/diy/ - Do It Yourself

>>1726960
this thread's digits brought to you in part by the MAX6960 4-wire serially interfaced 8x8 matrix bi-color graphic LED driver

Lol

>>1726517

>>1726988
So it's like a half-bridge but with parallel inductors instead of series caps? I think it makes sense, the chokes are acting as low-pass filters of a sort.

>>1726962
guess you'll just have to read the transistors themselves

I don't know if this is the right place to ask, looking for some help and/or advice on soldering.
My old ersa soldering iron is crapping out. I have a Hakko soldering station which is very nice to work with but cumbersome to take with me on sites, which is why I used my old Ersa 15W.
The Hakko FX-600 looks like a nice replacement but it only comes in 100V but I need one in 230V EU variant.
The chink ones are a no go since I need something with proper static protection and I work with sensitive electronics, parts that start at 2000 eurodollars. For this I am willing to pay top dollar for a good quality brand.
Anyone know of a good quality brand? New Ersa and Weller look too chinkified or too dated, no temp control means it will burn sensitive panels. Thanks in advance.

>>1727021
>chink ones are a no-go
lern2continuity tester
consider a TS100 with a grounded power brick, or a battery pack + ground strap. V- is passed thru to the tip metal by design, otherwise it couldn't heat up. you can find them as sexy as you like

>>1727046
I will not consider a chinkshit TS whatever because the instruments I work on cost $250000 to millions per piece. I am not going to use a chinkshit and jeoperdize an entire production by using low quality/non qualified tools. This is actual work, not your POS arduino hobby bench

>>1727050
ok boomer

>>1727082
Bought the Hakko and grounded adapter, stay mad zoomer
Also Tienanmen square massacre 1989, Vinnie the Pooh, Free Tibet, Free Hongkong etc...

Hey bros.

I’ve been reading the recommended books (Art of Elctronics and Practicle Electronics for Inventors), but they haven’t satisfied my desire to learn more about power supply design in depth (especially SMPS). I’m a senior EE major and we don’t really learn much about power supplies aside from transformers. Got any good reads for this kind of thing?

Are there any non boomers in /diy/?

>>1724276
Reporting success. Had a little trouble with the cheap braid but the solder sucker from the cheap soldering kit came through. $12 to bring this old thing back to life is pretty satisfying.

>>1727147
Pressman, Switching Power Supply Design

>>1727157
gen X reporting in

>>1727050
So does your workplace not supply you with the proper tools or what?
Where I work we're all basically just given JBC irons and hot air guns and any other tips/tools we need. What kind of ramshackle operation are you a part of where you need to supply your own tools?

>>1727157
>>>/diy/mcg

>>1727147
I really liked Christoph Basso's stuff on the subject but he deals a lot with the control and simulation side of supplies

>>1724276
bretty gud
now make sure you used an appropriate replacement capacitor series so you're not doing it again in three months

>>1727194
How would I know? by apperance they were almost identical, same 35v 470uF, same orientation.

>>1727157
Zoomer reporting in. Unlike the boomers I don’t have a seething hatered of micro controllers

>>1727200
you would have had to look up the old series and compare its specs to the new ones to know for sure
but similar size cans are a pretty good sign

>>1727206
there's a time and a place for everything. and it's called college

>>1727012
what difference does it even make?

>>1727239
aside from npn vs. pnp and max collector voltage, perhaps less than I thought

>>1727206
>>1727157
Based zoomer reporting in. Microcontrollers are alright, overused as hell though.

how do i set this up to be an inductor in my circuit? it is two inductor cores that slide into a thing with pins which i guess i plug into the breadboard. do i then just wrap some wire around it and then calculate inductance based on number of turns and the inductance factor of the cores? am i supposed to just wrap around the plastic part that holds the core, or around the core and plastic?

Forgive me for copying from /mcg/

Any problems transmitting spi from board to board, possibly multiple slaves, 1 on each board? The signal would be clocked pretty low <8 MHZ and no wires just headers or some form of board-board interconnects. I'm not sure if all the impedance changes will fugg me and can't find much online. Thanks boyos.

>>1727252
I can agree with that. Although Arduino, PIC and the like are an easy start for beginners, many of them get trapped into solving all their problems with them and ignore logic design.

>>1727280

take it all apart, wind coil around the plastic form (as per instructions or formula), re-assemble, solder wires to pins (or use loose wires temporarily), reassemble, hold the two core halves together firmly, and test. if it works ok, solder the wires and glue the two two half-cores with super glue. the plastic form should be able move slightly.

>>1727252
I'm only averse to microcontrollers because the arduino enviroment is pretty shitty, and moving away from it poses some difficulty with all these drivers and ides and terminal-based interfaces.

>>1727288
okay thank you, so just solder the ends of the coil to the pins? am i supposed to connect it to every pin, or just the ones i want to use? i have the data sheet but to me it doesnt seem to give much info on the pinout. i would guess i just use one pin as the input and the pin opposite of it as the output, but never used these before so not sure

>>1727315
>solder the ends of the coil to the pins?
Yes
or just the ones i want to use?
This. There's pins arranged as two sides/rows, so usually you'd have the transformer primary on one side and the secondary on the other, with the ends at the end pins and any centre-taps between them. In your case, it's quite possible that those pins are large and rigid enough to be used in a breadboard without soldering jumper wires from them, which is nice. Just wrap and solder the wires to wherever is convenient, it's not like you need any high voltage isolation.

Have you considered using a toroid instead? More difficult to wind, but less magnetically lossy.

>>1727315

just use 2 pins, for best solidity, one on each end, diagonally opposite each other.

>>1727315

just use 2 pins, for best solidity, one on each end, diagonally opposite each other.

for best results there should be no gap between the cores. that means dont put glue on the opposing faces, instead, press together and drop glue on the crack between the faces. this will give you zero gap.

>>1727323
>he doesn't mix ferrite powder with his superglue
>he doesn't use ferrite epoxy to customise his own magnetics

>>1727320
well i am just using what my universities EE shop has, i dont believe they have toroids but i will check again tomorrow

>>1727177
> What kind of ramshackle operation

the kind of ramshackle operation where they have guys working on expensive equipment who can't figure out how to tie a 1MΩ resistor from his soldering iron tip to ground.

>>1727206

Boomer here. We don't hate micro controllers, we just had a different name for them: "the company mainframe". We'd sometimes use it to simulate analog circuits (zoomers call them “quantum computers”) before we built them up.

Not sure why the world would need more than a few of your "micro controllers" though.

>>1727342
Analog circuits don't need to be cooled down to 0.1 Kelvin to work at all, retard.
Also fuck off, micros were already widely used 30 years ago. If you never designed anything with a micro in it you're a retarded LARPer who never actually did any serious electronics.

>>1727351
>If you never designed anything with a micro in it you're a retarded LARPer who never actually did any serious electronics
Or someone doing really high-speed stuff with FPGAs instead. Maybe low-power stuff too.

>>1727360
Nah. Micros are about 20 times as common as FPGAs. If a company can get away with doing something with a micro instead of an FPGA or ASIC they will.

>>1727157
<35 checking in.
>>1727342
You know you found a real boomer, when they talk about a minicomputer and mean PDP-11 type mainframe. We both know you used a PIC in the 90s and liked it.
>>1727365
Absolutely. With FPGA costs coming down every year though, soft micros are becoming more common.

So I want to put some Ni-mh batteries in the project im working on. They're wired in series, the voltage of them all fully charged is about 6v, but i'm wondering if i can just charge them from 5v from a usb chord.
will a constant 5v to the batteries damage them, or will it just top them off? It would be nice if i could leave it plugged in and not have to monitor it charging. I know that there are boards that you can get on ebay that will charge batteries in that configuration, but they're 2 or 3 dollars and i'm cheap.

Any of you bois work in HVDC?
Went in for an interview as a field engineer at a station and am pretty pumped about doing BIG power electronics lol

>>1727157
im a zoomer (22) just about to finish my EE degree and then switch into something else for masters because i actually came to find this shit boring as a job. will still work with electronics as a hobby though

>>1727374
make sure to take some nice closeup pictures of the switching elements for us

>>1727377
What's your job like? I honestly have no idea what to expect and really can't even imagine what I might end up doing after I graduate. I wasted like a decade from 12-21 on programming before I decided I can't stand it and would hate doing it as a career. Switched to CompEng and so far I'm enjoying the courses, which is pretty unusual for me. Shit's hard as fuck to me though, meanwhile programming was always pretty easy for me. I'm shit at math.

>>1727377
are you me?
going into ee as a job was a huge mistake, i love it as a hobby but hated it as a job, so now i do IT shit as my full time job and ee as a hobby

>>1726960
what do you think the emf is like near that thing?

>>1727295
ok boomer

>>1727252
>over
sometimes they're just cheaper and less hassle

>>1727281
once a wire gets to about 1/16 wavelength in length, you need to start treating it as a transmission line. 8MHz ~= 40m wavelength / 16 = 2.5m, but you will also be concerned about higher harmonics, probably up to 7f or 9f. I don't think you need to strictly consider controlled impedance etc. but you should think about loosely designing with those techniques, keeping those traces wide and keeping stub lengths short
for maximum success I'd just put buffers on every board. or reduce the clock frequency

>>1727324
>ferrite-loaded epoxy
this anon knows

>>1727368
>soft micros are becoming more common
doubt. every major FPGA vendor except maybe Intel has an offering with a hard-core processor on the side of their FPGA

>>1727369
no, you can't. step up the input voltage then use a proper NiMH charge controller. they aren't petulant little prima-donnas like Li+ or Hillary Clinton on election night but they still like to be treated with some respect

>>1727414
>this anon knows
I think you can also get ferrite 3D printer filament, though I can't imagine it's terribly good. Still, for testing motor designs its certainly better than nothing.

>>1727021
If you want quality then JBC is what you are after

>>1727414
>election night
"..but it was MY turn!" she said.

>>1727424
...as she smashed up the green room like a side of beef in a china shop

>>1727324
Top kek I did that once.

so i built something similar to pic related, but since my soviet-era house has ungrounded sockets i'm getting static and crackling sounds from the speakers when plugging in multiple devices. an easy solution would be to switch the ground signal off for the channels not in use. can anyone recommend a 3PDT/4PDT relay suitable for audio signal switching?

Are ladder circuits the most comfy circuits?

>>1726960
Iam trying to make a blow dryer for my hair, but my 5watt dc adapter isnt making the nichrome wire hot, wat do bros?

>>1727510
>5W

>>1727423
I think richfag's looking for temperature-regulated all-in-one pencils. JBC doesn't appear to have those

>>1727502
grounding is bourgeois
pretty much anything that calls itself a 'signal relay' ought to do the job. or, you could just double the relays you're using now

recommend variable soldering station?

>>1727548
chinkshit is extremely variable

http://sound-au.com/articles/bench-supply.htm

Is he right /diy/? Is it not actually worth building your own bench power supply?

>>1727577
>but they're SMD only
>t.boomer
to answer your question, it can be a useful educational exercise, if you're equipped and willing to test the hell out of it. Pic related. if you just need an output or two for whatever, fuck it, use a chinkshit buck board or cut the end off a USB cable

>>1727577
I did mine a year ago when I started with electronics. It was a good experience, it looks like a particle board box with some weird holes in it. I actually had to build something like 4 or 5 boards, because I fucked up them so hard it wasn't worth to fix it.
I have learned
>to solder
>placement>that you only notice stupid shit after they are in place and it's too much fucking work to remove them
>most of the times in discrete electronics back of the hand calculations and getting a feel for the behaviour of the device is more important than meticulous calculations
Mine can go from 0 to 20 V and put up a amp or so.

>>1727603
I have plans to make another one, with a switching pre-regulator but life got in the way and shit. I actually like analog design and I could not be bothered to write a PIC to reat the voltage and control and multiplex a 7 segment display.

