|
I'd use a bunch of ten amp supplies at regular intervals. Or even five amps. Easier, and regardless of rated current, you'll get discoloration by the end of the tails, so multiple taps fixes that
|
#
?
Oct 28, 2019 20:17
|
|
|
|
| # ? May 9, 2024 15:13 |
|
|
Sockser posted:I’m working on a big LED video wall type project, probably 576 neopixels— at 70% brightness that sets me at a max of <25A@5v. This is the biggest electronics project I’ve worked on and definitely pushing the boundaries of my comfort zone For wire AWG amp capacity, I use this worksheet I wrote a while ago: https://docs.google.com/spreadsheets/d/1Adq8C7qj4uQJbO45ggQ0-14KRVSZKUt-losqIGLVTvs/edit?usp=sharing
|
|
#
?
Oct 28, 2019 20:23
|
|
|
oh no insane neopixel projects are invading this thread too!
|
|
#
?
Oct 28, 2019 20:26
|
|
|
At least this guy has used a microcontroller before and built something similar and seems to be on the right track The other one I didn't want to touch with a ten foot pole, yeesh
|
|
#
?
Oct 28, 2019 20:52
|
|
|
This is a good opportunity to post a pic of a project I'll probably realistically never finish: Running an ESP32, scraping the power company's website and lighting up a map wherever there are power outages. Color changes based on percentage of affected customers. 338 individual WS2812b lights, hand soldered into a grid (really a single long and winding line). I never want to do that again. Original plan was to frame it and mount it on the wall, but then I moved across the country and lost momentum.
|
|
#
?
Oct 28, 2019 21:25
|
|
|
Do those nixie tubes count the number of times you've thought about finishing a project but didn't?
|
|
#
?
Oct 28, 2019 21:37
|
|
|
Cojawfee posted:Do those nixie tubes count the number of times you've thought about finishing a project but didn't? Shows the total customers on the map without power. I don't have enough tubes to display the number of starts and stops on this project.
|
|
#
?
Oct 28, 2019 21:40
|
|
|
Sockser posted:I’m working on a big LED video wall type project, probably 576 neopixels— at 70% brightness that sets me at a max of <25A@5v. This is the biggest electronics project I’ve worked on and definitely pushing the boundaries of my comfort zone First thing first: you have 40A available, so don't assume you only need to carry what you plan to use. What if a glitch in your code causes the neopixels to go 100% white? You want that to be an irritating annoyance, not a fire. You can scale your wires and connectors for worst case, and/or add protection circuitry like an in-line blade fuse holder to enforce your assumptions. I recommend both if it's not cost prohibitive (which for qty.1 it sounds like it wouldn't be) As for what gauge to use, if it's a public installation, it sounds like you could/should be in the realm of code and inspection, so maybe this shouldn't be solely at your discretion with advice from the internet? 16AWG can handle ~22A according to my favorite online rule of thumb table (https://www.powerstream.com/Wire_Size.htm) but the spreadsheet Stabby McDamage made looks more thorough, and skews more conservative, looking like you'd want 14-12AWG depending on your protection circuitry and distances. But again, check with the person in charge of signing off on the safety of the facility where it's being installed. 40A is nothing to sneeze at, even though the voltage is low. I agree with the idea of multiple smaller power supplies, but you said you'd already bought it, so maybe your hands are tied there.
|
|
#
?
Oct 28, 2019 21:59
|
|
|
Ambrose Burnside posted:i didn't expect reasonably-portable and affordable batteries to be able to keep up with a ~700W induction heater that can heat up a titanium dab nail to yellow-orange heat in less than 5 seconds, but apparently it's nbd with some decent-quality LiFePO4 cells:
|
|
#
?
Oct 28, 2019 22:06
|
|
|
KnifeWrench posted:First thing first: you have 40A available, so don't assume you only need to carry what you plan to use. What if a glitch in your code causes the neopixels to go 100% white? You want that to be an irritating annoyance, not a fire. You can scale your wires and connectors for worst case, and/or add protection circuitry like an in-line blade fuse holder to enforce your assumptions. I recommend both if it's not cost prohibitive (which for qty.1 it sounds like it wouldn't be) It’s a burning man type installation rather than an art piece living in the city park for the season installation, fwiw. I don’t think there’s anybody who’s going to inspect it Also I am throwing a 30A fuse in-line with that power supply but didn’t mention it in my previous post. I could swap out for two 18A power supplies or something, I’ve got plenty of time to rejigger plans, though I did already 3d print a sweet enclosure for the 40A supply, so that’s a bit of a shame. If I were to switch out for two smaller supplies, would I want to break each one out to four strips, or run the two power supplies into a bus that breaks out to eight strips? I imagine the former because combining them into one bus defeats the point of distributing the power in the first place?
|
|
#
?
Oct 28, 2019 22:09
|
|
|
Sockser posted:It’s a burning man type installation rather than an art piece living in the city park for the season installation, fwiw. I don’t think there’s anybody who’s going to inspect it Ah, okay. Well, if you don't mind thicker wires, there's no theoretical problem with the 40A supply. Definitely don't put two supplies in parallel, though. Your intuition is right about that: each supply has a tolerance for its output, and whatever that difference might turn out to be would be a voltage across a dead short. The main thing you want to avoid (after fire) is an appreciable voltage drop at the VCC of the furthest neopixel. This is one of the main advantages of multiple supplies: reducing the maximum distance. You never mentioned the linear distances you plan to cover, but you did use the words "big" and "wall" so I imagine you won't want to hand wave the way most 5VDC projects can afford to.
|
|
#
?
