 500hz off in-between signal channels, gives you 10 channels. Use one of those 500hz as a sync channel and you get nine addressed channels. Considering most brainwaves of interest are less than 40hz, you could get a lot more out of 200hz channels (don't forget the nyquist theorem)

Not exactly an EE question, but I'm an EE so fuck it. Anyone have experience with backing-up a bunch of machines? The company is used to FreeNAS+rsync, so I looked into the bacula-sd plugin and it was absolute trash. There's probably more complicated solutions (we're just looking at something which can initiate transfers and store them efficiently), and there's lots of stuff meant for like... mixed environments with shittons of machines that usually requires a specific OS config, but nothing really drag-n-drop. I think there's ways I can install limited software (the "jail" mechanism), but otherwise I'm just going to tell the leader guy we should stick with a double-sided rsync which asks the machines to push.

>>1200663
/3/ is probably the slowest

File: 1484338360626.jpg (56KB, 645x773px) Image search: [Google]

56KB, 645x773px

>>1201346
Ok, long post. I'm going to be using an 8051 microcontroller to run both the LEDs and the motor. I intend to have a closed loop feedback in that the μc will determine the motor position based on voltage difference and thus the μc will know what coil to energize at what time (this is the shit that's been driving me up the walls because I can't get it to work). The motor doesn't run for more than 15 seconds when I plug it in with the original harddrive's circuit so I'm making my own DC to 3-ph AC motor driver using a set of 6 n-fets and the μc to make the synthetic AC signal via PWM. The way I intend to get the "hand" position is by measuring the rpm (either through the motor controller circuit or by a separate hall effect sensor that would detect a small magnet glued under the platter), calculating how fast 1/60th of a rotation takes then just making a truth table in increments of 1/60th. Run some leds from the μC then get them to flash based on that truth table and running clock program. Once I have it all worked out and running, I'm going to design it all in Altium and get a PCB printed out through some supplier so that it will all neatly tuck in where the old HDD circuit used to be.

I only have so much space on my resume to fit this all in.

>>1201353
Lel this isn't my video. I got the inspiration to make my own from this guy's video, and since I don't have a working one myself (yet), I figured it was the best way to explain to anon what I'm doing. I know there are several iterations of this very project online, the way I planned to make this an "engineering" project was by doing it on my own as I described above. But I don't know if this is actually making any employer's neurons fire and if I should scrap it for greener pastures.

>>1201365
But if I think about it, the only thing that really sets me apart from the projects you can find online:
>http://www.ian.org/HD-Clock/
is that I'm making my own circuit and feedback loop to drive the motor, and them making my own PCB. And I guess everything else pretty much can be assumed as copied by an employer since the code and whatnot is all out there online. Which is why I'm having my doubts about bothering to finish this shit.

>>1201369

It is not a waste of time because you are learning some important concepts and design techniques.

I'm not very familiar with the 8051 except that it's pretty ancient, which is ok because it is still used a lot. I'd use an AVR but that's because I have years of experience with them. Someone else will say PIC, and then the ARM crowd will laugh and tell us to go with that. If it works, it works.

Learning how to drive the motor and know its position/speed will be fun. I'm not sure that gluing a magnet to the platter is a good idea; you might need two for balance, and you might not need any once you are controlling the motor, as you mentioned.

I once read that some employers preferred that hacker guy who was playing around with real stuff in his basement to that 4.0 recent grad who didn't have any real experience.

Another great thing about projects like this is that once you get things working, it's quite possible that you no longer want to make a clock but want to do some other cool thing. It's common to be doing dev work and notice something you had not anticipated and then you go off on a tangent and come up with something completely different.

>>1201372
It's been super discouraging, I've been on and off for the past 4 months purely trying to get this motor to run, but for the life of me I can't get it to work. I've been fiddling with other hobbies during my off time (which admittedly is much less stressful but nothing to do with electronics).

I picked the 8051 because it's what we worked with in school. I could make a switch over to another μc, but for now I can't be bothered to stress with coding on a different μc. I wouldn't mind using another one just to get the "experience", but for now I don't think its worth it until I can get something to work.

Trying to get this motor to run has been anything but fun my dude, I've been tempted to just buy a ready made motor controller and just plug that in. Good point, I guess having 1 magnet would cause balance issues, but I dunno how significant it would be depending how heavy it is. I mean, the other option would be to have an infrared sensor determine rpm via the slit. Again, this is if I cannot determine the rpm via the motor control circuit alone. I might need a way to track a "reference" point though to calibrate it anyway since otherwise the clock positioning will be all out of whack.

I dunno man, I feel like all this project has made me do is try find another project as I've been progressively losing interest in it.
>But like dude it's a clock... a rainbow clock! lol!
The idea just doesn't seem novel to me anymore.

>>1201376

I'm not a motor control expert; maybe one will show up.

But this is my feeling: Do you have an oscilloscope? I assume you do. If not, get one now.

Start simple. Apply DC to one phase. Does the motor lock in to one position? If so, you know you are powering a winding.

Remove that and apply to another winding. Does the disk jump a bit? You can use this to determine how many poles you have, btw.

You can drive a 3 phase DC motor with square waves, open-loop. You need to start slow and ramp the speed up. I don't know how they drive HD motors, or whether they are 3 phase or whatever. But if you have a scope and know how to play with pulsing the phases you can get it to move.

But a better method is closed-loop, which I assume is how they do it. But this either requires some sort of encoder that either matches the poles or is a higher resolution but integral multiple of the poles. If it's not an integral multiple you just have to do more math. With closed loop you can then easily control the speed from zero to max rpm. The open-loop thing I described above might be hit-or-miss.

And then finally get away from square waves and use PWM to simulate a sine wave for each phase. This should give you much smoother rotation; the square wave approach might not even accelerate properly since the rotor will try to jump to the next null much like a stepper motor, and if your ramp is too fast or too slow you can have stability issues.

As I said, I'm not a motor guy, so this could be a pile of nonsense, but if I had time, a scope, a motor, and drive electronics I think it would be fun to experiment.

File: BLDCControl.gif (39KB, 720x386px) Image search: [Google]

39KB, 720x386px

>>1201379
I do have a scope, and I've done those exact tests. I have 2 HDDs I'm messing with; one has 3 terminals, the other has 4 (this one has a neutral terminal if I recall correctly). I'm basically swapping between the two until one starts chooching. I haven't played with a closed loop system yet because I know that's a whole other headache to deal with so for now I'm just trying to get the motor to run by just spitting out fixed speeds. Closed loop is how you're "supposed" to do it because it's more efficient, but honestly I can get away with open loop (even though I said I wanted to do closed loop in my description). Measure the back emf and by comparing back emf to the neutral line you determine which coil to energize next (its been I while, I think that's how its done).

The sine wave simulation is, more or less, being done by firing square wave signals at the 6n-fet inverter circuit (pic very related). This has also been giving me headaches because things shortcircuit, some don't turn on, I get retarded outputs, it's basically always a new problem with that circuit.

File: 1496862528621.gif (2MB, 500x500px) Image search: [Google]

2MB, 500x500px

What's a good cheap fume extractor so I can avoid breathing in solder fumes? would prefer not to die of lead poisoning in my 50s thx

>>1201384
>Measure the back emf and by comparing back emf to the neutral line you determine which coil to energize next

That works, but it's a pain in the ass, and you of course know by now that you cannot use that while starting up. If the HDD motor has halls or some sort of optical encoder that would be infinitely better, but it might not be how they are driven.

As for short circuits, that sounds like the high and low transistors are on at the same time (like 1 and 4), which should never happen if your logic and circuit is correct.

So, try to get open loop working. You're basically using it as a three phase stepper motor in this case, which is fine. Most of what you will read about stepper motors might discuss two phase bi-polar or four phase unipolar. If one of your motors has a neutral, you might be able to drive it as a three phase unipolar, but three phase bipolar would be better and would use your existing circuitry.

I'd set up my system so that I could easily change the drive rate, and start at a very low speed and look at the disk. If it's just jumping back and forth, raise or lower the rate and at some point the disk will rotate if you are powering the windings correctly, which can easily be determined by applying DC manually to the phases until you figure out a sequence that works. Once you achieve rotation, experiment with ramp rates and you will eventually be able to start and accelerate and decelerate under control without any feedback at all.

>>1201253
So nobody sells like maybe 3 meter lengths of colored wiring?.. What the fugg how do people build custom harnesses then?! They always look to be color coded to me. There is nothing at all within a 100 dollar range?

>>1201257
Oh of course, LEDs are not ohmic loads, so the current will only be significant at the peak. I guess you could do some Fourier magic to switch a bunch of sinusoidal harmonics voltages to keep your power factor good if you really had to.