>>1727603
>most of the times in discrete electronics back of the hand calculations and getting a feel for the behaviour of the device is more important than meticulous calculations
under-appreciated point, thread title notwithstanding
>that you only notice stupid shit after they are in place and it's too much fucking work to remove them
rework is an under-appreciated art

>>1727510
You should put 200 hundred of those supplies in parallel. Should do the trick.

QRD on lithium ion battery packs?
Not sure if this is the right place for this but you guys seem sharper than, and probably avoid the folks on SQT - I have a few cordless tools that use 18V li-ion battery packs (pretty sure they are just a stack of 18650s) and am wondering about the best way to use / charge / store them. I did some searches and found stuff like https://www.powerelectronics.com/mobile/proper-care-extends-li-ion-battery-life which is really informative, but I'm not sure how to or whether I can even implement these best practices using the standard charger. I do store them cool (60 degrees F) but generally charge them full and usually use them down to 25% charge (according to the meter on the battery). How much of a practical difference would I get storing them in the fridge and recharging at like 75%? Also for storage of more than a few weeks, should I run them down to 50%? Basically I'm trying to figure out if it's worth the extra effort to be OCD with this. Also how bad is it to discharge a pack in like 5-10 minutes? (I've been doing this a lot recently with some high-torque applications). They are 5.0 Ah 18V packs and cost about $80 a piece. Thanks for any recommendations.

where do I find schematics for laptops of a certain kind?

>>1727672
Rothschilds bow to them
the main things to avoid are high temperatures, deep discharging, and long time spent at a high (or very low) state of charge. sounds like you go through them quickly enough that no special measures would be indicated. the fridge at 40%SoC is indicated for longer-term storage (like manufacturing), not really a major big deal for you. maybe you'd get 10-15% more cycles out of being OCD (ballpark). shame your SoC indicator is so coarse
if you have a decent spot welder for small applications, and you can find a reasonable supplier for replacement cells, maybe it's worth rebuilding your old packs once they lose their strength
thanks for the link tho, might add it to the OP

>>1727676
elektrotanya is worth a try, though they mostly do consumer A/V stuff last I looked

Zoomer student here. Grade my first design project, /ohm/.

It's an 8-bit adder/subtractor/and comparator without floating points. Takes in W-Z as the function code on clock count 1, S-Z as operand 1 on count 3, and then operand 2 on counter 4. Answer is stored on count 7, and count 0 signals the answer is valid. Also outputs what operation was done.

Is my layout shit?

Holy shit look at this. Just one extra sample and the spectrum leakage is massive. Massive. Take it out and boom, the THD is on the order of 1.0^E-15 thats 0.000000000000001 lol And the noisy phase graph corresponds to the corrected one on top. Once the noisy "almost zero" bins are zeroed, the phases look correct, too! Amazing technology. Changes your perspective on life. Of course you can use a windowing func but thats a different story. No window in this test at all.

>>1726960
I'm a squatter from >>>/out/ looking to DIY a grow light and just wanted to get some confirmation on things, I've never messed with this stuff before

So Samsung has there horticulture linears and I'm looking to use 2 of them, they are 2 feet long and consume ~51w each at the typical drive current of 1200ma, I know I can't get away without a heatsink and trying to find a premade one with the right dimensions is proving to be a nightmare, I'm wondering if I can just get a raw aluminum plate cut and just use that and if so what thickness would it might need to be?
Or could I even go a step further and get some aluminum angle bars or something from home depot or similar?

Also would an HLG-120H-48A be fine for driving 2 of them in parallel? That would give 1250ma to each strip and I wouldn't think the extra 50ma would hurt anything

>>1727720
Window rails can be used without problems as heatsinks. Did it myself several times.
>HLG-120H-48A
I wish I knew what those letters meant. But if something says 1250mA, it doesn't mean it will put out 1.250A regardless of what it connects too. It means that it was designed for a max current of 1.250A or it operates nominaly with that load.

>>1727604
>I could not be bothered to write a PIC to reat the voltage and control and multiplex a 7 segment display.
You know they have dirt-cheap 3-lead voltmeters on eBay, right?

>>1727725
>I wish I knew what those letters meant.
https://www.meanwell-web.com/content/files/pdfs/productPdfs/MW/HLG-120H/HLG-120H-spec.pdf
The 120 is the max output wattage and the 48 is the max output voltage, 120w/48v is 2.5A for the current
The forward voltage of the strips is 43V and from my understanding most meanwell LED drivers will operate in constant current mode when faced with an LED with a forward voltage lower than the max output voltage
Since the current is 2500mA I was thinking you can just run 2 strips in parallel which splits the output current to 1250mA per strip

>>1727732
I know, but bolsonitro fucked the Real and the mail mob charges you 15 Reais (a decent meal) for every package (this plus shipping) that comes through.
>Since the current is 2500mA I was thinking you can just run 2 strips in parallel which splits the output current to 1250mA per strip
Probably. Led lamps and strips usually have resistors in them (to limit the current) , so as long as the input voltage isn't stupid high above the rating, they'll usually work ok. Also, unless it comes from reputable brand and/or you tested a sample, usually LEDs are less powerful than advertised.

>>1727698
Thanks a lot my friend, that's exactly what I was wondering. Will keep abusing them and giving no (extra) special treatment. My first two are about 5 yrs old and I plan to re-pack them once they start degrading enough.

>>1727711
Zero padding can do some wonderful things to signals. Windowing is more about sidelobe characteristics and reducing leakage across bins IIRC.

>>1727595
Well I guess /diy/ bench supplies are always worth it as an academic exercise, can't argue with that. Maybe I'll go balls to the walls and /diy/ an SMU. That'd probably be a good learning experience and it might even be financially worth it too unlike /diy/ bench supplies which tend not to be financially efficient.

>>1727768
Just use an analog panel meter. The digital meters are shit anyway.

>>1727400
Not worse than near a standard secondary distribution circuit. Same voltage (110 or 220) and the taps are probably all made of extension cords. So the current loops don't enclose much area.

before I return these resistors for stupidly buying 1/4W instead 1/2W could I get away with a 1/4w resistor at R204? I'm gonna say no. (The last one snapped in half because of a cracked pcb btw, not from any overheating or melting)

favorite diy solder fume vacuum? do I rly need to spend 600?

>>1727834
i like the smell

>>1727720
It honestly really depends how hot it will get, there is a reason heatsinks are shaped the way they are, to maximize surface area for cooling ect. Honestly though an aluminum plate will probably surfice, just make sure you use some type of thermal paste so you get the most out of it. And if you can use the biggest peice of aluminum that will fit, it will make a better heat sink, angle bar would make it more effective too. Id wager a thickness of 2-5ish mm would work, not an expert though.

>>1727844
Me again, thermal tape with work fine btw, I don’t think there is a need for paste it’ll probably just make a mess since it’s quite large.

>>1727705
Use busses
Seems reasonable apart from that

>>1727822

no way to tell unless we have a voltage at the top of R204. if i assume 12V, it works out to 6mW. so a 250mW rating would appear to be adequate.

but you dont even need to calculate anything, just intuit it like this: the lil pot is probably rated 1/10 of a watt, and since it's bigger than R204, then R204 is gonna be burning less than 1/10 of a watt.

>>1727834
diy one with a 140/120mm fan and a filter if you really care

I need to make a circuit that lights up an led when a li-ion battery gets close to empty (3.4v sounds about right to me). I want to fit it inside a gameboy advance, so I have very little room to work with, i'd like to deadbug something inside the case if i could. I have some pn2222 transistors with me, could one of them be used for this purpose? I've done a bit of googling and I cant find anything useful. any help is appriciated.

>>1727876
fair enough, and yes you were right it was 12V

Spitballing making a nichrome resistance furnace. There's a nichrome calculator online that helps not having to manually calculate from resistance tables and temperature change and shit, but it has one confusing bit. You can't enter an intended voltage, it only has an output for it with the calculation output. You enter length, gauge and intended temp and you get out resistance, amps and volts. So for example you put in 10 feet 18 gauge 2000f, you get like 4 ohms 25 amps 100 volts (it was close to that, can't be arsed to check again). Now I'm planning on using 240vac (with a pid), and ohms law says 4 ohms and 240v is 60 amps. The 100v is coming from 4 ohms times 25 amps, but I don't understand why they would have it do that instead of asking your intended voltage, since that's what it's actually going to be. Am I being retarded or no?

>>1727909
Thermal power = electrical power = P.
P = temperature difference / thermal resistance (inside to outside)
resistance per unit length = resistivity / cross-sectional area
R = resistivity * length / area
P = V^2 / R * pwm duty cycle
Hence:
V^2 / R * dc = ∆T / thermal resistance
All in SI units, because all those proportionality constants that inch-standard/imperial needs really drags you down. If you're using a PID and plan to use a low duty-cycle in steady-state then you shouldn't need to worry too much about thermal resistance.

Use https://en.wikipedia.org/wiki/American_wire_gauge or picrel to get the cross-sectional area for your particular awg choice. In your case, 18awg is 0.823mm^2 or 823E-9m^2, and nichrome has a resistivity of ~1E-6 Ωm.

I simply wouldn't trust online calculators for this sort of thing. I'd instead do math in a spreadsheet or python program so I can change initial variables and see how the answer propagates.

>>1726960
I'm interested in building a synth of some kind, or playing with electronics to make music.

Anybody got books to recommend for that purpose?

>>1727720
>HLG
stick with Mean Well purchased from a reputable distributor. they are probably cheaper than stoner-badged crap
>angle bars or something from home de pot
I use the square U-shaped channels for lower-wattage residential lighting applications, in free air. you might be surprised what a little forced air flow across them can do. you should qualify with a measurement of some sort (thermistor, thermocouple, thermal camera, in increasing order of cost and reliability)

>>1727705
>7 clocks
y tho, if you have all those cycles you should be able to reuse a smaller adder and do the arithmetic serially
>two-input gates connected as inverters
y tho
>decoding a timer made of TFFs instead of using a shift register
lol
3/10 use real ICs or use an HDL

>>1727885
>lights up an led when a li-ion battery gets close to empty
there are voltage sensor chips for this that will work better than anything you can slap together, but you have to order them. why not use an SMD comparator like a LMV331/LM393 with some micropower voltage reference like an LM4040, Pic related

>>1727961
there are forums/web sites for diy synth makers which probably know the state of the literature better
>playing with electronics to make music
there are forums for circuit-bending weenies too

>>1727992
>7 clocks
But there's one clock, anon.

What inverters are you referring to? Do you mean the ones that convert to two's complement in order to do subtraction? How else would you do it? And how do you use a shift register for this instead of a TFF counter?

>>1727999
I think he's referring to the NOR used as a NOT that's connected to Count_3 which I'm assuming you did so you wouldn't need an inverter just for the one use.
You can also do the clock divisions with a counter or shift register.

>>1727992
>HLG
>stick with Mean Well purchased from a reputable distributor. they are probably cheaper than stoner-badged crap
HLG is a series in Meanwell’s led driver family, it’s unrelated to the company HLG

>>1727999
sorry I meant 7 clock cycles
the NOR gates you're using to make inverters. look up fanout to see why that is a bad idea
the 2's-complement circuit is fine
you string DFFs together. look up ring counters and look at how much of your decoding logic just vanishes

>>1728020
shit, thought he was talking about Horticulture Lighting Group, good catch

Rate my 5V regulator

Rate my 5V regulator

>>1727913
Damn that's pretty in depth. Mucho grassy ass. Only thing is it's not a pwm pid, just turns a plain solid state relay on and off to cycle, so I should drop it from the equation since it's just 1.00 right? Also what is the dc variable supposed to be?

What would be the best soldering iron to start out with?

