|
kid sinister posted:Why can't anything just work? original sin. next question
|
#
?
Dec 21, 2019 03:05
|
|
|
|
| # ? Jun 11, 2024 16:18 |
|
|
just think about how amazing it is that it ever works, and it keeps doing it over and over
|
|
#
?
Dec 21, 2019 03:32
|
|
|
Wait, how do I size a fuse? This mini power supply should make a max of 1 amp at 5 volts DC out from 7-7.5 volts AC in.
|
|
#
?
Dec 21, 2019 16:48
|
|
|
Are you trying to protect the supply or the downstream device? Size your fuse appropriately for the task. And check the datasheet for the fuse speed at various amperages.
|
|
#
?
Dec 21, 2019 16:52
|
|
|
I'm trying to protect the power supply board. How do I size it?
|
|
#
?
Dec 21, 2019 17:41
|
|
|
kid sinister posted:I'm trying to protect the power supply board. How do I size it? If you don't expect any inrush current, 125% of the design amperage. Transient inrush (caps and whatnot) 150%. Motors 200-250%. If you are trying to protect the upstream 7.5VAC from the power supply board shorting out, fuse the input to the supply. If you're trying to protect the power supply board from the load shorting, fuse the output. In your example, your max amperage is 1A@5V, so you'd want a 1 1/4A fuse. Always round fuse sizes down, so use 1A if you can't find a 1.25A. babyeatingpsychopath fucked around with this message at 19:32 on Dec 21, 2019 |
|
#
?
Dec 21, 2019 19:29
|
|
|
Along the same lines, I'm switching a fairly large load with an N channel MOSFET, up to ten amps or so. If I accidentally short that out, I'm trying to figure out how to save the MOSFET. I've tested both breakers and fuses, and they're not fast enough. The MOSFET fails. So I've got an Extremely Clever design with a few op amps to shut down stuff when a threshold is hit. Still untested so I dunno if it will work, but open to other alternatives
|
|
#
?
Dec 21, 2019 20:44
|
|
|
IIRC there are gate driver ICs that will do that for you, but also I respect being Extremely Clever because thats fun
|
|
#
?
Dec 22, 2019 00:37
|
|
|
I've been working on protection circuitry for the last few months and it's a bit of a nightmare, especially because I don't yet have the equipment to test a 10kV 1.2us/50us pulse so will it pass the standards testing? Who knows!
|
|
#
?
Dec 22, 2019 01:47
|
|
|
Make your equipment run at 20kV, maybe it can zap the test device back and win the battle definitively
|
|
#
?
Dec 22, 2019 01:56
|
|
|
ante posted:Along the same lines, I'm switching a fairly large load with an N channel MOSFET, up to ten amps or so. If I accidentally short that out, I'm trying to figure out how to save the MOSFET. I've tested both breakers and fuses, and they're not fast enough. The MOSFET fails. I'd look in digikey's "PMIC Load Switch category". Here is an example of a protected fet: https://www.st.com/content/ccc/reso....CD00002219.pdf Otherwise yeah, you need current sense and some sort of comparator. There are some options for integrating those: https://www.allegromicro.com/en/Products/Sense/Current-Sensor-ICs/Zero-To-Fifty-Amp-Integrated-Conductor-Sensor-ICs/ACS711 The hall effect sensor is a nice option since its resistance is extremely low and its isolated so you can sense anything with it.
|
|
#
?
Dec 22, 2019 16:56
|
|
|
CopperHound posted:Okay. You can't share a ground between your ac source and full wave rectifier output. You end up creating a short for half the cycle. You can however share the ground with a half wave rectifier. I talked to some old farts on a tube radio forum and sure enough, this was the solution they recommended. I'm waiting for my diode to come in the mail, with some spares. Serves me right for trying to be more efficient.
|
|
#
?
Dec 24, 2019 17:25
|
|
|
I'm laying out a big honkin' LED display and I just realized I routed an SPI line all the way around 3/4ths of the outside of the huge board, that's going to radiate like a motherfucker isn't it... should I spend the time to try to route that better or will it not actually cause problems? It's not like this has to pass FCC clearance or anything, it's just gonna be running all the time and I don't want it interfering with other stuff... e: I guess I could go up to 4 layers and run it on inner layers with ground planes on either side, but the price of the boards goes from ~$3 each to $19 each just to add the extra layers to this big a board, yikes. Shame Boy fucked around with this message at 15:31 on Dec 27, 2019 |
|
#
?