If you had a private direct AC (wind?) generator powering your household, what kind of power factor problems would arise if you used too many inductive or capacitive loads, or high frequency switching loads? What about with a DC (photovoltaic?) generator powering your house through an inverter?

>>1201396

Maybe someone knows of a source; if you are willing to pay $100 you can probably find one.

An alternative is to buy a trailer harness which is a ribbon cable with 4 or more different colored wires that are easy to separate. Then buy 3 or four spools of other colors at an auto parts store, and you have 7+ colors to work with, and with some planning you might be able to have a good layout that's easy to trace.

Another crazy approach would be to buy a replacement harness for a car that will have one or two dozen colors like you need, and will probably have long enough runs for your bike.

>>1201392
I don't think this has any halls, nor an optical encoder. I mean, I don't think it does. I did a resistance test on the 4 terminal HDD motor comparing 2 terminals together at a time, and I found that they all have equal resistances except for 1 terminal that when paired with any other terminal yields half the resistance, which leads me to believe its a 3 winding motor with a neutral line.

Yeah, I figured it had to do with how I'm turning on the fets. Don't even recall what the problem was exactly, but I somehow concluded my μc blew and that's when I called it quits. It's been a while now, I don't even know where I am on that circuit anymore.

I see what you're saying by running it as a 3phase unipolar, I can give it a shot since I've been skeptical about running it as a bi-polar anyway, could I get away with just plucking out 3 of the fets? I should set up a way to vary the drive rate, I guess I could hook up a potentiometer to the μc and vary it that way.

>>1201404
>could I get away with just plucking out 3 of the fets?

For unipolar you would keep 1 5 and 3 and ground the center tap of the HDD with 4 leads. Note that you have to add three flyback diodes across the coils now. Put one on each coil with the anode on the center tap (grounded) and the cathode on the other wire for that phase.

Unipolar is fine for what is virtually a no-load system like yours; once the inertia is spun up there is little load on the motor. If you were actually driving a frictional load of any sort the bipolar mode is much better.

File: filter_solder.jpg (27KB, 550x368px) Image search: [Google]

27KB, 550x368px

>>1201390
>good cheap fume extractor
About EUR 30 for the machine (1m3/min) and about EUR 3 for two replacement filters.

For industrial use Sn60Pb40 has been phased out here (EU) more than a decade ago because it's not RoHS compliant. Replacement for hobbyists is Sn95Sb5 or Sn95Ag4Cu1. Rumor has it that it erodes the tip of a (traditional) soldering iron and requires a higher temperature than leaded solder wire (adieu Weller Magnastat..).

>avoid breathing
I'm using the still available SN60Pb38Cu2 (mostly 0.7mm) and slowly exhale while soldering, as a zen exercise.. If you want lead, turn to tab water. Flint is everywhere.

File: WhatsApp Image 2017-06-27 at 12.18.10 (1).jpg (349KB, 964x1130px) Image search: [Google]

349KB, 964x1130px

So i have 2 identical boards, one works one doesn't. They from 2 separate DJ controllers (midi). They handle the input power and USB connection.
One of the boards is "not recognized" in windows, i tried switching them, its one of the boards fault.

My question is how do i fix the broken one. I visually inspected every component and cant see any defects. what could be at fault here and what else should i check?

File: WhatsApp Image 2017-06-27 at 12.18.10.jpg (460KB, 1320x1012px) Image search: [Google]

460KB, 1320x1012px

>>1201429

>>1201429
Discard the broken one and buy a replacement. It will be a lot cheaper than buying the test gear you'd need to identify the fault.

>>1201459
sadly a replacemnt costs over 150usd, the controller itself is worth barely that.

File: bd617bd3-16f8-43e0-a4f9-de786e1bcde7.jpg (237KB, 1600x900px) Image search: [Google]

237KB, 1600x900px

Hey mates, what do you think of our robot

>>1201465
>Hey mates, what do you think of our robot
> nrf24l01 my nigga

File: stereo wiring.jpg (627KB, 1756x1836px) Image search: [Google]

627KB, 1756x1836px

hey /ohm/ers

Have an old car, with a factory harness in it, which I dont want to chop off. the new stereo im installing has the little prongs i can plug into the harness, but when i do they look like this picture. Besides cutting off the wires and using butt connectors, is there an easy way to make this connection electrically safe? And after market harnesses are super-hard to get for this vehicle.

Secondly, the horn is not working. I got a new horn, and installed it, but it died within 2 honks. Fuse is ok; also correct voltage has been tested at the connectors that plug into the horn.. Do you know what i should check to prevent killing another horn? (also, horn ground at contact plate is not getting anything with test light reading, no matter key pos, or horn installed).

According to schematic, horn has no relay...

How do I test if the output on my bench top power supply is accurate? Will just measuring at the output with a multi meter suffice?

>>1201513

Yes, if the meter is calibrated; in my experience handheld digital multimeters seem to be pretty accurate at measuring DC voltages, but maybe not so great at measuring resistance.

And then there's that old saying: a man with one voltmeter always knows what the voltage is, but a man with two voltmeters is never sure.

>>1201516
Har har. They also say that one measurement is no measurement.

>>1201429

repairing a broken gadget is way easier when you have a second identical working gadget, coz you can take a bunch of voltage and signal readings and compare the two to get a hint as to the problem area. that said, if you dont have a DMM or a scope, this is all irrelevant. first thing i'd do is test to see if that crystal (X200) is oscillating and is at the right frequency.

File: Screen Shot 2017-06-30 at 3.12.13 PM.png (187KB, 1326x992px) Image search: [Google]

187KB, 1326x992px

If I wanted to make my own LM709 op amp, would any old BC548/BC558 work? I also only have common resistance values, so 24k instead of 25k, would this make much of a difference?

I want to make a Kraakdoos, and it tells me I cannot use a modern replacement.

>>1201496
put a bit of heat shrink over the exposed crimp
once you have all of the connections made and tested as working, encase them in silicone caulk

is the voltage to the horn correct while the horn is still connected?
high resistance somewhere could show 12v without the horn but drop to ~0v when the horn is attached
if the voltage is correct and the horn is securely connected to the frame, you just got a bad horn - try again
there is nothing in your car that can 'kill' the horn assuming you're getting a proper replacement

>>1201516
>a man with one voltmeter always knows what the voltage is, but a man with two voltmeters is never sure
You didn't see the pic the anon with six or eight Fluke multimeters posted a few weeks back.
The meters were from several different decades where he just bought new ones 'because'.
He had them all connected to a single source and they all read exactly the same.

>>1201698
Probably won't give the same results, the oscillator effects the kraakdoos gets come mostly from parasitics that'd be quite different on your discrete transistor simulation.

Buying NOS LM709s is still pretty cheap and easy.

>>1201302
If you haven't built your first computer from discrete transistors by the time you're eight you'd might as well give up

>>1201700
>they all read exactly the same
With a little help from Fluke? Multimeters can be adjusted. Smells like commercial propaganda to me.

>>1201710
It's not at my local stores, but I see a few results on AliExpress. Since it's only 8 integrated circuits and 1 hijab , I can't imagine they're terribly common, unless they don't just go by "LM709".

>>1201488
also lattice FPGA and STM32F301 micro

>>1201465
What's that overly large capacitor?

does it matter which side of polarity you put components on in dc circuits? for example a resistor for an led, does it matter if the resistor goes on the cathode side or anode side?

>>1201765
In that case: no. I'm not too sure how generally you're asking this, though.

>>1201765

You should look at Kirchoff's circuit laws, or at least understand the concept of electric current. In the case of one or more resistors and one or more leds in series, they will all carry the same DC current at all times, so the order cannot have any effect.

What are some good tools both hardware and software to use for reverse engineering hardware/circuits?

>>1201780
Or any practices online? Like there are programming exercises/challenges online?

>>1201769
>Kirchoff
*Kirchhoff

>>1201780
>>1201781
What I do is hang around on /ohm/ until someone posts a mystery circuit, it can be pretty fun if you've got the spare time. Though this is coming from someone who does algebra when they get bored. Last time it was to figure out the equation of the waveform of a diode and a resistor across an AC power source. I ended up with my dependant variable added to its natural log inside of an inverse sine function, at which point I gave up. The time before was trying to figure out the standard deviation and mean of the Standard Luminosity Function, which was going smoothly until I realised I wanted the measurements to be in mcd after all.

I have no friends.

File: hmm..png (5KB, 372x364px) Image search: [Google]

5KB, 372x364px

>>1201809
Would this count as a mystery circuit? What could it be good for?

File: blocking_oscillator.gif (3KB, 410x264px) Image search: [Google]

3KB, 410x264px

>>1201888
Looks like a blocking oscillator.