>>1728098
8/10 the cap on top leads are still a little bit too long

>>1728107
duty cycle (0..1)

>>1728137
show budget and anime tiddies
hakko 936d clone from ali, take your pick. TS100 is smol and cute but breaks easily

>>1728098
>>1728160
8.5/10 have an extra half point for excellent soldering. what's your gear rundown

>>1728160
>>1728137
I have a non temp controlled thing that just plugs into the wall and it works great.
The ones where the tip is secured with a screwable cover are better than the old pencil type because the tip lasts longer.
Get one of those brass shavings thing to clean and help tin the tip and apply non-clean flux to whatever you want to solder, makes the process go smoother in every way.
Tin the tip from day one and keep it tinned as best as you can, any non-tinned part of the tip will corrode and stop accepting solder. Try to leave the thing plugged in sitting idle as little as you can, or again, the tip will oxidize and stop working. Don't use the iron to burn plastic. Don't let the tip touch wet things (besides flux) or it will oxidize faster. Before buying check that it has a thick cable, otherwise the cable will break. And you want the plug to have a earth connection (and ideally check with a multimeter that it is actually connected to the metal tip), otherwise if the heating element breaks it might shock you when you touch the tip with the solder.

>>1728107
Duty cycle doesn't imply a high pwm frequency. Average out the rms voltage over a few on/off cycles over the minutes, divide it by 240V, that's your duty-cycle.

>>1728239
Honestly that sounds unnecessary, All I'm looking for is at least 2000 degrees and 30 amps or so (50 amp circuit but I'm not trying to get in a dick waving contest with the sun, nor do I want to use 000 battery cable for power leads). so long as it meets those requirements, how much it cycles is really of no consequence to me.

On a related note, I'm looking to get into SMD soldering. Do I need a hot air station or can I get by with the pointy tips for a standard soldering iron?

>>1728164
It's a JBC station and it's not really /diy/ since it's a picture I took at work. Work-grade gear.

>>1728285
How about a hot air gun and laser thermometer?

>>1728098
Looks like shit and probably functions like shit. Would never use in any application.

>>1728285
Depends on what kind of SMD. You would only need a gun for no-leads chips. For stuff like capacitors and packages with leads you need
a basic soldering station. Nothing fancy but just a soldering iron wouldn't be enough. And you would need a proper tip and some good, potent liquid flux. Here is the type of a tip I use to solder the entire side of a package like LQPF100 in a single sweeping motion. And if you screw up and leave a lump of solder and accidentally connect a bunch of leads together -- no problem! Just apply plenty of flux and sweep again. Quick and nice. And I am talking 100 pins with 0.5mm pitch. Sounds hard but it is super easy after just a little bit of practice. I can post some pics if you want. Oh and you would also probably need a good lens or better yet a visor. I have this one:
https://www.amazon.com/Donegan-OptiVISOR-Binocular-Magnifier-Lensplate-Magnifies/dp/B000BPWPRK
And this soldering station:
https://www.amazon.com/Weller-WES51-Analog-Soldering-Station/dp/B000BRC2XU
(oops, looks like it has been discontinued)

I’m working on an electromagnetic accelerator, a mix of a coil gun and rail gun. The first two stages are a coil gun, third is a rail gun. Haven’t seen anyone try this approach so I thought I would. I finished the coil gun part and it fires at a respectable (for me) 110fps already and I expect to get another 30fps out of it with some fine tuning and adding another 450v 1000uF cap. However, I need a strong conductive metal for the rails. I could use steel but I want to maximize current/conductivity, so I wanted to use Copper Beryllium. It’s hard finding 2mm flat bar in the USA and wanted to know what you’ll think I should do? I think normal copper might not last long as rails, I figure it might be too soft. The cap bank is a 13,200uF 250V bank so it’s a fair amount of energy that, I think, will run a high risk of welding if the metal is too soft and deformed from previous firing. If anyone has any wisdom to provide on acquiring copper beryllium or about the design in general I would be very grateful.

>>1728386
Do you have a particular alloy mixture you want to use? Could you work 2mm flat bar yourself if you had ingots or do you need something that's machined?

>>1728352
Thanks. I have a similar Weller station which will hopefully be fine. I'll look into those tips. Is there a particular type/model of flux you would recommend?

>>1727913
length/area is a fascinating variable

>>1728160
>>1728137
can I microsolder with that hakko? what's the tip compatibility like?

How much would it cost me to get my own vanity IC built?

>>1728463
I don't really come across it very often outside of thermal/electrical resistivity. Unless there's some other notable example?

>>1728471
Last I checked, vanities didn't need ICs. Fucking zoomers want "smart" internet connections to everything.

>>1728450
Oh man.... making alloys is way out of my skill level, I’d use steel before I get to that. I just need something very straight, conductive, hard and 2mm is height. I’m honestly just considering normal copper flat bar, it’s easy to find, and IF it doesn’t hold up, then investing in beryllium copper alloy if I can even find it. Worst comes to worst I’ll take the typical rail gun route and use steel (as inefficient as it is)

>>1728472
Bruh if I want to blink my led on a home WiFi network with my RaspberryPi and custom made phone app, let me be.

>>1728493
Have you contacted any of the major CuBe makers to ask whether they have flat stock available or if they can point you to small suppliers? I can't believe it would be that tough to source the stuff. With things like this you might need to actually get on the phone, you know?

>>1728471
No manufacturer would ever entertain that idea. They do large production runs with minimum order quantities in the thousands. Nobody does one-off runs of ICs. Since you're asking such retarded questions I'll assume you also lack the relevant skills and software to design the dies so you won't even get in the door. Beyond all that if you managed to find someone willing to do a one-off run it's going to be expensive, way more than you can afford. Large production runs have the benefit of economies of scale. One-off runs do not.

Just give up on this idea. It's retarded and you know it.

>>1728495
I’ve looked at a few suppliers, many only offer them in imperial sizes via their catalog or have way higher than I need minimum orders. I will probably get on the phone with a few of them that don’t have those limitations and hope their minimum order isn’t too crazy ect. Appreciate the help, will definitely ask a few suppliers where to go.

>>1728499
Worst case, ask about "samples". If you're doing a one-off project you may be able to get by on samples, which are usually a nominal fee for shipping with no commitment to purchase.

Also if they have MOQ restrictions you might ask if they can refer you to a local distributor or retailer of metal products that they sell to.

>>1728498
I was just curious what the smallest production run costs are. I thought if they could cut the dies small enough maybe they could include your circuit in shared masks/wafers that also include other people's designs, kind of what they do with PCBs.

>>1728502
christ, do you have any idea what the setup costs for a wafer are? and then you're talking about just getting a couple made? better multiply that by a few at a minimum to ensure you've got the yields.

and then the setup for the machinery to attach bond wires? you can't just mix that up with other people's shit because everyone's design is gonna have different bond wire locations.

what exactly are you wanting to get made anyway?

>>1728471

Easiest thing to do would be to get one of these, program it, sand off the numbers and etch your own vanity logo/identifiers on it. Manufacturers do it all the time.

...unless you had some analog stuff in mind.

>>1728503
>christ, do you have any idea what the setup costs for a wafer are? and then you're talking about just getting a couple made? better multiply that by a few at a minimum to ensure you've got the yields.
Well, at least the number of times the wafer would be used for a certain order size would be way higher than if the wafers had to be filled with a single customer's chip. For instance, if you want ten dies and you gotta use the wafer ten times, filling the thing with the same chip would mean a minimum order of 100, while if you mix it up between customers the minimum order would be of 10.
And besides, maybe there was a company doing e-beam lithography commercially and somebody knew the price.
>and then the setup for the machinery to attach bond wires? you can't just mix that up with other people's shit because everyone's design is gonna have different bond wire locations.
That would be easy to standardize. Just know that whatever you route to X pad on the die will be routed to Y pin on the chip. Or they could ship the bare die and let people solder on it by hand if it's big enough.
>what exactly are you wanting to get made anyway?
Nothing, it's just hypothetical.

>>1728386

Just nickel plate the copper. Nickel is pretty hard stuff and plates well on copper. You can do it in your kitchen.

I need to solder in order to replace the headband on my headphones, but I don't have a soldering iron. What's the cheapest soldering iron you know that satisfies basic standards of quality?

>>1728516
>inb4 he should use an FPAA

>>1728285
https://www.amazon.com/Aoyue-888A-Digital-Soldering-Station/dp/B01NBUJTTS

>>1728495
https://en.wikipedia.org/wiki/Berylliosis so it's being phased out

>>1728467
for some suitably large value of "micro", sure. uses the cheap, ubiquitous 900-M series tips. I do 0402s with them easily

>>1728498
>Nobody does one-off runs of ICs.
then how do babby chip designers in university learn anything
>I am 12 and what was MOSIS
https://en.wikipedia.org/wiki/Multi-project_wafer_service

>>1728471
see above, links in wookiepedia. shame Ferranti et al. aren't running semi-custom analog arrays anymore, those were a nice cheap way of getting chips (see Curtis Electromusic)

>>1728621
>then how do babby chip designers in university learn anything
They don't. There's no profit in the student market. Those are hands on skills you learn in the field.

>>1728628
>I am still 12 and I am still ignoring MOSIS
hese people have been around for over 30 years
https://www.mosis.com/
please kill yourself for being a larping kike on your way back to the flat earth sub

>>1728580
Hakko

>>1728526
Huh, wouldn’t have thought to do that. I’ll look into that and see if it effects conductivity and the like. Thanks

>>1728500
I will do that, thanks for advice anon. May your resistor leads stay straight.

>>1728675
Good luck. Honestly if you've not been on the phone to local supply houses yet, you should. Find out the UNS code or ASTM spec code for the alloy/format you want and ask them to quote you a price for that alloy. Of course they won't have any in stock but ask if they can order it for you. You should be able to get a price—even if it's a high price it's better than it being unavailable.

And since I'm guessing you don't have the equipment to machine the stuff precisely you should consider whether you can build more flexibility into your designs; like you say you want 2mm, can your design accommodate 1/16th inch? Just be ready to consider compromises. Obviously you've already considered compromising on the alloy itself by going to steel, but can't the dimensions change slightly? Just my 2¢.

can I use a pulse pattern generator as a generic psu?

>>1728741
How much are you drawing from it?

>>1728743
5v, also what is a pulse pattern generator used for? I can't find any satisfactory answers, testing/engineering circuits?

>>1728744
I meant current.
Pulse pattern generators just make square waves to my knowledge. You'll need a low-pass filter to use it as a power supply or perhaps it'll let you set a 100% duty cycle.

>>1728744
>testing/engineering circuits
Basically. Suppose you've got a memory chip and you want to figure out its timing characteristics, like the minimum length of pulses on the address bus have to be in order to get the correct data. Or how it responds to excessively low or excessively high voltages, or if there's weird shit happening with the signaling, like long rise or fall time, or weird shit with harmonics.

>>1728741
No.

>>1726960
How hard would it electrocute if water got in? I believe that much amount of water has got to have some resistivity to it.

>>1728788
pretty hard, especially if you're unlucky and the hot wire touches it first (assuming 110v country, in 220v both wires are hot)
but it's hard to say, depends on resistivity of the water, the human skin and capacitance to ground

>>1728793
>220v both wires are hot

Whoa... what?? That's not true is it?

I live in NA, and I can get 220 across alternate kitchen sockets (each one is on a separate circuit) where both are "hot" with respect to ground but I didn't think that was true of 220v countries. Otherwise, there would be little point in making them brown/blue colors would there?

>>1728806
Yup
I was surprised too when I measured it. But at least over here (undisclosed location in a third world country) it's 110 to ground on both wires.

So theoretically if I didn't have a ground fault breaker I could run American appliances by connecting one of the wires to ground

>>1728621
>babby chip designers in university

Okay, I took VLSI design back in the day. The classes were held in the research park which was a big PITA. Eventually you'd get to make your own chips (after about 2 or 3 years). Yeah, I dropped out before that.