Dec 27, 2019 15:27
|
|
|
It'll probably be fine. I've seen plenty of really really bad routing work, even stuff like part of a dc-dc switcher going the long way round the board. Is it under 100MHz?
|
|
#
?
Dec 27, 2019 17:43
|
|
|
taqueso posted:It'll probably be fine. I've seen plenty of really really bad routing work, even stuff like part of a dc-dc switcher going the long way round the board. Is it under 100MHz? Yeah the bus is going to run at like 5-10MHz tops, and the microcontroller is only 16MHz, so I guess it's not really that big a deal...
|
|
#
?
Dec 27, 2019 17:57
|
|
|
Shame Boy posted:I'm laying out a big honkin' LED display and I just realized I routed an SPI line all the way around 3/4ths of the outside of the huge board, that's going to radiate like a motherfucker isn't it... should I spend the time to try to route that better or will it not actually cause problems? It's not like this has to pass FCC clearance or anything, it's just gonna be running all the time and I don't want it interfering with other stuff... The best bang for the buck with signal integrity is source series termination. Just a ~33 ohm resistor at the driver source.
|
|
#
?
Dec 27, 2019 22:08
|
|
|
Shame Boy posted:I'm laying out a big honkin' LED display and I just realized I routed an SPI line all the way around 3/4ths of the outside of the huge board, that's going to radiate like a motherfucker isn't it... should I spend the time to try to route that better or will it not actually cause problems? It's not like this has to pass FCC clearance or anything, it's just gonna be running all the time and I don't want it interfering with other stuff... This will be fine. SPI busses always end up a horrible mess but they're not really high enough energy to radiate. You're more likely to have the opposite problem, and your SPI may be impacted by noisy stuff it runs near. Termination resistors can help, and it's easy to put them in the design and just put a 0 ohm link in during assembly until you need them
|
|
#
?
Dec 28, 2019 06:11
|
|
|
There was something I saw ages ago and can't remember what it was called, maybe someone here will know. It was some kind of plastic or glass that when you applied a current to would turn black, take the current off and it goes back to clear. Anyone have any idea what that stuff is called? I had the idea of using it as a film camera shutter, which would be pretty cool if doable.
|
|
#
?
Dec 28, 2019 15:14
|
|
|
E paper?
|
|
#
?
Dec 28, 2019 16:39
|
|
|
Cholesteric LCD is the name of one such technology. But I'm fairly confident it's too slow to be a camera shutter -- much slower than an active LCD at least. Edit: misread your request and confused myself. The thing you're talking about is "switchable glass", "smart glass", or "e-glass". I don't know how fast it switches KnifeWrench fucked around with this message at 16:46 on Dec 28, 2019 |
|
#
?
Dec 28, 2019 16:42
|
|
|
Parts Kit posted:I had the idea of using it as a film camera shutter, which would be pretty cool if doable. There's an old macrophotographer who puts together some fairly sophisticated rigs to photograph bugs in flight, and I think that one year he used the lenses from a set of active-mode 3d glasses to do just this because the physical shutters on his SLRs are too slow to achieve the necessary speeds. e: Found it! https://www.flickr.com/photos/fotoopa_hs/17265864892 https://www.photomacrography.net/forum/viewtopic.php?p=169134&sid=b8256ab95dd526db6c601d785599a6e1 Dia de Pikachutos fucked around with this message at 07:27 on Dec 29, 2019 |
|
#
?
Dec 29, 2019 06:50
|
|
|
here’s a stupid + frivolous question you shouldn’t waste too much time on: could you modify the pulvermacher chain battery concept:  (a galvano-healing quack device from the 1800s that is really just a flexible/articulated modular primary wet cell battery in the form factor of a chain made using dissimilar metal wire/sheet elements; for electrolyte, either temporary use w sponged-on vinegar (for copper-zinc designs) or (w magnesium-copper)continual operation as long as the user sweats on it. it continually shocked the wearer in operation, but they had a popular secondary application as an early adjustable-current power source for experimenters.) ...could you adapt the sweat-electrolyte magnesium chain design to not shock the wearer and instead provide general-purpose trickle-charging for your electronics, power for wearables/LEDs, etc
|
|
#
?
Dec 29, 2019 08:48
|
|
|
That's hilarious
|
|
#
?