File: hint.png (1KB, 258x152px) Image search: [Google]

1KB, 258x152px

>>1201903
Close, but it does not oscillate on its own, the 0.47µF negative feedback prevents that.

>>1201919
Second and last attempt: kicker circuit for clock/toy/etc. pendulum.

File: maneki neko.jpg (31KB, 600x409px) Image search: [Google]

31KB, 600x409px

>>1201936
BINGO!

I kind of want to make a mechanical ten-key pad calculator. There is one in the market but its overpriced and I don't like it. Is it possible to use the circuitry of a membrane calculator and solder in the mechanical keypads? Whats the best approach to take here?

>>1201962
Wire the switch ends to the sides of where the membrane makes contact, yea I don't see why not, granted you have some sort of mounting solution in mind.

>>1201400
>buy a replacement harness for a car
For what a car spare part would cost, he could just buy a replacement motor bike instead.

>>1201761
It's the power source.

File: 172372359.jpg (34KB, 612x442px) Image search: [Google]

34KB, 612x442px

I'm paranoid as fuck because a roommate has been sneaking into my room and stealing loose change and I'm too much of a pussy to confront them. I have a knob like picture.

I have locked my door but they use q-tip or some shit to open it. I don't want to be robbed or pay a fortune for a remote door lock that would also cause drama because it will look different.

Was thinking of waiting till I'm alone to bore a hidden compartment for a 9v battery in the door and putting an attiny85 and small solenoid over the unlocking pin inside the outside knob. Tiny momentary buttons soldered to the inside of the knob. The unlocking pin can't be simply pressed but to be pushed when the solenoid activates. And switching back to low power if I take to long or right after pushing the unlocking pin. This way I can still look like I'm just poking the unlocking pin when I unlock it with any implement, but they will get deterred by their perceived self stupidity.

It would also be nice if I could easily remove attiny85 to program a new combination if they catch on. If they do, I want to upgrade to a NFC transceiver so I can just have Tasker program a randomized combination when I wave my phone at the knob before I go out and enters it when detected again.

I need it to work in low power mode, wake on an interrupt from any of the buttons, then using the buttons to enter a specific combination of presses.

>>1202102
>stealing loose change
>they use q-tip or some shit to open it
so many schizos on 4chan

>>1201735
ebay/ali is fine

>>1202211
I guess so, but Ali's free shipping takes a month where I live. I've never bought an IC that's a can before, I wonder why they don't make them anymore?

>>1202213
was more expensive and harder to automate assembly than with plastic packages

File: Screenshot_20170626-043653.png (2MB, 1440x2560px) Image search: [Google]

2MB, 1440x2560px

>>1201988
Okay I figured out the answer to my question... For anyone to know from now on. The thing to search for color coded wire used in car and motorcycle electronics. Is GXL automotive wire.

-the guy who asked originally.

File: I seriously hope you guys don't scan this.png (2KB, 295x295px) Image search: [Google]

2KB, 295x295px

>>1202193
Alot of projection there.

Has anyone here used a PCB for a faceplate or other aesthetic purpose? I usually use aluminium plates I pull off an old heat sink, but they're too small for my current project, so I was considering setting the potentiometers and LEDs behind a PCB because I'lll have some left over from the project's PCB afterwards, and I don't think the copper look will be bad on the face of a wooden box. I'll try to set it in a dado so you can't see the fibreglass, which will be interesting with only a handsaw and rasp.

I could also etch lettering in it, but for that I'd need to convert almost the entire board with resist, or try to make some silicone walls that I can pour etchant inside. Do I have to use the specially made etch-resistant marker, or are there solutions closer to a paint? Liquid electrical tape might work, if it dissolves in a common solvent.

>>1202265

dont do it! it's impossible to make a PCB look good. you have tons of other artsy-craftsy options, like paint, stencils, stamps, color photos, resins, marbling kits, contact paper, etc.

File: dash_cam.jpg (8KB, 232x304px) Image search: [Google]

8KB, 232x304px

>>1202102
I would suggest to focus on silent, hidden, evidence-producing intrusion detection. A cheap ali dash cam is all you need.

>>1202295
How would it look any worse than copper sheet?

File: IMG_0807.jpg (2MB, 2448x3264px) Image search: [Google]

2MB, 2448x3264px

>>1202295
>>1202303
Or did you mean the etching part? I want it to look similar to pic related, where it says "SUCC" at the top, but inverted. I guess I'd need that press film to do anything accurate, but I was thinking maybe just a few tick marks for positions around a potentiometer. I've seen people covering their board in resist and scratching the unwanted resist away with a craft knife, which is probably as good an option as any.

Does anyone know of an IC for creating a buch converter with a wide input range? I have a 20-100V one at the moment but I really need 40v-200v input. I only need 1 amp and I don't mind having external gate drivers and fets but I can't find a simple controller.

File: maxresdefault[1].jpg (171KB, 1920x1080px) Image search: [Google]

171KB, 1920x1080px

>>1202265
Professionally made PCBs with soldermask and silkscreen can make pretty nice looking panels. Can do capacitive sense buttons or reverse mount things like LCDs for fancy. If you don't mind waiting a while getting stuff like that printed in china isn't too expensive.

>>1202323
For super wide input range going flyback would be easier since you can run the controller off the transformer modulated. See TI's calculator recommendations https://webench.ti.com/webench5/power/webench5.cgi?origin=ti_panel&lang_chosen=en_US&VinMin=40&VinMax=200&O1V=10&O1I=2.0&op_TA=30

>>1202323
If you want to be real lazy, any regular buck controller + regulate controller power with big zener + power resistor or high voltage linear reg

>>1202340
Flyback is just massive compared to what I'm using at the moment. I just want a synchronous buck converter controller that can run at 1mhz+ I will add a GaN half bridge and compact inductor. There's small 700V LDOs so starting it is easy enough.

I set up a PIC microcontroller to blink an LED. The code I'm using is this:

void main()
{
TRISB=0X00;
PORTB=0X00;
while(1)
{
int i = 0;
PORTB=0XFF;
for (i = 0 ; i < 10000000 ; i++)
{}
PORTB=0X00;
for (i = 0 ; i < 10000000 ; i++)
{}
}
}

I'd expect a 50% duty cycle and consistent timing, but the timing between blinks varies greatly, and the LED is never on for more than a tiny fraction of a second (~1-5% duty cycle). Any ideas why? I'm using the internal oscillator since I forgot to buy enough capacitors to stabilize a clock crystal.

>>1202434
Try disabling watchdog

>>1202434
10000000 is larger than can fit in an int, you want a long

>>1202438
Doesn't seem like it made much of a difference, aside from maybe slowing down the pulses. Why should the watchdog timer affect it?

I started getting into the software for these things two hours ago so sorry if this is a stupid question.

>>1202444
Timing is still very inconsistent but the duty cycle seems fixed.

I don't expect it's an issue with the internal oscillator being that inaccurate. Is there much room for variation in the amount of time it takes to count to 10 million?

>>1202421
Something like MIC2111B then?

>>1202447
If working properly you should be getting something like 0.02-0.1Hz frequency, not sure exactly without knowing your clock frequency, is it going that slow?

>>1202449
It's a PIC16F716 with a default 8MHz internal oscillator. I didn't set any frequency division bits, so I'd assume it's still running at 8MHz.

Sometimes it'll change states 4-5 times in a second, sometimes it'll be stable for 4 or 5 seconds. I'll try to put a video on youtube or make a gif or something.

>>1202449
>>1202452
https://www.youtube.com/watch?v=LpdBtG2S71A

>>1202452
Is that also with watchdog disabled?
got anything attached to the /MCLR pin?

>>1202456
Watchdog disabled, MCLR is at 5V = VDD from my USB ICSP

>>1201699

thanks for the info.

Super late, but may as well try..

Should i check the voltage by bridging the fuse panel with my mutlimeter, or just inline bridge from the connector to the horn?

Also, with the new horn i bought (actually used from junk yard), the horn has been unbolted from the frame when i tested it, i thought since it had 2 wires one was ground... Is the frame bolt acting as a ground, and did i possibly fry the horn by not having it bolted if thats the case?

Trying to fix an old console (PS). How bad would it be to solder to the pcb a wire from an old IDE cable for the 5.6 amps power supply input?

>>1202445
Watchdog timer is sort of a failure prevention tool. Once its counter overflows, it restarts MCU and program starts from new (software error, electrical malfunction etc causes it to loop). So these very short flashes could be result of this - pic powers up, initializes software, lights up LED and shorty afterwards watchdog overflows, restarts mcu and starts from beginning.

>>1202452
Have you properly defined clock frequency in definitions? Though I'm familiar with PICs, pretty sure it still applies

>>1202489
Yeah, I threw in
#pragma CLOCK_FREQ 8000000
based on some example code which included a header I don't have, system.h

Full source code at
https://pastebin.com/gLFWie2t

>>1202458
Interrupts via RB0 (6) disabled?