There are lots of places that can do small runs of things, but those are mostly the research divisions of companies and universities that inherited old-process equipment (homotaxial, huge process size). Apparently IBM can still make powerPCs at their research labs in small quantities after they got rid of their east fishkill fab.

For analog stuff, they used to spend a lot of time dicking around with prototypes. On pic related, those things that look like transistors are called "process parts" where they take wafers, drop the weird transistors they can make on them with their process, and stick pins on them. Nowadays, you'd do a lot more simulation. Fucking up is very expensive.

>>1728824

And theoretically I can run 220v stuff across the two 110v circuits I have. All without the expensive 20 kg transformer!

Jeez... don't let them find out that we defeated the region locking for small appliances.

>>1728471
https://electronics.stackexchange.com/questions/7042/how-much-does-it-cost-to-have-a-custom-asic-made

The IC development something something lab at my uni had a batch of analog ICs made it was arround 1 million Reais which at the time would be 400k trumpos

>>1728824
>anon discovers single wire earth return

>>1728831
>>1728831
Also, assuming you have everything ready for your IC. If you ship that to a foundry, the only way you get 1 or 10 ICs at your doorstep and not a boxfull is if you ask them to throw the rest away, because the way the process is structured, YOU CAN'T JUST DO ONE. Anyhow I'm a shitty 8th semester EE and ICs are not my area.

>>1728831
what did the IC do?
did it help in making sopa de macaco?

>>1728833
No, I knew about it, but I thought we had one neutral and one hot wire like in north america, so I'd get 220v to ground

>>1728837
From what I understood the page you linked says they generally only manufacture one or two wafers per mask set
>>1728843
kek

>>1728823
No it really fucking isn't. At least not where I live with polarised sockets, could be different in euroland but I doubt it.

>>1728843
Brazil, Yes.

I really don't know, from what I know of them they usually wait until they have a lot of projects done, make some guidelines and put them (no shared grounds, supplies etc, make them like little hotel rooms) on the same order and wafer. Sometimes parts of packaging are done in house sometimes it comes from china.

Can someone pls explain to me how the LED strips work? They have 4 wires going in, I presume ground, power and two signal wires? So how does an individual LED know how to light up with a specific color?

>>1728852
Sounds like it was uma delicia whatever they were doing.

>>1728854
you have red, green and blue LEDs, and red, green and blue wires coming out. Seems pretty straightforward to me..

>>1728854
Three signal wires.

>>1728854
That one isn't individually addressable so it's just GND, R, G, and B lines. Give each of them 0-12V (analog or PWM, there are pros/cons to each) and you're good to go.

5+5m of these led strips just arrived, this is going to be fun. Still need to get the driver board done, and then the ugly gloominess of dark winters shall be banished. I love how the different color temps show in the pic

>mixed color temp light sources

>>1728867
>you have red, green and blue LEDs, and red, green and blue wires coming out. Seems pretty straightforward to me..
See the strip in the pic I posted? It's actually RGB+W. Each of the colors is addressable. I don't see how that's 'straightforward' at all.
>>1728868
>Three signal wires.
So how does it address the W?
>>1728869
>That one isn't individually addressable so it's just GND, R, G, and B lines. Give each of them 0-12V (analog or PWM, there are pros/cons to each) and you're good to go.
OK... what about the W?
Also, would it be possible to make each LED addressable?

>>1728870
>5+5m of these led strips just arrived, this is going to be fun. Still need to get the driver board done, and then the ugly gloominess of dark winters shall be banished. I love how the different color temps show in the pic
I'm the guy who asked the LED strip question above. Which set of LEDs did you get? Also, which controller did you get?

>>1728878
>It's actually RGB+W. Each of the colors is addressable.
I dunno then. Put 5 volts across various pins and see what happens.

>>1728878
Well if it's S0,S1,S2,GND then you've got more than enough bits to address four LEDs.

Though honestly the way I'd expect it is that it's R,G,B,GND and the way you get white is by turning on the red, green, and blue LEDs simultaneously.

>>1728881
says 12V on the strip my friend.


>>1728854
Nope. If those are chinesium, they probably aren't RGBW. Also, the chinesium strips don't always set the R to red wire, B to blue wire. "White" is just turning R, G, and B on all at the same time. You have RGB.

Take a macro photo of the LED diode, high res, so we can see it clearly. And post what you bought.

>>1728884
>says 12V on the strip my friend.
So? Put 5V and if it still doesn't work try 12. You're not gonna break it by running it low.

>>1728884
>Also, the chinesium strips don't always set the R to red wire, B to blue wire. "White" is just turning R, G, and B on all at the same time. You have RGB.
based china
reminds me that i need to post in /csg/ more often

>>1728882
>and the way you get white is by turning on the red, green, and blue LEDs simultaneously.
no. it has a separate W. there's 4 LEDs in a single chip. you can never get pure white by using RGB only.
>>1728884
>Nope. If those are chinesium, they probably aren't RGBW. Also, the chinesium strips don't always set the R to red wire, B to blue wire. "White" is just turning R, G, and B on all at the same time. You have RGB.
retard. the post.

>>1728888
Just fucking power the thing and see what happens. Or do you not have a power supply?

>>1728891
I do. It works fine. How would powering it on explain how it works? I don't have an oscilloscope so I can't decode the signals so I'm asking on here.

>>1728888
>retard. the post.
That's a totally different LED than the one OP posted. You OP, retarded, or both?
the 12v strips aren't addressable, so there's no data lines, it's just one line per color.
now again, post what you bough so we can look at the listing and laugh at you for being suckered by chinesium marketing strategery, but also help you light your little blinkies

>>1728894
You have four wires. What happens when you put +12V across wires 1 and 2? 2 and 3? 3 and 4? 1 and 3? 2 and 4? 1 and 4? Do you get light? Do you get the desired behavior?

>oscilloscope
The fucking thing isn't an analog computer that's going to turn on different LEDs because you sent it +12, +5, -5, and -12.

>>1728897
you're a moron. go back to watching that retard on YT so you can get even more retarded.

>>1728899
post the fucking listing or shut up and get out

>>1728901
it's a generic 5050 LED you mouthbreathing retard.

>>1728902
>5050 LED
>four wires
I don't think so, Tim.

>>1728903
you're even dumber than AvE.

>gets BTFO
>hurr durr every youtuber is le dumb

>>1728899
Nigga i have all these variations in front of me right now, and this is just at my apt. You absolutely did not buy what you thought you did if they match that first pic.

>>1728903
holy shit retard.
5050 just means 5mm x 5mm. it's the package size. And most have 6 pins. Not all pins HAVE to be connected. You are dumber than dirt.

>>1728909
Reckon your W pin isn't connected then lad
KEK

>>1728911
you are confused as to whom you are replying. not my LEDs, I'm trying to help t. retard

>>1728908
I'm not your nigga.
And you're a moron. These are clearly 5050s with RGB+W. And If I choose, individual R, G, B and W on a remote, they light up different areas of the cheap. Which means you're a retard.

>>1728854
>>1728888
>buys chinesium
>gets BTFO
>lashes out
i fucking love it when the nobeards post

>>1728916
>chinesium
quit samefagging and go back to watching that retard on YT.

>>1728915
Where's the W pin if that's RGBW? There's not even a conductor visible there. Even presuming you've got a white LED on that chip, what you bought doesn't have the wiring for it. It's just left unconnected.

>everyone who says chinesium is a samefag
he's on the ropes boys

>>1728837
>Anyhow I'm a shitty 8th semester EE and ICs are not my area.
the value of a degree, the post

>>1728852
yep, same idea as pcb pools

>>1728870
not sad. one such case!
don't feed the leavers

>>1728888
as pure (((white))) as you, sojaboja
stop trolling and get back to work

>>1728907
there's maybe three that aren't

>>1728919
>being this mad about chinesium

>>1728917
You've shown how dumb you are at this stage. You got bamboolzled and you are literally blaming people trying to help you. You being a dumb-ass isn't going to make those lights RGBW, tho.

>>1728915
measure the voltages on those pads, anon, when you select white.
> lights up different areas in the 'chip'
you are fucking dumb as fuck, seriously. Post a pic and prove us wrong.

>>1728927
He doesn't even know how to select white. He won't even power the damn thing up to show how it works. I'm honestly thinking he's a troll at this point.

Do any of you nerds have opinions on potentiometers?

I am looking at 12-15 volts input and somewhere from 100-200 ohms base resistance and 2 series pots of both 1k and 10k feeding a capacitor on a 555 circuit. (so I need a 1-3 watt potentiometer to handle that range, which I understand limits my choices)
I can also use 500 base resistance, a more expensive capacitor and use 1/2 to 1 watt pots if that would be more reliable.

Target frequencies are ~100hz to ~30khz, I do not care about the exact values so much, hence why it is adjustable.

Is there a particular type of variable resistor suited to this?

I have it narrowed down to hot pressed carbon and conductive plastic, I have scoured the audio nerd forums and I care more about the sharpness of the carried impulses more than specific fidelity.

Most pots are rated for 1000 hours under load, is it going to be worth the trouble for me to invest in some high end ones, or should I buy some bulk cheap stuff and be prepared to swap them out? I have also considered using a rotary switch and some fixed resistors for a portion of it, there are many ways to skin this cat.

Once I have the circuit dialed in I may have to just get fixed resistors for the task, but if I could keep it adjustable and reliable all the better.

I plan to run this circuit continuously and the potentiometers are the bottleneck for part ratings.

tl;dr what variable resistors last the longest when pulsed continuously with about 12-15 volts?

>>1728932
>1W resistors
>for a timer
use a micro, boomer

>>1728940
No, I need to build this thing to withstand fairly powerful transients because of what I am doing with it, the lower the base resistance I can get away with before the timer capacitor the stronger and sharper pulses will get through.

For you sir:
https://vault.fbi.gov/nikola-tesla

https://www.newyorker.com/news/amy-davidson/donald-trumps-nuclear-uncle

https://www.sciencedirect.com/science/article/pii/001600329390308H

On Light and Other High Frequency Phenomena
The Problem of Increasing Human Energy
Nikola Tesla

>>1728879
> LEDs
yuji

> controller
homebrew; basic esp32 and couple of mosfets with gate drivers.. board is still in the works. all will be powered by off-the-shelf meanwell powersupply

>>1728920
>the value of a degree, the post
>tfw classmates that NEVER failed a class ever
>just rote learn all the way
>meet them later in the course, can't do jack shit and never soldered anything
all of my rage

>>1728965
also
>exchange to another country next year
>all my junk
>all my equipement
RIP

>>1728966
find someone shipping immitation crab meat and hide it deep in the back of a shipping container

>>1728932
> Is there a particular type of variable resistor suited to this?
Use a Rheostat instead of a pot. I don't think the extra inductance will impact you enough to need to compensate at AF frequencies.
>>1728965
I know a few EE's who do nothing related to electronics, or their degree for that matter.

>>1728461
I tried this one and I liked it. It leaves some sticky residue so I do clean my boards after using it. But it is the most aggressive type I've ever tried. I bought it on ebay. (It is in the middle next to my food in the picture).

ok so the pulse generator isn't what I need. what I need is an arbitrary waveform generator with modulation. basically a digital FM synth, like a DX7, but for electrical current that I can pump into certain substances. what generator will give me as much control over its signal as a musician instrument? can I somehow use the synths ability to generate waveforms to control an electrical signal?

>>1728997
It is called a D/A converter with a power amplifier.

>>1728977
symmetric dual keks

>>1728997
>implying musicians have that much control over the output
a power opamp. back in the day, the LM12 would have been the shit. today, maybe look up some of the stereo amplifier modules by Sanken

>>1728977
>Heineken
fuck that shit

>>1728998
Or a full fledged inverter for tha tmatter.

>>1728977
>>1729004
>they don't brew their own at home
grim

>>1729017
right, this is a diy board after all. we should brew our own flux, too.