Dec 29, 2019 09:55
|
|
|
Ambrose Burnside posted:here’s a stupid + frivolous question you shouldn’t waste too much time on: This is going to be one of the wackier goon deaths Don't take that as discouragement though! Either you'll revolutionise wearables or you'll die an internet legend.
|
|
#
?
Dec 29, 2019 10:04
|
|
|
KnifeWrench posted:Edit: misread your request and confused myself. The thing you're talking about is "switchable glass", "smart glass", or "e-glass". I don't know how fast it switches spongepuppy posted:There's an old macrophotographer who puts together some fairly sophisticated rigs to photograph bugs in flight, and I think that one year he used the lenses from a set of active-mode 3d glasses to do just this because the physical shutters on his SLRs are too slow to achieve the necessary speeds.
|
|
#
?
Dec 29, 2019 15:40
|
|
|
Hello thread! I am working on a pulsed plasma device for my startup and while I’ve done these things for a decade or so I’m having a pretty hard time conceptually with one of my designs. The system uses a Inductively Coupled plasma antenna that is effectively just a large three turn antenna that is about 10 cm tall. This runs at 13.56 MHz through a matchbox. I’m not worried about this, it works and the generator and match are the gruntiest and best protected part of the system by far. Typically runs on the order of hundreds of amps. Directly below it, I want to attach another strap antenna and run it at an inductive heating frequency for crucible work. Since I have some flexibility in frequency I’ll be trying to do it via zero-voltage switching, but what I am concerned about is the heater strap antenna seeing the ICP field and blowing up the ZVS. I say this because I also have a pair of electromagnets in series next to the ICP and I’ve managed to blow out two pretty grunty amplifiers in about 100 hours of runtime. I think I can solve the magnet problem by just clamping the gently caress out of it using a line filter; even though it’s common-mode, the frequency change is different enough it shouldn’t matter, but what should I be doing to filter the ZVS heater?
|
|
#
?
Dec 29, 2019 19:44
|
|
|
Splode posted:This is going to be one of the wackier goon deaths i was at a bar w some (decidedly non-electrical) engineers last night and the table was divided as to if you could use skin-contact sweat as an electrolyte while not getting shocked or if the two are inextricable. we couldn’t work it out ourselves, shortly after finishing pitcher # ???, for some weird reason said it was frivolous (...altho i think sth like a sweat-powered watch w pulvermacher chain band would be a cool one-off art/concept piece for a design portfolio, real-life practicality aside) Ambrose Burnside fucked around with this message at 20:12 on Dec 29, 2019 |
|
#
?
Dec 29, 2019 20:04
|
|
|
I don't know the exact chemistry byproducts etc, but like any chemical battery I'd expect the electrodes to get depleted over time, so its not like you're literally harnessing unlimited energy from sweat or anything. But anyways, I think that each coil should act like parallel plates so using longer coils and less "links" in series would be one way to adjust voltage vs current. I'd mostly be worried about whether sweaty metals would absorb through skin and lead to some toxicity issues. Interestingly I did a quick search on zinc poisoning and found this quote:Zinc and copper compete for absorption in your small intestine.
|
|
#
?
Dec 29, 2019 20:51
|
|
|
I just want to back up and say maybe we shouldn't just throw out the shock part right away. There could be a market for this, maybe we can concentrate on underwear.
|
|
#
?
Dec 29, 2019 21:48
|
|
|
taqueso posted:I just want to back up and say maybe we shouldn't just throw out the shock part right away. There could be a market for this, maybe we can concentrate on underwear. What a coincidence, I just got to this part in the wikipedia article: quote:Pulvermacher even had a model designed to attach to the male genitals in a special sac which was claimed to cure impotence and erectile dysfunction.
|
|
#
?
Dec 30, 2019 06:31
|
|
|
Ambrose Burnside posted:here’s a stupid + frivolous question you shouldn’t waste too much time on: Goon Project!
|
|
#
?
Dec 30, 2019 13:06
|
|
|
Ambrose Burnside posted:i was at a bar w some (decidedly non-electrical) engineers last night and the table was divided as to if you could use skin-contact sweat as an electrolyte while not getting shocked or if the two are inextricable. we couldn’t work it out ourselves, shortly after finishing pitcher # ???, for some weird reason Since nobody's really answering this question: a single copper-zinc cell is 1.1V. That's your voltage. Full stop. You will never get more voltage than that. You can get even multiples with more cells. Three cells = 3.3V. Wire gauge affects internal resistance, larger wire = less resistance. Surface area in contact with electrolyte is your current delivery. More surface area = more current. The easiest way to get more surface area is longer links. The other easy way is smaller gauge wire. Make a couple and see what the life looks like. Copper and zinc both corrode pretty rapidly when in contact with wet skin and exposure to air, so total useful voltage life may be time-limited. Air-zinc cells ave a fairly impressive self-discharge once exposed; hearing aid batteries with this chemistry have about two weeks of life whether they're in heavy use or not. For your copper-zinc cell, I don't think skin toxicity is a big problem, but I have a cavalier attitude about industrial hygiene.
|
|
#
?