>>1202495
Not explicitly, but it's an output, right?

>>1202455
put a 100nf cap across the supply rails

>>1202499
...as close to the pins as possible

>>1202455
Also try a less powerful led.
Looking at the compiled machine code might reveal some compiler optimization shenanigans that might cause problems.

>>1202499
>>1202500
I'll do it when I get home in a couple hours. I noticed that my Vdd is only getting 4.3V and fluctuating quite a bit, with MCLEAR at ~4.6V, so that might be the culprit. I'll separate it from the ICSP and decouple.

>>1201681
how expensive could repair be from a shop?keeping in mind the average income is over 4 times lower here than US.
Should i even bother if i want to get it fixed under around 90usd?

>>1198606
I've used lab view in a robotics project. It's a decent program, but you have to have a very intimate knowledge of computer logic and knowing how literal object programming works. I liked it.

>>1202499
>>1202500
>>1202504
Added decoupling cap and put MCLR and Vdd right on the +5V rail instead of going through the ICSP. They're both sitting pretty at 5.05V. Doesn't seem to have helped the timing.

Video showing that it's much slower, even after lowering the count-to value by 10x, and still not periodic:
https://www.youtube.com/watch?v=QjdIOjBRUGg

>>1202595
Got it working. The PIC16F716 doesn't have an internal oscillator I can use for the clock. Stole a crystal and some caps off a different MCU and it works fine.

It was set up to use the oscillator pins as crystal inputs and outputs so letting them float around made it think it had clock cycles every so often.

>>1202511

it'll likely be very expensive coz a local repair shop wont have the specialized knowledge to actually repair the dead components, like the manufacturer can, so they'll just order a brand new board. so you essentially end up paying for a new unit PLUS labour costs.

the only exception to that is if it's a trivial, easy to spot, problem.

>>1202484

IDE cable wires are around 26 gauge and should not be used to carry 5 amps. for that you want 18 gauge. or you could use several strands of the IDE cable, say 4 of them in parallel.

>>1200327
That's so much BS. In the US all of the shit is certification AFTER you get your degree or some other qualifications (like ten years of work). Electrical Engineers actually get into all the shit we need today. For my BS I took photonics courses, because some stupid fucks thought a 50 person class for power stuff was going to be enough for 100+ EE students. I just set up a server for my company and am currently trying to replace some goddamn cisco hardware from 15 years ago with a linux solution for a fax or international calling company that serves tens of thousands of people. It's bullshit, but you need people with high learning ability and electrical knowledge to do it. If I wanted to go back to school, I'd be in an EE program for solar panels, lasers, even bullshit like flash memory, etc. An actual EE degree covers a lot of physics and practical math, electronics knowledge, only one mandatory power class, and a few more classes you can choose from - I chose math and photonics, most of the other people chose power or signal classes.

If you want power some places do offer Power Engineering degrees, or you could do Robotics or Mechatronics, but literally NO UNIVERSITY DEGREE, NOT FUCKING ONE IN THE ENTIRE COUNTRY, is an actual Engineering certificate. You could be a goddamn PhD in Power Distribution and it could still be illegal for you to wire a damn light fixture.

just fixed a 55 inch samsung tv I got off of craigslist for 20 bucks by changing a few capacitors. I had to replace a surface mounted capacitor with a tantalum. Why is it that electronic component stores always have every kind of resistor and capacitor except surface mounted? Do people just not buy those? I hate that the only place to get them is online.

>>1202689
Surface mounts are a fucking bitch to solder
t. regards from someone trying to mount 3 extra small I2C ADCs (and yeah extra small or something is part of the form factor - like how DIP actually stands for something)

>>1201390
Fan with or without a window if you don't like rosin fumes. Wow!

>>1201428
Do you know what kind of solder companies 'recently' started using? I have a USB iron that works really well on lead, pretty good on most solder I find, but some modular PSU from 5 years ago or so required a stupidly hot iron. Like even a 30W I had wasn't enough to do a good job, so I was using an "etching" iron from a crafts section (actually had the temp on it iirc, which is ironic considering it's meant for goddamn crafts)

As for eating away at irons, I've found it comes down to the tips and heat more than anything. Shitty tips with high heat will carve stupidly quick, but I've used others that lasted 10x of times longer without disintegrating. If you wipe off the tip (wet sponge or dollar store steel scrubber) and use an in-line dimmer to turn down the iron to a more suitable level for what you're working on, that alone makes the tip last longer. Related, since an eaten tip gets shorter, it'll be closer to the heating element, so you don't need to use as high heat. I've used irons where the tip gets eaten more than halfway down, but those were probably some cheap shit iron.

File: Untitled.png (21KB, 768x768px) Image search: [Google]

21KB, 768x768px

i couldnt find anything from a google search, so i would like to make an electronic to mechanical tach converter for a vintage motorcycle. its supposed to turn an electronic pulse from the ignition coil into a pwm signal to drive a little dc motor that spins the tach via a drive adapter rather than using the cable directly. of course it would have to be adjustable to tune it in and as responsive as possible. only thing is ive never tried something this complicated, ive only done basic projects so im not sure where to start. im thinking finding a simple dc motor controller could be easy, but then i wouldnt know how to make a thing to tell the controller what to do via the ignition pulse or how bad the response would be.
>just use an electronic tach
i want to keep it as stock as possible including the tach, if someone made an aftermarket one that looked the same then i would go for it.
>just use the regular cable drive
my bike was the base model that didnt have a tach and so no drive installed. finding the tach is easy but the parts to actually install the drive to the engine are harder to find.

>>1202448
>Something like MIC2111B then?
This has a 3.4V output limit and some pins connected to the high voltage side though.

>>1202714
Can't you just run the tachometer off the engine? Whether it be on the camshaft or wherever, Surely it will be better than trying to tune a PWM DC motor to bike-RPMs. A synchronous motor would actually get you a linear relationship between RMP and frequency, but again getting it up to the tens of thousands will be difficult and probably bulky.

I'd just disassemble your tachometer and replace its internals with a mechanical voltmeter, and then convert the input frequency into analogue voltage with a microcontroller. Even better if you can replace the internals with those from a tachometer that's meant to run off ignition pulses, but tuning it would be more difficult.

>>1202739
>getting it up to the tens of thousands will be difficult and probably bulky.

a mechanically driven tach does not have a cable spinning at the motor RPM. It would have a substantial gear reduction. And a "vintage" motorcycle might not be turning "tens of thousands" of RPM anyway.

But bike guy: if you are altering the bike in the first place, why not just figure out how to take the guts from an electronic tach and put them inside the mechanical tach. Then all you have to do is convert the frequency. the whole DC motor plan is not a good one, in my opinion.

>>1202628
DOH

>>1202748
Either way he'll have to mess around with the internals of his tachometer, unless he puts a gearbox on the electric motor to make it faster.

what is the best wire gauge to use with perfboards?

>>1202683
"In the US" it's almost like there are other countries other than the US.... my mind is blown....

As for me, I've had my BEng in "Electronic Engineering" (reread that first word a couple of times until it sinks in...) for two years and have been working towards a PhD. My professional accreditation (IET) lists me as.... oh.... "Electronic Engineer".

I don't know why you're getting your knickers in a twist over light fixtures. That's got close to nothing to do with engineer. Electricians are the ones who install stuff like that.

>>1202692
The RoHS directive came into effect in July of 2006. Because of the problems with higher temperatures and tin whiskers there are exceptions for medical, military and measuring devices. Private use of leaded solder wire is not regulated, only industrial production. I guess they use the alloys already mentioned.

What you describe is most likely due to heat transfer to the ground plane, not the solder. I had to replace all the caps surrounding the CPU on a main board. I pre-heated that section with a 50W halogen desk lamp at short distance for about half an hour and used my 'on-the-road' soldering iron that can be temporarily switched from 20W to 130W by pushing 'the red button'. This finally worked and I got the board running again.

>wet sponge
Yes, but not too wet. According to Weller, the thermal shock can degrade the coating of the tip (fine cracks because of different thermal properties of the copper core and the thin steel coating). If it hisses, the sponge is too wet.
>in-line dimmer
Thats the way to go. When the flux turns brown the tip is too hot. Normally I use a 24V/25W iron on a self-made adjustable power supply, usally set at half of the max power. It can be switched to 12V for my tiny 8W SMD iron which I rarely use because I have pencil-type tips for the other one. The 12V setting now serves as the stand-by power.

The good news is that one manufacturer (Stannol) confirmed that the traditional variants of leaded solder wire will probably be available 'forever' and you should never mix lead-free and leaded solder for repair work.