>>1729019
isn't it just fucking rosin? get you some pine tar and boil it down like a man.

I so got used to STM32 chips I didn't know there were modern 32 bit micros with NO internal flash. Like some NXP chips. Wtf? How do you even flash them without flash lol. Do you first need to buy external flash (possibly from the same vendor, right?) and attach and configure it before you can flash them? Or whats the deal. Perhaps thats some kind of trimmed "budget" versions of the same chips that actually have flash?

>>1729031
Read the datasheet and app notes. Getting a controller to work with memory is not going to be any harder than connecting the address bus to some ROM chips.

>>1729033
So basically I just wanted to research what are the fastest MPUs currently on the market that are easy to work with , i.e. not in a BGA package and preferably should have flash memory. Apparently my best bet is STM32H7 at 480Mhz.

>>1728932

simple. just use a mosfet, and control it with a potentiometer like everyone else in the world.

Like, when you have a mechanical relay and you don't want it to burn out any more you get a solid state relay.

>>1728972
>Use a Rheostat instead of a pot. I don't think the extra inductance will impact you enough to need to compensate at AF frequencies.
Inductance interferes with the waveform I am trying to generate.

>>1729052
>simple. just use a mosfet, and control it with a potentiometer like everyone else in the world.
I don't have any loose MOSFETs, I do have a bunch of other stuff.

If you know anything about MOSFETs, what can you tell me about driving two n-MOSFETs source to source?

Assume I learned electronics from a boomer a long time ago and that my aims and methods are based on my redneck environment, I can order shit online, but it needs a good reason as opposed to using what I have.

My understanding of n-MOSFET behavior is that both sides will activate from a positive voltage pulse on two gates tied to a common source and give a really low combined on resistance, is this correct or will I get the voltage drop from the intrinsic diode on one side or the other no matter what when using them back to back? Discussions I have seen are unclear on this point, some claim a low total on resistance in both directions when used as a bidirectional switch, others claim the intrinsic diode voltage drop matters.
I am entirely uninterested in p channel anything as it is unsuitable for harnessing the aether in the manner Tesla described, they could be used to make a bootleg nuclear fission reactor, but the inverse is way more interesting.

>>1728685
Yeah I have considered just changing it but I've already put a lot of work into the 2mm round projectile aspect (the coil gun I mean), and I'll be honest I'm a little to lazy to change all of it at this point now that I have the coil gun just where I want it and a lot of parts I've made or ordered are centered around the 2mm. I had the idea to make the plates a little thicker, keep them 2mm apart and make up the extra space a the top with some type of plastic. I really appreciate the help, learning a lot about metals and manufacturing, learned my lesson on planning mechanically more before hand too.

>>1728998
I need the frq at min 40Mhz

>>1729067
>I have junk, what do?
don't be this guy

>>1729086
prove it

>>1729050
I.MX RT10xx have some 500MHz options in QFP. shouldn't be much harder to work with than anything else
but you're getting to the point where leadframe inductance really starts to matter. worth considering taking that step in fabrication technology

>>1729022
>just
not just. for harder jobs it needs to be mildly activated. so add almonds

>>1729116
>that image
oh god, i'm dying

>>1729116
>I.MX RT10xx
Yeah thats what I found that didn't have flash, maybe some others do but it doesn't matter to me at this point... because
>you're getting to the point where leadframe inductance really starts to matter. worth considering taking that step in fabrication technology
hehe i was just wondering about that. in the past, i was able to get away with incredibly sloppy prototypes based on <=100Mhz chips with literally zero precautions. well maybe, just making sure grounds were all connected to a single point and all decoupling capacitors were in place, but thats about it. And I never had any stability issues. But apparently ~500Mhz is a different story... And also, kind embarrassing to admit but a few years ago I coded my own HAL from scratch for F3 and F4 chips. Why? Just for the heck of it. Stuff was pretty straightforward and it was a good learning experience and a fun project. But now. looking at the programming manual for H7... Holy shit. It is 10 times more complicated than even F7.

>>1729067
A lot of inductance sure, just cancel it out with some C somewhere else, once you figure out what frequency the thing is running at. Rheostats are already wound to minimize inductive effects- waveform shape matters a lot less than power handling, since at most you're going to get some dampening effects.
>I am entirely uninterested in p channel anything
Don't tell him that nearly all semiconductors contain both P and N type material, and in fact need it to function....

>>1728850
Then what? The multimeter was playing a prank on me? I happened to conjure 110v out of nowhere? Fuck off retard.

>>1728932
Use a power transistor in common collector mode

>>1729199
>undisclosed location in a 3rd world country
I'm guessing you weren't grounded properly or the electricity was otherwise hooked up out of spec.

>>1728932
>and I care more about the sharpness of the carried impulses
I don't understand how bandwidth works. A sharp 30khz square wave could need a bandwidth up to the megahertz

>>1729022
Rosin leaves lots of residue. No clean flux doesn't. Dunno what they make it of. Also like the other anon said, you need some kind of corrosive agents sometimes.

>>1729022
To be honest I have something along the lines fo pine rosin gummed up on the side of a part of my deck. I've scraped it off when it's hot enough to be malleable and used it for soldering before. It's not as good as my RMA, but it works.

>>1729221
No-clean still leaves some residue, it's just not as corrosive and not sticky. Ace solder reworkers still clean it off. And cmon, who doesn't have a bottle of IPA lying about?

I'm trying to make a temperature controlled heating element.

I want to use a POL converter to deliver the power, and use the PMbus protocol to control the temperature, which will change the voltage +/- 25% with nanosecond delay

Here's the idea:
1 Start with setting the temperature resistance coefficient dataset of the desired heating element
2 discharge a known voltage value into the element using DAC, ADC read the voltage to calculate temperature
3 offset/adjust for ambient temperature
4 set the digital potentiometer connected to trim pin to the appropriate resistance based on the PMbus adjustable range to maximize range
5 power on, PID algorithm (because i don't know anything else) using the measured voltage values from PMbus and ADC
7 consider digipot, trim, and PMbus response time to change voltage range as needed

Tell me it's shit

>>1729232
I don't understand half of what you said, but are you trying to calibrate temperature by using the changing resistance of the heating coil?

>>1729233
I'm measuring the voltage before and after the element, and the amperage leaving the DC DC converter to calculate the resistance of the heating element.

The converter has a built in voltage readout, but it has ripple and I'd rather read it out through the higher resolution external ADC

>>1729210
The ground is fine, a rod buried about a meter and a half. Had to install it myself because the house didn't have any. Don't remember how thick the wire is but it's fairly thick. Granted I didn't measure the electrical characteristics of the ground, but as the input impedance of a digital multimeter at mains frequency is multiple megaohms, even burying a probe in the dirt should suffice to take a voltage reading, and I confirmed the reading by measuring it on another house too. I probably made a mistake assuming all 220 grids work that way, true, but apparently it's just how they run things here (Uruguay). On an unrelated note, at the previous house I lived at voltage at peak hours went down to about 180v, which caused my outdoors led lights to dim a lot. Safety is not very good, just this year there was a guy disfigured by arc flash not to mention this https://en.m.wikipedia.org/wiki/2019_Argentina,_Paraguay_and_Uruguay_blackout

>>1729209
Nope. He'd want common emitter to not have the control current go through the circuit.

>medium PICs
>10 bit ADC
>8 Bit registers
top leek

>>1729221
rosin (non-activated) residue can be left on without cleaing. it's just sticky
>no-clean
usually, fractions of rosin (esp. abietic acid) or other organic acids

>>1729228
>RMA
you didn't mildly activate your rosin. some halide salts are good for that. just not NaCl
>a bottle of IPA
sry I drank mine

>>1729239
>nanosecond delay
>bit times in the microseconds
your idea is a meme

>>1729242
not actually a problem, brownouts give you some time to enjoy the dude lmao

>>1729315
Fucking ZnCl2 is best electronic flux. Just don't be pussy and wash it away.
Liquid damage board? No fucking problem, shit is strong to remove any oxides.

What's the difference between rosin and resin flux? Why is "organic" flux harder than rosin flux? Isn't rosin an organic substance to begin with?

>>1729324
Doesn't it remain inside the joints?

>>1729328
Not an issue for through-hole, or some SMD packages.
You won't be using it for BGA, obviously

>>1729324
this is how incels fuck
this is also why incels don't fuck

>>1729327
1:literally just the spelling. 2:not always the case. 3:organic acids.

>>1729337
>1:literally just the spelling.
Doesn't seem the case. For example see this flux's datasheet http://www.inventecusa.com/assets/nc-559-asm.pdf
It says ROL0, which according to https://www.eptac.com/wp-content/uploads/2007/09/webinar_eptac_09_19_07-2.pdf?__cf_chl_jschl_tk__=5dfeb55b3b7ec3bbcb0802b7b034c25afc7eae26-1575640958-0-AQDC5A11fB_mk_7xRLVvjYabTVRBKIEpthLLraXwq3k_cgPagtN_FL0K_45D44odqFoP0xVCF9zRYkoCxIA9Od9TzIMPTlG8oqyWqgpmuAM5Ly1sqJW0e0mCtBM_F2NpV3JYCGMXjE4ngeLfsQz_EEzCofUVeCeEVgwHA2RPd9vT5H5u1VnDg3aIs7bxFOIk5AvcwrwuYlNv3_rh_KiI_cXL0r-Se_IZpsjs4CFFYcW5ahowFgckNlSeZAiTDQjibRwdrq02wSEistMcHau2vDQPTln5VS9Bkj0jrlWrcIpOjCaJiQJ_sdBPXwbDTMSh1NtchfODNhdcS01n61fx2EUjTzwsTenho6WyVIo6pYezBjAH59PlHRk3lkAL52F47R3_z-uLjiO8ALfY4otT9Po
page 12, means it's a ROsin as opposed to REsin flux (REL0).
>2:not always the case.
Ok but then it seems whether a flux is organic doesn't mean much.
>3:organic acids
According to >>1729315 the main active agent of rosin is an organic acid, so all rosin based fluxes are organic then?

>>1729353
>all rosin based fluxes are organic
OA means it's not plant sourced, so it's not organic in the foodie sense
have a recipe with some discussion here
https://northcoastsynthesis.com/news/homemade-traditional-resin-flux/
tree sap + citric acid scraped right off your Sour Patch Kids is probably a good flux if you're innawoods and need to make a thing

Any advice on testing whether a relay is sticking or poorly functioning? Is it enough to just check that the relay cycles in response to current, and that there's continuity when the relay's supposed to be closed, and no continuity when it's supposed to be open? Or are there other electrical parameters worth checking?

>>1729475
current carrying and arc quenching are important too, but a failure is pretty easy to see.

>>1729475

connect a load near the max capacity, turn it on, and measure voltage across the contacts. if it's anything substantial, like a volt, then the contacts are dirty and will get hot and die soon.

Can't post this, so had to take a screenshot.
Any idea what word could potentially trigger their buggy algos?

>>1729515
Maybe SPI/I2C? My thought is it might be a partial match to a commonly-spammed URL variation.

>>1729517
Hmm guess not.

>>1729315
Literally doesn't know what bit time is
>you are a meme

>>1729515
part of the justification for 4+ layer boards is solidity of the power plane, especially (but not just) when multiple rails are involved. when you interrupt the power plane fill with a trace, you are turning that part of the plane into a hairpin curved wire aka a loop antenna. any power supply dI/dt carried by that loop antenna will induce currents and shit all over your local electromagnetic spectrum, which isn't acceptable for a production device. also, you are also adding inductance to any currents carried over that bent part of the plane which means your Vdd/Vss aren't going to be exactly equal. if you're using analog circuits onboard the big chip, they'll be electrically noisy
your power is still an external signal. you've got the right idea keeping the caps as close to (or under) the power pins. just be careful of current flows and keep it solid

>>1729515
Also you have to think about the harmonics, not only on the first. A square wave has significant power in the harmonics, the faster the edge, the further it'll go.