Dec 31, 2019 02:11
|
|
|
if i were to actually prototype sth (not actively planning on it, fwiw) i'd skip zinc cathodes and use magnesium; the zinc ones need to be charged with vinegar to do anything + only operate in short bursts until the dowel sponge cores dry out, magnesium works w salt water so no electrolyte sponges are needed + your body automatically waters the galvanic garden, so to speak. also, if im reading the standard electrode potential table right, that bumps the single-cell voltage up to a more convenient ~2.72V irt toxicity i don't think anything on the table is a substantive issue- zinc and copper are both omnipresent in coinage and jewellery that spend long periods in direct skin contact, no hazard is indicated for skin contact, and magnesium seems equally-innocuous in that specific regard the REAL problem with this as a seriously-considered practical wearable is one you already deal with in jewellery, but dramatically magnified here- the metal oxides that rub off of uncoated tarnish-prone metals stain the skin, transfer to clothing, and are frequently itchy/irritating/can cause contact dermatitis if you fly close to the sun for long enough. if we add a galvanic reaction to the mix that oxide production is going to increase dramatically. it would just be horrible to have to actually wear one of these things on an extended basis, you'd be tattooed green and black from the verdegris + Mg ox and have a rash in the post from the scratching. idk what a coherent and really 'useful' application would be... maybe as a component of a wearable emergency transponder beacon or sth. the design has an infinite shelf life while in airtight packaging, functions as long as it's attached to a living/sweating human being, and would reliably provide a small amount of power for many days or weeks regardless of environmental conditions. the downsides of skin irritation and eventual destruction of the battery/chain itself become acceptable Ambrose Burnside fucked around with this message at 02:52 on Dec 31, 2019 |
|
#
?
Dec 31, 2019 02:43
|
|
|
some sort of membrane over the metals?
|
|
#
?
Dec 31, 2019 02:46
|
|
|
Ambrose Burnside posted:idk what a coherent and really 'useful' application would be...
|
|
#
?
Dec 31, 2019 09:27
|
|
|
If I'm building a battery powered device is there any need for filtering on the battery power? I found a paper that examined battery noise and found that essentially 100% of the noise from batteries is from Johnson-Nyquist noise in the equivalent series resistance, which makes me think that it's stable and clean enough that there's no need for any filtering.
|
|
#
?
Dec 31, 2019 10:20
|
|
|
I have done a lot of battery logging with lithium, and also with CR2032 coincells, and it's all been very clean.
|
|
#
?
Dec 31, 2019 10:51
|
|
|
Batteries should be clean, I guess there could be some low frequency noise caused by chemical effects depending on the battery type? Anecdotally an audiophile friend once told me that lead acid batteries were noisy, but I've never attempted to verify that. The battery ESR/ESL is probably not very good at higher frequencies so you'd still need bypass capacitors for most designs.
|
|
#
?
Dec 31, 2019 12:15
|
|
|
So at work I'm diving into some FPGA poo poo for the first time in ten years, which is cool. But I'm working with VHDL for the first time and holy crap I can't believe how strongly typed this language is. What's that, you want to add a boolean to an integer? Well gently caress you, I have no idea what the "+" symbol means anymore. Also the lack of distinction between blocking and non blocking assignments is melting my brain. Yeah I think I get it, VHDL translates more directly into synthesis than Verilog. But compilers have improved in the last 30 years, which is why most people don't write poo poo in assembly anymore.
|
|
#
?
Dec 31, 2019 16:30
|
|
|
|
| # ? Jun 11, 2024 16:18 |
|
|
ANIME AKBAR posted:What's that, you want to add a boolean to an integer? Well gently caress you, I have no idea what the "+" symbol means anymore. See I had to program in Javascript for years so I'm very very thankful for strong typing now. Why yes Javascript, I definitely wanted you to convert that to a string and append it, thanks
|
|
#
?
Dec 31, 2019 17:15
|
|