>>1202766
>Electricians are the ones who install stuff like that.
Exactly. Chances are if you're going to university to learn something, it's not going to be a field job.

>>1202774
When I bought a cheap supermarket sponge and cut it up to use as a soldering sponge, it turned brown and let off a smell even with plenty of water, so I bought copper wool instead. Now I've got a dozen pieces of sponge lying about.

I see some things say that tin solder is worse for your soldering iron tip, and others that say leaded solder is worse, and I'm not buying into either. I've got a 40W iron, so it should never get stupidly hot in the first place, and when the tip does finally go on me it's hardly expensive to buy another.

I have 5 volt power supply and a li-ion battery.
Is it ok to charge li-ion battery through 1N4007 diode and a resistor? Or better buy TP4056 modules?

File: received_10158996575820083.png (18KB, 623x596px) Image search: [Google]

18KB, 623x596px

So I'm hoping to make a little circuit to allow my 3-pin computer case fans to be be controlled by the 4-pin pwm signal.

Mainly because the noise of the 3-pin fans being permanently on 100% is driving me crazy.

A) would this work without borking my motherboard and/or fans

B) am i missing something in terms of controlling the fan speed without having project boards littering my case

>>1202763

24-gauge solid-core telephone or ethernet wire. easy to find large quantities in the dumpster wherever a new building is being erected.

>>1202795

you never charge a LiPo with a fixed source, unless it contains its own protection circuit, coz it will burn your house down.

>>1202813

buy a new 4-pin fan. recently made fans tend to be considerably quieter than older fans, whether 3 or 4 pin, even if both are brand new.
if the fan is especially old or especially cheap, then the lubrication will have dried out making it a much noisier than when new. dumping oil in it will fix it temporarily.
if you're worried about screwing your mobo, then you shouldnt be playing with it.

File: sponge.jpg (6KB, 300x225px) Image search: [Google]

6KB, 300x225px

>>1202790
>cheap supermarket sponge
Suitable sponges look like pic and expand when wet. Avoid finely pored plastic sponges, they melt.
>40W iron, so it should never get stupidly hot
40W is a lot without regulation. Do you get dark, burned flux resin around the solder joints? If so, try the in-line dimmer mentioned above. More is not always better, ' just enough' is the optimal temperature. You can 'feel' it.

Is there a way to make like a reverse VCO?

I basically want to take a signal that will very in frequency and convert it to a DC voltage proportional to the frequency (in my case it's the lower sideband of a doppler radar signal, the signal will be in the range of 1kHz to 8kHz roughly and will be about 2.5Vp square wave regardless of frequency and RF amplitude)

>>1202748
>a mechanically driven tach does not have a cable spinning at the motor RPM

I worked in the motor pool in the army.
Mechanical tachs for our vehicles had a cable connected to the end of the camshaft without further reduction.
The 2:1 reduction of the camshaft was all the gearing involved.
If the tach was connected directly to a shaft (electric motor, whatever), the tack would indicate double the actual shaft speed.

File: Stewart Warner 1932-1934 V8 Ford Tachometer.jpg (226KB, 600x467px) Image search: [Google]

226KB, 600x467px

>>1202748
>a mechanically driven tach does not have a cable spinning at the motor RPM

>>1202864
use an LM29x7. if you want cheap and generic you can also make a V to F with two or three schmitt inverters that will cover your band just fine.

>>1202874
cable gauges almost always have worm drives in the 1:5 to 1:10 range so that the cable and gauge internals dont have to spin at sanic speeds

File: figure 47.png (19KB, 423x290px) Image search: [Google]

19KB, 423x290px

>>1202864
Data sheet lm2917-n.pdf

>>1202877
>drives in the 1:5 to 1:10 range
1:2 >>1202869
1:1>>1202874

stop generalizing

>>1202766
>>1202790
You guys are pretending to be retarded right? I don't know where you come from, but in most places where people don't shit in the streets electricians are certified. Let that sink in, that in civilized countries people don't enjoy their house burning down, nor do they think it is responsible to give any retard who gets an Engineering degree the permission to sign off on a multi-billion dollar skyscraper or waste dumping.

Even a goddamn medical doctor needs a license to practice medicine, and lawyers too.

File: generalizing.png (452KB, 768x7500px) Image search: [Google]

452KB, 768x7500px

>>1202881
boy r u dum

Hey guys,
I just fixed the antenna of my old alarm clock and have seen that it has some easily accessible PCB.
I was wondering if I could add a microcontroller to output my own alarm sound? Actually just have to find a pin which outputs a signal and then just wire an esp8266 or shit up to the speaker.
Where would you start to find that alarm pin where I maybe could get a signal?

File: IMG_7318.jpg (644KB, 3000x2000px) Image search: [Google]

644KB, 3000x2000px

>>1202925
forgot picture

>>1202714
Buy an electronic tachometer and move the innards to your preferred casing.

A DIY tachometer isn't rocket science, but it's not trivial either (a frequency-to-voltage converter followed either by a galvanometer or a voltage-to-PWM converter and a servo).

File: IMG_7321.jpg (554KB, 3000x2000px) Image search: [Google]

554KB, 3000x2000px

>>1202927
Nice.
Just pulled out my multimeter and found this little transistor which apperently is only triggered when a alarm goes off! Puts out 8V. So that's more than enough.
Not just gotta figure out what I want to bring to the speaker

>>1202927
I'd assume there's some MCU on the clock that'll have some kind of enable pin that'll go high or low when the set time is reached. That probably drives a transistor that'll saturate and and allow the signal from the radio/amp to reach the speakers.

>>1202942
nvm I see you found that out for yourself. Just put your signal source in line with the collector for your transistor and you should be okay.

File: 1499015055464.jpg (1012KB, 3000x2000px) Image search: [Google]

1012KB, 3000x2000px

>>1202942
wait no, I think that's the wrong transistor. I believe that is either an RF amplifier or mixer/demodulator. It's hard to tell with pics but the IF cans and the inductor suggest that.

Try this transistor.

>>1202946
Thanks mate! I'll look for that one.
I still have to learn shittons of stuff on how to figure out which part of the PCB is doing what.

>>1202948
it doesn't help that the board is an unholy 80s clusterfuck

>>1202948
It might help you to remove the screws and actually take the board out. It looks like a single sided board so all the traces will be on the other side. That way you can trace it out and see if one of the pins on those ICs is driving that transistor's base or not.

File: IMG_7322.jpg (3MB, 4000x5333px) Image search: [Google]

3MB, 4000x5333px

>>1202954
yup I just did that.
If I am not completly mistaken the resistor 26 (right next to the big cap) is connected to the transistor. It also only has ~10V when the alarm is on, so that's seems not to be too bad.
Gotta look tomorrow a little bit closer, not enough time right now.

But thanks guys, this seems to be a cool little project and it doesnt even fuckin matter if I fuck that alarm clock up, so there is only going to be profit for me!

File: radarAnalogFrontEnd1.png (3MB, 4012x2379px) Image search: [Google]

3MB, 4012x2379px

>>1201809
Boy do I have a challenge for you then.

In April I reverse engineered the front end of a K-band radar gun to assist in my defense against a speeding ticket in Jew York. I was acquitted by the way. Anyway I still have very little idea how this works and I'm curious.

What I do know is there is a Gunn diode oscillator and cavity resonator that functions as the local oscillator which is at 24.15GHz. This is transmitted via a conical antenna. The signal strikes a moving target and is shifted in frequency slightly, the doppler effect. The return signal is then mixed with the LO via a single balanced Schottky diode mixer. The intermediate frequency range is above 1kHz and under 10kHz and which corresponds to speed readings between about 20mph and I think around 150mph. Anyway this is all before the board in the pic, a bunch of waveguides and stuff. Now this signal is fairly low amplitude coming into the board, it's so low it's hard to measure, probably microvolts. The output from this board to the counting unit (not pictured) is about 2Vp square wave at the intermediate frequency. I know there's amplification somewhere in here, can't tell if it's with the transistors or the op-amps or both. The circled part is the part in the schematic. Above it is just power circuitry to drive the Gunn diode from a cursory glance. That u723 is just a regulator.

Nevermind the fact that that output connector is missing connections. The first one is the only relevant one, the IF out. The second idk about, might have something to do with the switch between the moving and stationary modes for the radar gun, the LDR off state resistance is very high 10MΩ vs 55kΩ@1mA decreasing with increasing current. Middle is 12V in, fourth is 8Vish, and last is ground.

File: alarm detector.gif (6KB, 681x404px) Image search: [Google]

6KB, 681x404px

>>1202956

the simplest, most likely-to-succeed, least-likely-to-destroy-the-unit, easiest-to-restore-original-functionality method is to cut off the 2 wires going to the speaker, then rectify them with a diode and capacitor, and use it to turn on a transistor switch.

the second best method is to google the part number of the big chip underneath the ribbon cable. in the datasheet see which pin is the alarm output.