I'll have to live with someone that has somewhat severe anxiety, panic and PTSD for a while. One of the things I've noticed is that she asks me to plug her laptop charger because it can cause some sparks and make that "pop" sound and that makes her anxious. Anyone know what I can do to fix the inrush current problem without affecting performance that much?
>blublublu irrational fear tell her to get over it
Well, I tried. But if I can do something to help I guess I should.

>>1729315
No I mean the tub of store bought rosin I've got is almost certainly mildly activated. And I'm not going to start burning halides with my iron any time soon. And I'm unsure how well polar salts would dissolve within IPA. so I'll stick to organic activators.

>>1729575
>I'll have to live with someone that has somewhat severe anxiety, panic and PTSD for a while. One of the things I've noticed is that she asks me to plug her laptop charger because it can cause some sparks and make that "pop" sound and that makes her anxious. Anyone know what I can do to fix the inrush current problem without affecting performance that much?
Put switched outlets in her place like they have in the UK. That way she can plug something in while the power is off.

>>1729616
>switched outlets
How are these not mandatory in burgerland

>>1729627
It's really not that much of a risk. I have only been shocked once in my life, when I was a child, and I somehow managed to touch one of the contacts when I was plugging something in. It has never happened again. It hurt a little, I cried, and that was it.

GFCI outlets have also more or less eliminated the risk from having wet hands when plugging things in.

I wonder if the fact that we're on 120V makes a difference.

>>1729627
I'm in south america and never saw one in my life. Maybe I'll add a switch to the charger instead, might be easier.

>>1729650
Nvm, these are quite common because wiring is passed through brick and mortar walls and adding new outlets/switches is not trivial. So they are used to switch lights and still connect something but I guess it's trivial to connect the socket in series with the switch.
Thanks anon.

the wiki article. on diodes has left me with more questions than answers. does any1 have a book rec for the diode design function? something with actual math

>>1729660

>>1729616
>polar salts
my mistake, it's not the salts you want, those are already there. citric acid is good. compare to glutaric acid, which is used in some commercial no-clean fluxes

>>1729639
actually, it was capitalism

>>1729660
the Shockley diode equation wasn't enough?

>>1729662
Pic related

>>1729662
It's simple. When A and B are at 5V, no current can flow through the diodes, so the pullup resistor also has no current going through it, meaning the voltage across the resistor is minimal. Hence Y is 5V. When any of the inputs A or B is set to 0V, current will flow through the resistor from 5V through the diode to ground, meaning Y will be at ~0.7V (1 diode drop). Adding another diode in parallel with this by pulling the other input to 0V doesn't make a big difference and the output Y remains at about 0.7V. Important to note is that the device feeding the inputs A and B needs to be able to sink current, no need to source it, so you could use an open-collector output of a comparator or something.

>>1729682
they're diodes. a push-pull source will be fine

>>1729684
I'm not saying you need to use a common-collector, just that if you don't want to bother with a pullup you don't need to.

>>1729554
just checked the prices at jlcpcb 5pcs 10x10mm

>2 layers: $2
>4 layers: $30
>6 layers: $87

>>1729678
>>1729682
I guess I'm looking for the physical chemical side of it, the material science, not just the electrical. I know some of the general chemistry involved, but not the electrodynamics(?) particularly of the crystalline materials and precisely why it is certain substances are chosen over others for certain applications and how the complexities of the structures of those substances relate to the function of the electrical components

>>1729700
No shit it's more expensive. If you're that worried about EMI shit and conformity, go read Henry Ott, Electromagnetic Compatibility Engineering (2009) and see if you can solve your problems without multilayer boards.

>>1729702
Oh, you want to know how semiconductors work?

>>1729700
Are 2-layer PCB's the $5 roast chickens of JLC?

>>1729639
>It hurt a little, I cried

Jeez... I touch mains all the time, even when I was a kid. Even plugging in my hot comb (it was the 80s) with wet hands into the bathroom outlet (with no ground or gfci, again, it was the 80s). We used to do it for fun in our electricity classes (preparing us for blue-collar work in high school).

I touched live 120 on electrical outlet that I was replacing the other day that I forgot to shut off. No biggie, just carried on.

I wouldn't put it across my chest, and I sure as hell wouldn't ever want to touch 240 tho

>>1729710
yes, and they're every bit as delicious

>>1729700
I mean, unless you're using the onboard ADCs or comparators or shit, it might not even matter. just digital should be fine as long as it's decoupled well

>>1729721
>he's never gotten poked by 480
virgin

>>1729554

Ha, open an old TV, you'll find SINGLE sided paper boards riddled with jumpers, and they're running frequencies on the order of 200 MHz. Maybe they are just that much smarter. Doing more with less.

>>1729721
This, what a bitch.
120V doesn't even really all that painful, it's startling if you accidentally contact it, or dare I say, SHOCKING... I'm sorry I'll stop. But it doesn't really hurt much. Would probably be more painful, and more lethal if you solidly gripped it and held it before for accidental contact it's not bad at all. I had two incidents, one where I brushed up against the live side of a main switch in a piece of test equipment and one other time where I was unplugging equipment at my old job with a frayed cord and accidentally touched the exposed wires. It startles you but I'm still alive and no worse for wear.

>>1729708
yes

>>1729726
>those trippy hand drawn curvy traces tho

>>1729726
>implying I haven't
ok zoomer
>and they're running frequencies on the order of 200 MHz
you're about two orders high by the time the signal gets out of the tuner can. also, televisions are not a quantitative device. a 5ns blip isn't even going to make the picture tube blink but might well make the micro reset. context, you autist

>>1729725
>>he's never gotten poked by 480
>virgin

No, cherry popped at a young age. That time when I was about 10 where I was going to make a "jacob's ladder" so I hooked up a 8v transformer in reverse. Everything looked fine (I thought it might explode). Then I reached across to grab something and my arm brushed across the wires. It left this cool white skid-mark about a foot long on my arm (when I got up off the floor to examine it).

It's weird, it took a long time (like 20 minutes) for the sensation to go away. It's kind of like when your arm goes to sleep and it't gets tingley.

for me lads? it's shorting out a 2.4 kV motor starter through my cat 4 600V meter

>>1729703
Nah I am not. Just exploring different options.

>>1729751
Imagine the smell

>>1729749
>No, cherry popped at a young age.
hmm
>I hooked up a 8v transformer in reverse
oh

>>1729749

What do you guys think of Port Mapped (In/Out) IO vs Memory Mapped IO?
I'm considering cannibalizing my In/Out instructions for faster load/store instructions that have short immediate offsets for faster stack operations.
Since I'm planning on doing this with 74 series logic, I want everything to be as fast as possible, but on the other hand I'm debating if this is even a worth while change to make.
Currently all register-register and load/store operations take a single clock cycle to execute, or two if the use an immediate value.
That's already faster than most TTL CPUs, which tend to be implemented with 4 to 8 clock cycles per instruction.

>the virgin port-mapped I/O
>the chad memory-mapped I/O

>>1729820
Hey, when you only got 64k of address space, every byte counts.
Having the IO memory space allows for that much less main memory to be taken up by peripherals.
Still though, I guess I could just map a bus bridge somewhere in memory.

>>1729824
I'm playing with an Apple //e lately, the memory-mapped I/O is fine there. Just build a form of bank-switching into your system and it'll be awesome.

>>1729832
Yeah, I figured I was going to need to incorporate bank switching sooner or later.
I just kinda want to keep the memory space as flat as possible.
What do you think a good way to lay out the banks is?
I'm thinking of dividing the memory into four 16k chunks.
the lowest 16k will be ROM containing the OS, the next 16k is fixed, and the upper two 16k chunks are dependently bankable
I'll probably redirect $FF00 to $FFFF to memory mapped IO.
Oh well, if the whole thing becomes to burdensome, I'm planning on a 32 bit extension to implement on a FPGA anyways.

>>1729818
I don't see any point to a separate I/O space unless there's some major reason for it, like actually split buses. the 6510 didn't have any problem stealing the first two bytes of the memory address space to GPIO. is this addressing mode applicable to all register indirect loads/stores, or just sp only?
besides, if you do want a separate I/O bus, you probably wouldn't need or want things like add and subtract and you surely wouldn't need a full 2^16 address space. a more limited access paradigm of reads, tests, and writes of bits and bytes is more in keeping with the idea of an I/O bus
>Hey, when you only got 64k of address space, every byte counts.
meh. but 256-byte pages do add up

>>1729840
>I just kinda want to keep the memory space as flat as possible.
usage would determine that. will you be using overlays? will they be loaded from disk or ROM? will you be running gigantic sieve algos? another look over CA:AQA would be advisable. let the program's needs guide you
off-the-cuff I'd think about a fixed region at the top and/or bottom of memory (wherever the reset PC is, so that you can keep the bank switching continuously enabled). maybe 4kB, with an I/O window carved out of that and the rest of the fixed region used as a kernel ROM and system service jump table. the other 60kB area would be a single window (maybe two if I could justify it) which could be mapped anywhere
but also think about whether and how you would get to the lost 4kB in each bank. you might have enough nanoseconds left to insert an adder onto the address bus, for a stripped-down Intel-like system with a single shared 60k code+data segment. this is nice because you don't have to have your loader consulting a table poking relocation entries in your program text, nor your assembler constructing that table for the loader's use

>>1729840
>the lowest 16k will be ROM containing the OS, the next 16k is fixed, and the upper two 16k chunks are dependently bankable
The way the Apple II works is that the lower pages are for specialized purposes (pages 0 and 1 are both used in the architecture for specialized tasks, page 2 is used for the keyboard input buffer, part of page 3 is used for interrupt vectors), one set of screen buffers are from 0400 through 0BFF and another in the 2XXX-5XXX range, I think. User code and the OS reside someplace between 6000 and BFFF. I/O is in the CXXX range. Everything from D000 to FFFF is ROM that can be bank-switched to RAM, with some bank switching in that area (iirc, some of the most important built-in routines are in the FB00 and up region). See pic related for some info on how they organized it.

It's actually a pretty neat system that's got excellent documentation. It's worth looking at for an example of how it's been done previously.

>>1729840
>>1729858
>16k will be ROM containing the OS
an even better target for stealing 256b for peripheral address space, giving a 48k user code+data segment. which you could split if you prefer, but think of the children first
speaking of children, are you having hardware IRQs? if so you'll need to think of how to treat them. my initial ass-monkey-tier idea would be to force them to live in a dark closet in bank 0

>>1729863
agreed. but maybe systems with video aren't such applicable examples if his console will be a UART (?)

>>1729873
>are you having hardware IRQs? if so you'll need to think of how to treat them. my initial ass-monkey-tier idea would be to force them to live in a dark closet in bank 0
You could probably also put the bulk of the IRQ handler code in another bank and just have a short IRQ vector table somewhere reserved in non-switched memory that hits the right bank switch right before JMPing.

>>1729858
It's not a completely separate bus, they would share the same address and data lines, just with an extra MEM/IO signal to select which devices should respond.

>a more limited access paradigm of reads, tests, and writes of bits and bytes
I've only got a single memory addressing mode, register base+offset, In/Out would simply set the Mem/IO line high, while load/store don't.
The idea being that the whole video buffer and other IO devices could be confined to the IO space, leaving main memory free to just be RAM, and without slowing down access to the video buffer by having to go through a bus bridge.

>usage would determine that.
Right now just a simple 80's style 'Home Computer', nothing serious, but I would like for it to capable of doing more than simple text mode stuff.
Once I get closer to a final design I'll probably try porting Wolfenstein 3D to it.

I get what you're saying about segmentation, but that kind of is the one thing I absolutely want to avoid.
Besides that, I don't have any opcodes left to implement any segmentation.