File: WhatsApp Image 2017-06-22 at 16.02.26.jpg (334KB, 900x1600px) Image search: [Google]

334KB, 900x1600px

>>1201761
>>1201990
Actual original poster here.
It is NOT the power source. It is a high voltage capacitor which is used to activate the kicker mechanism. The robots are supposed to play robot football, and to shoot the ball they have a high energy coil to kick the ball.
The three circuit boards you see here are in order from top to down:
- Main processor board containing FPGA, processor, IMU and communications
- Motorcontroller bridges
- Kicker high voltage board

pic related, some of the horrible soldering work one of the soldering volunteers did on the motorcontroller board.

>>1202987
jesus christ

why did you pay for enig?

File: IMG_4650.jpg (158KB, 1600x1200px) Image search: [Google]

158KB, 1600x1200px

I left my job to get into college because I was tired of doing meaningless stuff, just to find out that college was more of the same.
What I want to do is make stuff with my hands, just get started with anything.

I like: drones (have a hubsan lying around), planes and rockets in general (couldn't afford aerospace engineering, got into mech end instead), programming, automation (PID was hella cool).

I live in my grandma's, so there's a shitton of old shit in here. For example pic related. What could I do with that, any component I can salvage?

>>1202992
salvaged components aren't good for anything except power supplies or making the very basic circuits described in introductory books. dismantling things also won't teach you much without already knowing what to look for. it's more "did they implement this flyback in an interesting way?" than "i examined this board and now i understand flyback converters".

>>1202987
shouldn't you use an oven to solder that kind of stuff?

>>1203001
you can hand solder those easily, reflow is just better. it only really matters for fine pitch packages like tssop and obviously qfn

>>1202992

it looks like the remote control uses ultrasound instead of infra-red. search youtube for ''ultra sound projects''

another thing worth salvaging is the fly-back transformer.
Some Awesome Things You Can Do If You Have High Voltage : https://www.youtube.com/watch?v=asa-MGJQYBs

or this, Oscilloscope From an Old Television : https://www.youtube.com/watch?v=nAzQsl3gKMY

you can salvage the speaker if you're really desperate, and the deflection coils will given you several yards of magnet wire, which is infinitely useful. (or maybe not; i still havent finished the spool i bought in 1996)

>>1203007
>that reel of magnet wire everyone has
>that unfinished reception coil rotting in my parts drawer
CNC solenoid winder for hobbyists when

File: V-22_Humvee.jpg (52KB, 744x593px) Image search: [Google]

52KB, 744x593px

Ayyy /ohm/, what's a good project I could work on?
I'm an electrical engineer student and next year we'll be asked to work on a project for a lot of hours. By default some projects will be offered for us to pick but I want to do something fun and challenging. It would use programming, electronics and why not mechanical engineering.

I was thinking about doing pic related, a tricopter/plane hybrid but I figured the fine tuning would be too hard and I would never get it to fly properly?

What are you guys working on? Any idea of a cool project?
My other ideas included an app driven airboat, an equalizer, an electric skateboard and a Zeppelin

Thanks

>>1203313

A vest you wear that pounds your body with the Bass line so that when you listen to music wearing headphones it's as awesome as when your stereo is disturbing everyone within a mile.

I'm assuming you could get the effect without the sound being loud enough to disturb your neighbors. If I'm wrong about this, it pretty much defeats the purpose of wearing headphones.

>>1203316
Couldn't it just vibrate? I think a company had done a bracelet that does that. Thanks for the idea, I'm writing it down

>>1203318
>Couldn't it just vibrate? I think a company had done a bracelet that does that.

Sound is nothing but air vibrating, and if you're familiar with a chest "bracelet" that provides bass then yeah, it's already been done. It's such an obvious idea I should have googled first.

>>1203321
Well sure, depends of the frequency I guess. I didn't mean to be rude or anything btw, that's a good idea even if it exists and I have no problem doing something that already exists.

I don't ask you to find the next big thing and if you do find it you should probably not tell me

Is it bad to be pulling 34mA from a 74HC164 8-bit SIPO shift register? I want an LED for each bit, for which purpose I bought a bunch of 3mm red LEDs and an equivalent bunch of 820Ω resistors, which give me a current of 3.8mA. They're still surprisingly visible. So I'll be pulling 3.8mA out of 7 of the 8 outputs, and double that out of the other, and the datasheet only has "maximum current per output = 25mA" and "maximum total current out of all outputs = 50mA", but no recommended values. It states "±4-mA Output Drive at 5 V" in the features section, am I to assume that it's fine to run all outputs at this at the same time? I don't want to be smearing heat-sink compound on top of my PDIPs.

>>1202990
Hmm we got the PCBs from a sponsor, but I think for 2x 10 PCBs we 'paid' 1700 euros
However, their service was horrible. Next time we will use good old elecrow again.
>>1203001
>>1203002
You can use an oven, but you will need to get a good reflow profile for that. In my experience i have found it can take more time doing that, than actually soldering them by hand. Especially if you can easily find extra (free) hands to help.
Also, the boards are double sided loaded, so we would only be able to reflow one side with an oven.

>>1203350
>so we would only be able to reflow one side with an oven.
Just flip it into a bed of green sand for when you do your second oven run.

>>1203362
of course it is possible, but seeing the limited number of components, it is just as fast as hand soldering (at least if you are a bit skilled)

>>1203371
Just a suggestion. Do people do this at all for larger boards, or am I the first to think of it? I imagine the factories have some sort of custom-built heat-resistant form that the other side of the board sits into; they might even be able to solder both sides that way in the first place.

>>1203372
I think they generally either glue the parts in place with a tiny drop of epoxy, or use a higher temperature tin for the first side, then flip it over and used a lower melting point tin for the other.

>>1203339
It's fine. The stated maximum power dissipation is 500mW, which is 100mA at 5V.

The data sheet doesn't even state thermal resistance, which would be necessary if it was possible to exceed temperature limits without exceeding current limits.

>>1203339
>fine to run all outputs at this at the same time?
Quick guess: at Vcc=5V you can expect about 4.5V at the output with 4mA load. Red LEDs have about 2V, so a series resistor of 620Ω will get you the 4mA. Power dissipation of the chip: 0.5V*4mA=2mW per output, 16mW with all 8 LEDs ON, that's next to nothing. You can use a lower value if you want brighter LEDs, like 560Ω or even 470Ω.

>>1203350
the relatively tight tolerances on reflow profiles are to minimize the probability of failures when manufacturing thousands of boards. for a one-off a hot air gun dialed to whatever has always worked for me.

>>1203313
look into High Altitude Ballooning, it can be as simple as sending a balloon up with a phone in it to text a GPS coordinate when it lands, to including a live video downlink. Balloons require thinking about the electronics, as well as ensuring your electronics can survive a harsh environment (so thinking about materials etc) within tight limits (low mass, etc). Furthermore theres a whole bunch of regulations you need to comply with to legally fly your kit, which adds an extra dimension to the project which is often neglected by students, ie look at muh iot flashing light circuit, vs look at my iot flashing light circuit which I could start selling tomorrow with CE, FCC approval etc.

File: Screen Shot 2017-07-04 at 10.23.37 AM.png (520KB, 1548x520px) Image search: [Google]

520KB, 1548x520px

>>1203434
Current limits do anything other than being temperature limits? Even if it's for isolated heat sources within the chip, surely there's nothing else that over-current would cause, right?

>>1203465
Actually I'm using 820Ω resistors because they're the only ones my local store had at least 50 of. But they're only for small indicators, so it shouldn't matter.

Soon I will begin construction on my needlessly complicated variable frequency divider, wish me luck. Top are the 8 red LEDs, one for each register output, the yellow is for the output selected by the rotary switch, centre, and at the bottom will be a banana binding post, all mounted on copper PCB faceplate. The control panel will have a few more buttons and switches on it, not sure how it will be mounted relative to the main faceplate though. Red had better go better with the copper than blue or green.

File: dot-shift.jpg (23KB, 480x360px) Image search: [Google]

23KB, 480x360px

>>1203631
>wish me luck
I certainly will. That's a lot of soldering and a single 'cold' joint is one too much. Let it run slow at first to see that all works as expected.
The diagram has some interesting details, e.g. the initial clear (G3) because the state of the register is _undefined_ after power-up. It has a common 150Ω series resistor because only 1 LED is ON at a time. The first H is inserted by G4 and the recycled. G1/G2 is the buffered clock generator.