>>1729863
I'll have to read up more on the Apple //e.
I've looked a bit at the C64 and GBC so far.

>>1729873
>16k will be ROM containing the OS
>an even better target for stealing 256b
Oh, true.

>speaking of children, are you having hardware IRQs?
That's the last big thing I need to finish on the CPU.
I have a rudimentary IRQ interface, but really crashes and burns if a higher priority interrupt tries to pre-empt a currently running lower priority ISR.
It's still an early work in progress though.
>my initial ass-monkey-tier idea would be to force them to live in a dark closet in bank 0
Yeah, I was just going to put it in the first 128b of memory on bank 0.
User interrupts get handled first by a ISR in ROM, then they get forwarded to a table in the bottom of the User RAM.

>>1729876
I'm not too worried about memory consumed by the IVT and ISRs.
I don't have anything elaborate planned at the moment. Not a bad idea though

>>1729882
>I'll have to read up more on the Apple //e.
>I've looked a bit at the C64 and GBC so far.
Check these out: http://www.applelogic.org/files/AIIEREF.pdf
http://www.classiccmp.org/cini/pdf/Apple/Apple%20II%20(Redbook)%20Reference%20Manual%2030th%20Anniversary.pdf

There's a good chapter on hardware implementation in the first and some good schematics in the second.

>>1729882
What architecture are you using?

>>1729876
>when the IRQ vector hits that bank switch just right
possible. esp if the lower bank is fixed and the ISRs in ROM do the user ISR some favors before/after calling it like a subroutine

>>1729882
>just with an extra MEM/IO signal to select which devices should respond.
sounds like an address bit. on the Z80 the bus cycles were different (automatic 1WS, slight sequencing differences) so it made sense to have separate loads/stores for them. maybe show opcode list, it might be time to Muntz that fucker
>really crashes and burns if a higher priority interrupt tries to pre-empt a currently running lower priority ISR.
hmm, might need to mask all ints until the ISR has moved the current mask to somewhere safe. I imagine you're not having the CPU do much or any automatic stacking when vectoring

>>1729892
>Z80
based but bluepilled
6502 4 lyfe

>>1729894
have fun with your 24 bits of general register, $01 b01

>>1729897
I've got a stack and plenty of work RAM to use if I have to juggle values around.

>>1729887
Thank you.
I love how well documented these old systems are.

>>1729888
It's a homebrew architecture I've been working on to apply the concepts I learned in senior Computer Architecture.
It's a 16-bit load/store architecture with three operand instructions.
Nothing fancy, but it's very simple and clean, I think.
It's not another Ben Eater SAP clone, that's for sure.

>ISRs in ROM do the user ISR some favors before/after calling it like a subroutine
I was planning on it, OS gets the first crack at it, then it passes it to the User.

>sounds like an address bit.
Well, yeah, I guess so, just one that's only set by In/Out.
I didn't really want to have any 'bank' registers in the CPU core, just makes for one more dependency to keep track of in re-entrant code.
I guess I'm not really bothering with any bus timing differences, since I'll just let myself be spoiled and use modern fast SRAM.

>maybe show opcode list, it might be time to Muntz that fucker
Ok, I'd be impressed if you can muntz this much more, it's already pretty dang lean.
Current Opcodes:
Add (Rd = Rx + Ry)
Sub (Rd = Rx - Ry)
LSL (Rd = Rx << Rd)
LSR (Rd = Rx >> Rd)
Inc (Rd = Rx + 1)
Dec (Rd = Rx - 1)
And (Rd = Rx && Ry)
Or (Rd = Rx || Ry)
Xor (Rd = Rx ^ Ry)
Not (Rd = !Rx)
In (Rd = Rx[Ry].io)
Out (Rx[Ry].io = Rd)
Load (Rd = Rx[Ry].mem)
Store (Rx[Ry].mem = Rd)
Compare (T = Rx ? Ry), where ? can be 'less than', 'equal', 'zero', etc
Halt (Halts the system)

Registers:
R0 - K - Constant, normally zero, but immediate values are loaded here
R1-R4 - General Purpose registers
R5 - Stack pointer by convention, otherwise identical to R1-R4
R6 - T - Test register, true if the value in T is not zero
R7 - Program Counter

All instructions may be predicated based on if T is true or false.

On the chopping block are Inc, Dec, In, Out, and Not, which I might replace with add short immediate, load/store w/ short immediate offset

>>1729908
are load/store the only instructions that support memory access? do they support indirection?
also i don't see branch or jmp instructions here

>>1729745
> about two orders high

You are right if you live in a world where everything is a perfect and periodic sine wave or square wave. Out here, though, things are a bit different. It will become clearer when you understand why you can't be an effective TV repair guy with a 2 MHz oscilloscope.

>>1729927
Correct, Load and store are the only instructions which can access memory.
>do they support indirection?
They only support indirection, in the form of register base + register offset.
You get absolute addresses or immediate offsets by having immediate values, which are represented by the K register.

>also i don't see branch or jmp instructions here
Not as such, instead you use a conditional execute instruction.
a PC relative branch would be done as this:
+Add PC, PC, K
The '+' means execute if T is true.
A unconditional jump would simply be
OR PC, K, K
K||K is K, which then gets stored into the PC.

All of this is hidden behind assembly psuedo instructions, of course.
Like I said, it's already pretty Muntz'd.

>>1729894
> 6502 4 lyfe

The 6809 is the penultimate expression of the 6xxx form, finally culminating in the 68HC11 (which many of you probably have in your basement in one form or another). It was a true joy to code.

It's a pity that just enough of the grotesque zombie once known as Motorola survived to become a troll and prevent 6809 cores from taking over entirely. The 6502 (created by Moto's incompetence) was highly Muntzed, however the modern WDC 6502 clones actually have more transistors that the 6809 did.

You could probably stuff something like 100,000 of these cores on an epyc threadripper die(s). That would be something to see... kickstarter anyone?

is kicad the bestest pcb designer in the hood?
it must be since cern uses and develops it right?

>>1729998
Well its certainly not a bad one, but it apparently has a few issues. Not that I've seriously tried using Altium or Eagle. It's by far the best piece of FOSS I've ever used in any case.

>>1730001
>but it apparently has a few issues.
such as?

>>1729964
>motorola CPU
i'll stick to my MOS/WDC chips thanks

>>1729908
hm, well
>In
>Out
you probably don't need all that complex addressing in the I/O instruction. if you have address mode bits in your instruction format, you might want to reuse them. or consider I/O space for load/store as an addressing mode
>Not
xor with ~0
also, if you have small constants, do you really need inc/dec

>>1729998
best? not necessarily. best bang for the buck? fuckyeah
CERN has definitely helped do some of the harder parts like the interactive autorouter. but Wayne Stambaugh takes FOSS to the grape koolaid level and has farded about giving KiCAD an IC design mode, as if that will make MPW any cheaper or practical, or is even a reasonable goal given the multiple already FOSS IC design packages have failed to do it. I mean, his generosity and his work is appreciated, but this is why upper management should serve at popular pleasure

>>1730072
>autorouter
i actually never used that and did all traces manually, but if the cern cucks made it then it must be really good and maybe i should learn it

>>1730072
>or consider I/O space for load/store as an addressing mode
I actually had that in the 32 bit version, but I didn't have enough space to encode it on the 16-bit version.

>Not
>xor with ~0
Yes, I know that. I added Not since it's one cycle faster than xoring with an immediate 0xFFFF.
I'm just debating if it's actually worth keeping around.

>if you have small constants, do you really need inc/dec
Currently, I don't have short constants.
I'm debating if they're worth adding, because I only have space to encode a 3-bit short constant.

>>1727515
>>1727650
How about i use a 12w adapter then?

>>1730132
>tfw commited the mistake of taking a ASIC design class next semesters
>fml
I honestly just took some random classes until the classes I want open up on the semester after that one. (Energy related shit)

>>1730147
Welding machine

>>1730178
Next semester*

>>1730093
it's not really an autorouter, proper. it's just the push'n'shove mode added to trace placement. nothing really to learn, just enable it in the routing options and push other tracks out of the way as you route. quite handy

>>1730132
but if you have an immediate 0, you can XNOR. not gonna push you on it
>don't have short constants
oh
I would say yes, they are useful. if nothing else, expand inc/dec to addq/subq (a la M68k), maybe consider doing the same for those other instructions where the 3rd op is ignored
also I'm thinking halt is a waste of 12 instruction bits, and you should de-integrate clock control into a peripheral, or make it a special case of the compare encoding

>>1730132
>I'm just debating if it's actually worth keeping around.
Honestly I'd keep it unless you have a reason (other than muntzing) to get rid of it. Opcodes that work like macros (but include a cycle savings) are a good thing IMO, even in a RISC architecture. Like on the 65C02 they included a STZ (STore Zero) that replaced a load immediate followed by a store of that value, at something like a one-cycle savings. Most people pooh-pooh it because it only really makes a difference in loops, and makes your code incompatible with the NMOS 6502s. Even so I think these are a good thing in reasonable quantities.

>>1730234

Usually — in the initialization — we detect the presence of "STZ" or "TRB" and backpatch the code accordingly so everything stays compatible. Sometimes we'll replace a whole chunk of code (like a loop) in line before the hauptprogram. If the initialization was big, we'd bank-switch it out after it was done since you only need it once.

On something like the 6502, every cycle counts.

The eevblog guys "came a gutser" because their fricking USB libraries were written by zoomer interns at STM (or wherever) and used the entirety of the MCU's memory. https://www.youtube.com/watch?v=SFrUINyYcEA.. The level of helplessness, ignorance, and malfeasance nowadays is just ridiculous.

>>1730230
>but if you have an immediate 0, you can XNOR. True, but I'd have to add a XNOR opcode, and then I haven't really saved any opcodes.
I guess XNOR has more uses than just NOT, but i dunno.

>expand inc/dec to addq/subq (a la M68k)
Oh cool, that's almost exactly what I was thinking
I didn't know if a 3-bit immediate was too small to be useful or not.

>I'm thinking halt is a waste of 12 instruction bits
I can see that. It kinda started as a general control instruction, but it turned out there really wasn't that much in the CPU to really configure.

>>1730234
Yeah, I've got a few more macro-like instructions on the 32-bit version, but the 16-bit version has so few opcodes that I'm trying to cut everything I can without negatively impacting performance.

I know I can get more opcodes by cutting the third operand, but when I tried that I had to use most of the new opcodes to make up for lost functionality from the three operand design.

Honestly though, at the end of the day, this is mostly just fussing over minor details. I've done a bit of programming with the current ISA and it feels pretty competent, I'm just trying to see if I can make any last alterations before I freeze the ISA and start building the CPU with TTL chips for real.

>>1730272
>zoomer interns
bet they're actually poos or the failsons of ukrainian oligarchs

>>1730278
>XNOR
yeah, maybe never mind that one
>too small to be useful
lol no, you'll be using it all the time. show me how you'd stack saved state at the beginning of an ISR, for example
>general control instruction
consider the following: any write to address 0 halts the processor. congrats, you've just detected many null pointer dereferences without an MMU and freed up an entire page of opcodes for further use. thank me later
those 4096 extra instructions might be a good place to encode software interrupts a la the original M68k, if you can't think of anything more productive to do with them
what's your instruction word format look like these days?

>>1730039
Well personally I can't get the 3D viewer to show me parts on the PCB, let alone having the convenience of assigning 3D models to footprints, which should be fucking automatic. I've heard complaints about it not having a few nuanced features that other products do, like about nesting sheets or something along those lines, but there are addons for kicad that might do that.