>>1203881
>the recycled
*then recycled

>>1203881
I'm gonna have a reset button that I press each time I change division, preferably a spammable one. Is that diagram what I should be doing? Because I'm not using nearly as many inverters.

>>1203631
>Current limits do anything other than being temperature limits?
You can overheat/melt/kill one output without necessarily overheating the whole chip.
Excessive output current can also trigger latch-up in some cases. Which then ultimately kills the chip via overheating, but it isn't something you prevent by improving cooling.

>>1203885
Its benefit is that you can "easily" get different output duty cycles. If you want 50% duty cycle, the one inverter version does better job. Also, the single inverter version works equally well at all clock speeds, while that contraption requires you to carefully match G4's delay to input clock and reset pulse length. That becomes increasingly difficult as the clock speed is increased.

>>1203881
Oh and I'm using a twisted ring counter because I think it gives me better division choice, though I may be wrong. In the end I plan on being able to switch between a ~4MHz crystal, a 32.768kHz crystal, and a button push.

>>1203912
Can't I just make a modular "duty cycle component" that I can plug into any of my outputs? Preferably one that works at any frequency of square wave. I'd just throw the waveform into an integrator, possibly with the help of a 2.5V rail, and feed that into a comparator with a potentiometer voltage, but the frequency effecting the amplitude of the triangle wave remains an issue. Is there a better way to generate a triangle wave?

I was thinking about creating a google glass knock-off, connecting a small screen to an rpi and positioning it to eye level. Thing is, I can't find a screen small enough anywhere. Has anyone attempted this before? Where would I find such a screen?

File: 23984763.png (108KB, 724x344px) Image search: [Google]

108KB, 724x344px

Why are DI boxes so expensive?

File: ouch.jpg (18KB, 459x261px) Image search: [Google]

18KB, 459x261px

>>1203885
No, merely educational. Like what's the clock frequency, what about the timing of G3 and G4, what is the diode good for? What if you want to shift two LEDs through the register? Mostly rtf datasheet stuff/calculation and basic understanding. This is a simple toy after all, not a high-end ring counter. Diagrams are like short stories - or crime novels occasionally.

>>1203916
>Can't I just make a modular "duty cycle component"
This can be done either in frequency-independent manner or in monostable-like manner where each output edge triggers a fixed length output pulse. The drawback of the first one is that it needs higher input clock and has limited (fixed) resolution for duty cycle. For example, if you want 10% duty cycle setting resolution, you'll need 10 times higher input clock. The latter requires adjusting the duty cycle whenever you change the frequency.

Both problems can be avoided by designing the oscillator itself so that it can produce the wanted duty cycle directly, but then you can't use post-dividers.

>Is there a better way to generate a triangle wave?
All options have their own drawbacks. The "better oscillator" option above would handle this, too.
If you insist on using a crystal oscillator, then there's the option of phase locking a triangle wave oscillator to it. It's bit tedious, though, and has limited range. Or you could use an up-down counter feeding a DAC to perform the conversion in frequency independent manner, but this requires high input clock to produce decent output.

File: dibox.jpg (7KB, 262x167px) Image search: [Google]

7KB, 262x167px

>>1203932
Niche market. DIY box is cheaper.

>>1203631
> Current limits do anything other than being temperature limits? Even if it's for isolated heat sources within the chip, surely there's nothing else that over-current would cause, right?
It can cause migration, but that's a long-term effect.

But for the most part, they're not "temperature limits" insofar as they're largely unaffected by the overall die temperature (i.e. ambient temperature, heatsinking or immersing in LN2 won't significantly affect the current limit), even if the failure mechanism is thermal.

Where can I salvage a vibration motor from, something that would be easy to install in a device like a game controller for force feedback

>>1204021

on pretty much every visit to the thrift store i see one or more old controllers with vibrators for $2-$5.

what would be a good op amp for general use nowadays? I've used lm358 and lm741, but they are quite old

File: Intuitive_Op_amp.jpg (42KB, 370x500px) Image search: [Google]

42KB, 370x500px

>>1204078
>op amp for general use
Typical pseudo question. There's no such thing as 'general use', it's alway specific use. For audio applications e.g. I would not use the 358/324 or 741 and prefer the NE5532. It really depends on the specific requirements. Interesting newer opamps are the rail-to-rail types where the input range includes (or even exceeds) the supply range. The inputs of the old 358 can also operate around 0V which makes it suitable for single supply designs. If you ever come across the book Intuitive Ic Op Amps by Thomas M Frederiksen, grab it. Frederiksen was the designer of the 324.

>>1204097
>There's no such thing as 'general use', it's alway specific use.
autism

>>1204098
general babble

>>1204078

the 1458 and its cousin the 4558 are probably the most-used general-purpose op-amps found in commercial and hobby projects. they're like dual 741's but a bit better. the LM324 is the most popular for single-supply use. for low noise, a ne5532 or tl022 are used. if you get those, along with the LM386 and the LM301, you're pretty well covered for most purposes.

these work fine, but they're old and dont have the super-duper specs of newer units. but 99% of the time they're good enough.

>>1204097
>>1204166

One problem that I have is that there are so many op-amp part numbers.

There is probably a guide somewhere that shows the best and newest op-amp for typical applications.

Meaning, is the 324 still a good go-to device for many simple applications, or has it been superceded by something that is cheaper, more robust, or whatever?

I can't imagine anyone actually using a 741 any more; am I wrong about that?

>>1204166

>these work fine, but they're old and dont have the super-duper specs of newer units. but 99% of the time they're good enough.

and that's what I'm talking about. if a new design is available for a good price, and it has super duper specs, how can I keep up with the state of the art without subscribing to Op Amp Monthly?

As a non-electronics guy, I wonder how you guys get any work done without falling behind on the latest power FET or whatever.

>>1204173
Quite often people will have a list of parts they are familiar with and if the part meets the required spec for the design, it's used. Otherwise, using part finders on either big distributors like Farnell or Digikey is a quick way to find a component that meets spec. Finally, you may browse individual manufacturer's websites, which often list the very latest up to date components. Not very many people stay up to date on the latest x, because not very many people need to.

>>1204173

we can use, for example, digikey.com, which has a parametric search function. so, if we need an op-amp that works at 3V and draws 100uA, and can do rail-to-rail outputs, and be below a certain noise figure, we just enter those specs, then pick the cheapest unit from the list of results.

>>1204173
>the best and newest op-amp for typical applications.
Linear Technology published such guide some years ago, at least for some applications.
Art of Electronics has a decent comparison table, too. That said, it's not application specific.

>324
>or has it been superceded by something that is cheaper, more robust, or whatever?
It is one of the cheapest opamps. So if you want cheap shit, it's worth considering. It isn't impressive in any other way.

>I can't imagine anyone actually using a 741 any more
It is still in production, made by several big players like TI and STMicro. They wouldn't make it if it didn't sell pretty well.

File: LM358|324.png (41KB, 568x402px) Image search: [Google]

41KB, 568x402px

>>1204173
>a good go-to device for many simple applications

I'm currently working on something quite exotic for an equally exotic purpose: a x10 dc-to-lf preamp for my multimeter, powered by a single CR2032. This is sub-millivolt stuff. Guess what I use? A 358, designed in 1972 and first introduced as the quad version 324, which then arguably became the most popular industry standard opamp ever.

Of course this cannot work because the 358 has a worst-case input offset of about 10mV and no pins for balancing. Or can it? I found a way to solve that problem. Should it drift too much I will have to zero it now and then, why not. It's not for mass production, it's for me and it's DIY.

The good thing is that almost all (?) DIP8 dual amps have the standardized pinout 1out2-3+4GND5+6-7out8V+. So if nothing goes I can easily upgrade to one of the modern low voltage rail-to-rail amps like the MCP6002 my dealer offers for 43 cents. So why then try the 358 at all? Erm.. I have a bag full of them. 5 cents per piece, original NS and TI, couldn't resist..

File: medium-PIC16F877A-PDIP-40.png (355KB, 1021x460px) Image search: [Google]

355KB, 1021x460px

For one of my classes, I have to build a frequency counter. I was going to originally use an Arduino, but then thought hey why not try using a PIC since I want to learn how to use those.
So, I found this:
http://www.best-microcontroller-projects.com/pic-frequency-counter.html

And this is really embarrassing to ask, but how does one program the PIC (as in how do I get what I've written on the computer to the microcontroller)? Everything I've ever used has been a development board that always have a header for usb.

I have an old 600W pc power supply that is fully functional. It is capable of outputting 12V at 50A. Im an electrical engineer. What should I make?

>>1204349
PICkit 3. You connect 4 pins to the IC and plug it via USB to your computer. I recommend a separate small breadboard for this purpose.

>>1204358
Ah, I see. Thank you.

Hey guys, I was looking for some advice, maybe you can help me out:
For a project at my university I need to develop a circuit for a line-following robot, which has one (or two) light sensor and two motors (left and right wheel) that can be controlled with +/-1 Volt (going forwards and backwards).
I am new to electronics, I've read "Getting Started in Electronics" and wanted to know which other books/youtube channel I could read/watch to learn more about this specific project. I've been making small experiments, but for now I want to learn how to "read" the black line the robot needs to follow.
Thanks in advance for any help!

>>1204414
>line-following robot
Have you tried googling this exact phrase? No? Do it.

>>1204078
>tfw I have 3 OPA2134s sitting in a component box that I use for general use
>tfw I paid $10 for each

>>1204430
>intended for high performance audio
>priced accordingly
>inputs and outputs don't cover either supply rail
>unimpressive offset voltage spec
>5V minimum supply

Why would you pick this as your general purpose opamp?

>>1204440
I bought one for audio and it worked well (still in use), then I bought another because I wanted to mess around with audio, then I thought I lost it so I bought another (this happened twice), and now I've got surplus.

What do people here use for a good rail-to-rail amp?

>>1204452
Dunno what others use, but
- if you're willing to pay $10 per amp, want wide supply voltage range and are fine with bipolar inputs, OP184/284/484 might be worth considering.
- if 5V max supply is ok and you want fet inputs and/or lower current consumption, AD8603/8607/8609 is much cheaper (but still quite expensive). It is also much slower and has higher (voltage) noise than the previous amp.

File: Screen Shot 2017-07-05 at 10.46.44 pm.png (60KB, 264x256px) Image search: [Google]

60KB, 264x256px

What sort of cooling would I need to cool a 200w peltier?
/g/ here, so something that could cool 200W of tdp I take it?
Noise is no concern. Let's just say I'm some evil villan who's current plan is to turn on a doomsday dehumidifier and steal all the water in the world for myself.

>>1204419
I did. And all I got was powered by 5V with chips, which I can't use (I believe) with 1V.

>>1204486
Where does that 1V come from?

>>1204483
A heat sink and a fan. Dimension adequately.

File: 200w.png (72KB, 1012x663px) Image search: [Google]

72KB, 1012x663px

>>1204483
A 200W peltier dissipates more than 200W of heat. How much more depends on the required temperature difference, but it can be over twice as much at zero temperature difference.
Pic related suggests that if you keep the hot side at 50C and the cold side at 10C, you'd need to dissipate around 390W. This would be (theoretically) enough to keep RH at 50%, assuming 20C room temperature.
The available cooling power would be around 100W, which would be enough to steal around 150 grams of water per hour. You'd need to run your doomsday machine a while for anyone to notice it.

You'd need 0.08K/W total thermal resistance from the peltier to ambient for these calculations to apply. Apparently this is possible with 125 x 150 x 75mm heatsink and 56m^3/h fan. Even then, these calculations assume 100% efficiency for everything except the peltier itself.

>>1204490
>watts of heat

>>1204490
Thanks anon, that clears things up
>>1204501
Energy is energy. Deal with it.

>>1204501
energy exchanged/time?

File: ohms_law.jpg (12KB, 300x435px) Image search: [Google]

12KB, 300x435px

>>1204507
Yes, "200W peltier dissipates more than 200Ws of heat per second."

and this incarnation of /ohm/ seems to be on its way to the archive

>>1204565
he mean that a 200W cell can pump more than 200W of heat from its cold side to the hot one.

>>1204565
It would be joules of heat per second

I'm trying to understand a few things in the circuit posted here: electroboom.com/?p=101

1. I still don't understand what D1 is for. If I'm not mistaken, the voltage across the transistor is 25V, and when it turns off it's 0V. Why is that a problem?

2. Why the 1uF capacitor? Couldn't he just leave it out? I also don't quite get how currency is supposed to flow through a capactor, while we're at it.

3. Could one use a microcontroller like an ATTiny to control the coils? I'm pretty sure it's fast enough, but I'd rather ask.

>>1204600
True, J=Ws=Nm=kgm^2/s^2

File: Capcharge.gif (2KB, 385x165px) Image search: [Google]

2KB, 385x165px

>>1204611
>1. I still don't understand what D1 is for. If I'm not mistaken, the voltage across the transistor is 25V, and when it turns off it's 0V. Why is that a problem?
The coil is an inductor, for which dv=L di/dt, so I guess the schottky diode is there to protect the transistor Q1 from the sudden high V it will get when you cut the current to the coil.

>2. Why the 1uF capacitor? Couldn't he just leave it out? I also don't quite get how currency is supposed to flow through a capactor, while we're at it.
It's a debouncing circuit: https://en.wikipedia.org/wiki/Switch#Contact_bounce. You have current across a capacitor until it's fully charged, pic related.

>3. Could one use a microcontroller like an ATTiny to control the coils? I'm pretty sure it's fast enough, but I'd rather ask.
Not sure what you mean, anyway the speed of the electronic depend on the speed of the projectile; you want turn off the coil when the projectile pass through its center.

File: 91vmO2zB5TL._SL1500_.jpg (326KB, 1355x1500px) Image search: [Google]

326KB, 1355x1500px

Started a thread about it but I figure I'll copy it over here as well since I love you guys:

If you don't know what Soylent is already, it's a powdered food substitute. Measure it out with a spoon, mix it up, drink it for lunch. I want to make a machine that will automate the measuring, and maybe the mixing later on.

My plan is to have a small 8 segment display and a few buttons to allow a certain number of calories of Soylent to be selected. Once the selection is made, the machine pours the powder onto a scale until the right weight is on it, based on the nutrition information. Typically this is around 109g for one meal, which has a volume of half a cup or so.

The electronics side will be done with a PIC microcontroller, some basic IO stuff, and pressure sensors to measure weight. The main issue is going to be slowly dispensing powder. The existing powder dispensers I've seen are for biomedical companies who need microgram precision and are way outside my budget. So my thought is to use something like the food dispensers in the picture, and have the crank be attached to a motor. Issues that I can foresee with this are:
>powder dispensed in too granular of a quantity, e.g. one compartment of the waterwheel-like mechanism holds 30g, can't dispense a partial compartment, so I'll be over- or undershooting
>hard to attach the motor, since I'm not a mechanical guy

Anyone have any thoughts? Are my fears unfounded? Is there a better way?

File: SCH.jpg (25KB, 512x361px) Image search: [Google]

25KB, 512x361px

>>1204611
>what D1 is for
When Q1 is OFF the voltage across it (the drain-source voltage Vds) is V1, 25V. When Q1 switches ON, Vds goes (close to) zero and the current in L1 ramps (!) up to 70A or whatever current it reaches during the ON phase of Q1. Now the coil collects and stores energy in form of the magnetic flux it generates. When Q1 is now switched OFF, e.g. because Q2 switches ON, the stored energy has to go somewhere, else it would immediately destroy Q1. The fast diode D1 now offers a (current) path the stored energy can follow and 'discharge' the coil. The energy stored in a coil follows (the square of) the current that flows through it.

You may want to read this a few times because your assumptions are far from reality.

>>1204633
>It's a debouncing circuit: https://en.wikipedia.org/wiki/Switch#Contact_bounce.
I'm wrong, it's obviously not a debouncer.
From his description I get that it is necessary to prevent the user from holding Q1 open for too long, which would fry it:
>If there is no bullet inside and the trigger is pressed, the gate voltage of Q1 drops as the capacitor at the gate is discharges by the 1M Ohm resistor. At some point this results in increment of the Q1 drain-source resistance, meaning a high power drop across Q1 that will blow it up. So there should always be a bullet.

is soft switching like ZVS always reasonable when designing DCDC converters? independent of power etc

>>1204649
>I'm wrong, it's obviously not a debouncer.

yep, it's a differentiator. it'll turn a square pulse from the switch into a short spike of around a second. i would think that it could be much shorter, like 10x or 100x shorter, and still work.


>>1204611
> Could one use a microcontroller like an ATTiny to control the coils?

you can use any kind of logic gate or CPU that outputs 5V instead of the switch. you might need to feed that logic signal thru a transistor which can dump, say, 100mA into the gate of the FET.

>want to make a kite camera thing to map my university
>cool I just need a micro controller, camera, battery and image processing data libraries
>lets use a raspberry pi
>using a a raspberry pi, cheap camera and lithium battery is more expensive than just buying a camera, tying to a kite and set to record.
>without taking the time coding the data I/O and conversion from camera to storage.
>if I use a cent range micro processor I'll have to code in assembly from scratch
just fuck my life

>>1204633
>>1204645
>>1204672

Thanks guys! I think I got it now. I'll try coming up with my own design and see what happens.

bump limit reached

new thread >>1204692