>>1730292
Here's the current opcode format:
>2 bits conditional execution control
>1 bit immediate value flag
>4 bits opcode
>3 bits Rd
>3 bits Rx
>3 bits Ry

The conditional execution bits aren't really a huge deal for the single cycle implementation, but they'll be a good performance booster if I ever get around to pipelining it, since short bits of conditional code would incur less of a performance penalty than a branch mis-prediction.
The 16 opcodes seems kind of limiting at first, but due to the PC being in the register file most instructions can perform a variety of tasks that would have otherwise needed to be separate opcodes.

>how you'd stack saved state at the beginning of an ISR
Oh yeah. An offset of 8 is just enough to save all the registers.
I guess I got hung up on functions trying to access more than 8 words of local variables.

>consider the following: any write to address 0 halts the processor.
Seems like that would be harder to recover from than just using issuing a trap on writes to mem[0].

I might just turn Halt into a reserved opcode and deal with it later.

I've been learning about electronics using music stuff. Making oscillators w/ 40106 and 555/556 ICs and stuff. CD4040s to change tonal qualities and low/high pass filters. Its been tons of fun.

But I'd like to make a little "test station" sorta thing with a mixer and amplifier. I bet a 386 IC and 7805 would be involved based on my research, but I'm too stupid to find a great way to build it.
Does anybody have a schematic or know what to search to find what I'm looking for? There's the "OMSynth" that some tutorial guy made. I'm looking for something similar to that.

Thank you very much.

>>1730470
most datasheets for chips will include an applications section, which will generally provide fully or almost-fully worked examples of how to use the chip

>>1730544
Thank you very much. That's one of those things that makes perfect sense. Still learning to recognize those patterns in this new hobby.

>>1730544
they will also include important information such as power supply voltage range. the 386 will actually be a little more performant if you run it closer to 10V than 5V but not over 12. for full details see the full datasheet
>but what about the mixer, Yogi
that's a fairly usual opamp application, Booboo. Pic related is typical

>>1730550
no prob, cheers. when starting out in this hobby, let me promise you that there is no such thing as reading too much

>>1730550
in fact, Ray Marston did a few series in Nuts & Volts magazine on different kinds of opamps. it's a bit dated and some of the part numbers aren't available anymore (unless you're lucky and chinese recyclers didn't just place a different label on some LM324 out of the scrap heap) but the principles still apply. start here
https://www.nutsvolts.com/magazine/article/op-amp-cookbook
in particular you will eventually want to read the two-parter on operational transconductance amplifiers, which are remarkably useful in making somewhat precise voltage-controlled audio functions
https://www.nutsvolts.com/uploads/magazine_downloads/11/May%202003%20Ray%20Marston%20-%20Understanding%20And%20Using%20OTA%20OP-Amps.pdf

>>1730560
>>1730569
Awesome. Thank you very much. You guy(s) are very useful. I'll remember to keep checking the datasheets. Probably the best way to gain understanding, anyway.

Which program do you guys like for circuit simulation? I find LTSpice to be clunky as fuck.

>>1730656
KiCAD has a SPICE interface now but I haven't really used it.
tried MicroCap?

>>1730656
>clunky
That's spice for you. But real-time sims are comparatively worse. I'd try KiCAD, it has a really good part library in there, and I assume most of the ICs and transistors and the like have Spice models. Not to mention the convenience of designing your circuit board in the same software.

>>1730560
Omfg, those potentiometers! Must be the Limeys again.

>>1730697
That's how you draw those.

why is it bad to have 90 degree turns in pcb traces?
is it because electrons are flying so fast they can't make such sharp turn and fly out from the trace and crash into some smd component nearby?

I’ve got a problem with a board. It’s a power board from a washing machine. It has a boost converter as the supply for digital stuff. I found the resistor in series with all the circuit burnt. I found a L7905 burnt too. I removed it, and I can get -10V, but if I replace it, with only the voltage input and the gnd pins (for testing the output of the pins) voltage drops from five to 0V in a loop once a second. I can hear the buck converter switch each time the voltage rises to it’s normal level, but doesn’t seems to work properly. The buck converter is a VIPER013HS. My first approach is replacing it, but is a pretty expensive part (15€ in ebay for one piece, it’s cheaper in digikey but I am not making an order with only one part), so I want to be sure, and know your opinions.
I’m in mobile and I’m not a native English speaker, sorry for any errors in my text.

shorting together input gnd and output gnd on a buck is fine right?

time to find out if my smol mill can handle 0.2mm traces

>>1730703
That section of the transmission line will have different impedance characteristics which will cause reflections.

>>1730750
I just love watching this thing work, carving shit with incredible precision

>>1730717
the 7905 and resistor may be burnt for another reason. if it's backing off like that, the boost is probably good. look for shorts on the load side of the 7905
>boost
you probably mean buck but ok

>>1730746
sometimes. if the buck converter is synchronous (i.e. uses a MOSFET instead of the catch diode) or has an output discharge feature, you'll fry it

>>1730697
limeys
poos, same thing but with 150% more self-absorption

Making a single-sell LiPo cutoff protection circuit with DW01 + 8205A, and just ate a bag of dicks

DW01 has a (((feature))) that when your battery is disconnected if won't turn the shit off until you either

1) charge the thing for a bit (I don't even have a changer since batteries are to be charged before inserting into the device), or
2) short GND and CELL-

https://www.eevblog.com/forum/projects/pain-and-suffering-getting-the-dw01-and-8205a-protection-circuit-work-right/
https://www.eevblog.com/forum/projects/dw01-g-battery-protection-ic-issues/

That means I can not just blindly add a switch to cut the whole god damn battery because it will enter in the limbo state. I have to redesign the circuit to instead disconnect the EN pin of my LDO. Also I have to add a fucking button which you are supposed to press when you have switched the battery.

The PCBs are already manufactured, took a fucking month. Now I have to either hack like 10 of PCBs or rebuild fucking everything.

Legit wanna kill myself.

>>1730945
Schematics of my shitty protection circuit

>>1730803
fuck!
so during the pcb milling the mill sometimes resets, or when drilling holes the holes are wierdly offset, this never happens when milling other stuff like wood or alu
I watched it go and when the bit makes contact with the pcb it releases a blue spark and this must be fucking with the arduino uno which is running the mill
is there something i can do about this? like grounding the pcb or something?

>>1730980
Grounding and shielding should do the trick, perhaps adding some caps in places. Using twisted pairs or coax for your wires might also help but I doubt it's necessary.

>>1730983
>Grounding and shielding should do the trick,
so i connect the pcb i am milling to the ground of the powersource that is powering the arduino?

As for shielding i don't think there is much i can do since brain arduino is attached to the frame of the mill and i can't really move it away

>>1730986
Connect everything to ground, but I'm not sure if you want to put a resistor in series with the chassis to stop current spikes from going through the spindle-bearing. So connect the mill chassis and PCB to ground, and also the arduino's 0V rail. An easy upgrade to make that may make a very real difference would be putting ferrites on the wires going to and from the arduino. Also see about solidifying the reset pin (i.e. tying it to a rail), as I think it might be capacitively coupled to something on the board that could cause a problem.

How do you calculate the SNR of an arbitrary set of data when you barely know what the signal looks like? I know SNR is the square of the ratio of the RMS signal to noise amplitudes. But in this case, it's hard to say what's signal and what's noise. I'm thinking low pass filter to get the triangle wave as the signal and high pass filter to get everything else as the noise.

>>1731011

how do you know a girl is ugly? do you have to calculate her index of beauty? no, you just look. same here: you dont need numbers to tell you the signal is uggo.

>>1731039
Yeah, but I want to know how to find the number that tells me how ugly it is in a way that other guys will understand

>>1730995
Ferrite beads on the motor connections too. Most printers have them.

>>1731011
What follows is but a guess:
I'd get the normalised FFT of your ideal/transmitted data set A, the normalised FFT of the received noisy data set B, subtract A from B to get C, the noise data set and get the normalised FFT that too. Then integrate/sum over C's FFT and integrate/sum over A's FFT to get a total signal and noise numbers. If you do it right, the actual length of the data set won't impact the result. I guess you'd just sum up the absolute value of every data point in each FFT.

You can apply a filter to B in order to increase the effective SNR, but you'll still be stuck with the noise at and around your transmission frequency.

>>1731083
I don't have a transmitted data set, though. This is data from an ECG.

>>1731087
Wow what leads are those? You're getting a holy fuckload of noise. towards the end as well. Are you using good electrodes?

>>1731094
To be honest I don't know much about the setup. A friend got this dataset in one of his classes and asked me how to find the SNR. I told him what I wrote above, I'm just curious if there's a more correct way.

>>1731087
Then FFT the received dataset and figure out what frequencies (and their harmonics) you're expecting and what ones must be from noise.

>>1731095
Ah okay. I was also gonna say there's an obvious Q wave on there, which suggests this patient has had a heart attack.

Even on the first beat, the P wave is getting lost in the noise. Whatever the SNR is in actual values, it's unacceptable IMO for medical diagnostic use.

>>1730946
how'd they taste?
I mean you gotta RTFD m80
the fast boat from LCSC is worth considering

>>1731041
if you don't know signal you can't calculate SNR. you need a known reference, even if that means resynthesizing the data set based on other knowledge about the data, or just crudely dictating signal or noise membership based on frequency
also consider sedating the patient

Speaking of EKGs, have you guys ever looked at how they work? Like what they measure and how the different leads give the clinician a three-dimensional look at the heart by tracking the depolarization of the cardiac muscle?

I'm still astounded that someone came up with that.

>>1731133
they'd been fucking with frog legs and electricity since early days. a reciprocal relationship was almost to be expected
>he says in complete hindsight

>>1731149
Haha fair enough. I'm still stunned at how it works. Especially the chest leads.

There are 6 different terminals that go on the chest (in addition to 4 that go at the wrists/ankles or around the hips/biceps). Since the EKG is voltages, if you put one of your scope leads on a chest electrode, where do you think you put the other lead to measure the potential difference?

I still love this.
>Frank Norman Wilson (1890-1952) investigated how electrocardiographic unipolar potentials could be defined. Ideally, those are measured with respect to a remote reference (infinity). But how is one to achieve this in the volume conductor of the size of the human body with electrodes already placed at the extremities? In several articles on the subject, Wilson and colleagues (Wilson, Macleod, and Barker, 1931; Wilson et al., 1934) suggested the use of the central terminal as this reference. This was formed by connecting a 5 kΩ resistor from each terminal of the limb leads to a common point called the central terminal, as shown in Figure 15.5. Wilson suggested that unipolar potentials should be measured with respect to this terminal which approximates the potential at infinity.
>Actually, the Wilson central terminal is not independent of but, rather, is the average of the limb potentials. This is easily demonstrated by noting that in an ideal voltmeter there is no lead current. Consequently, the total current into the central terminal from the limb leads must add to zero to satisfy the conservation of current (see Figure 15.5).

>>1731154
nah you gotta hook up your body in wye configuration

i need to replace a capacitor in my old TV, bit they only one they had at the store was 10V, which is much bigger than what i need. it doesnt even fit the case unless i solder it to some big ass cables instead of directly into place.

os this all right? or should i try the next electronics store, which is a 30 min drive away

Do i understand it correctly that the switch above should move by itself when the relay turns on or off
https://datasheet.octopart.com/HFE10-2/12-HT-L2%28257%29-Hongfa-datasheet-12510254.pdf

>>1731342
What capacitor? You'll have much better luck ordering online.

>>1731342
Order online. You don't have to exactly match values if it's expensive or hard to source. You can usually go up in voltage and capacitance.

>>1731342
>i solder it to some big ass cables

99% of the time this will not be a problem.
you can also solder it on the opposite side of the PCB if there's room for it.

#### ATTENTION ####

bump limit and page 9.
where is faggot OP for to make new thread.?

maybe this?
simple machines
fully functional
easy to intel

>>1731468

>>1731520
>>1731520
>>1731520

>>1731463
>you can also solder it on the opposite side of the PCB if there's room for it.
great idea thanks, i'm embarrassed i didn't think of this earlier