Oct 28, 2019 22:25
|
|
|
My understanding has always been that trying to achieve high current by using multiple power supplies in parallel is not a great idea. However, if you actually have multiple, separate rails at the same voltage with common ground, then that's fine. That also lets you use more, smaller fuses on each rail too, which is safer (the problem with a 30A fuse is that it only blows at 30 A, and you can do a lot of damage with 29 A).
|
|
#
?
Oct 28, 2019 22:26
|
|
|
The individual strands will be stretching 8ft4in, approximately, and then like 13in tall. I’m running 8 strands of 64 (50 on a strand + 14 on a second strand) rather than one strand of 512 which should help with power injection; not sure if I’ll still need to add power in the middle of the strands. Worst case I can put the power supply halfway between the two and just run power 3 foot each direction or something, if I really need to inject power midway, but I think if they’re not all set to 100% white they should be mostly fine E: My biggest previous project ran ~250 leds off a 10A supply using a breadboard as a power bus which was probably not the smartest call but it made it real easy to run JSTs to everything Sockser fucked around with this message at 22:35 on Oct 28, 2019 |
|
#
?
Oct 28, 2019 22:29
|
|
|
This may not be in line with forums values, and if it's not let me know and I'll take this post down, but I'm on the team working on a small multimeter/oscilloscope project, which is live on Kickstarter for those who are interested. http://kck.st/32TV1Gz I only mention it because "I want a really really cheap scope" comes up here a lot, and while the sample rate is lower, this is way cheaper than a full sized device. I don't make any extra money if you buy one, though it will be cool seeing something I worked on out in the world.
|
|
#
?
Oct 28, 2019 22:34
|
|
|
Sockser posted:My biggest previous project ran ~250 leds off a 10A supply using a breadboard as a power bus which was probably not the smartest call but it made it real easy to run JSTs to everything LED array and space heater, all in one!
|
|
#
?
Oct 28, 2019 22:36
|
|
|
Splode posted:This may not be in line with forums values, and if it's not let me know and I'll take this post down, but I'm on the team working on a small multimeter/oscilloscope project, which is live on Kickstarter for those who are interested. Is this going to be like the Vion but hopefully not suck then?
|
|
#
?
Oct 29, 2019 00:31
|
|
|
Sockser posted:I’m working on a big LED video wall type project, probably 576 neopixels— at 70% brightness that sets me at a max of <25A@5v. This is the biggest electronics project I’ve worked on and definitely pushing the boundaries of my comfort zone I like Molex standard connectors. Their .093" series is good for a bunch of amps. Unfortunately molexkits.com is down, but they had a great selection of stuff; whole kit with pins and housings and a decent crimper (in the deluxe kit). From my desk reference kit: Pins p/n 42478-0001, sockets p/n 42477-0005 good for 14A each, will take 14-30AWG wire. Standard Molex crimper. Then just grab part numbers for housings (plugs and receptacles) from the offical Molex website and go to town. babyeatingpsychopath fucked around with this message at 01:43 on Oct 29, 2019 |
|
#
?
Oct 29, 2019 01:35
|
|
|
mobby_6kl posted:Is this going to be like the Vion but hopefully not suck then? Yeah, we've already got a miniature app driven multimeter/oscilloscope that works in the original pokit meter. That can only handle 60V though, so this one will be a larger unit for higher voltages
|
|
#
?
Oct 29, 2019 02:27
|
|
|
Splode posted:If you're not exceeding the rated continuous discharge for the battery they really should last their normal life. But are they really rated for 50 Amps continuous? That's insane, I'm used to working with much smaller lipo batteries that are usually only rated for 1 or 2 times their capacity i didn't mean that i'm worried abotu the total lifespan of the batteries (lifepo4 cells seem ideal for this particular application so i wouldnt expect an issue there) but rather the actual working time a battery pack would give me. i forgot how "watts" and "watt-hours" work, naturally, and now know i'd get... a little under 2 minutes' work at 1000W output from that $100+ w shipping battery pack. Assume 5-10 seconds per heat, that's ~12-24 working cycles, which, assuming a 5% duty cycle, keeps me working for... about half an hour. So that's answered, I guess, although that'd be acceptable for demos, legit induction coil fieldwork like heating nuts/bolts/seized poo poo, aaaaaaaand ofc hittin that wax on the go 
|
|
#
?
Oct 30, 2019 00:36
|
|
|
So I was thinking of making a pulse function for my TIG welder, I checked out this youtube video and it looked doable: https://www.youtube.com/watch?v=hk02102nefk Then I went on ebay and I found this, looks like a complete unit for pulsing, works down to 1hz and it sets the pulse using percentage. I'm not sure but to me it looks like this could be hooked up between the pedal and my welder without much fuss, it varies a voltage (don't remember how much, 10V max iirc) to change the output of the welder. https://www.ebay.com/itm/Square-Sig...oqxdWPGwhTdYxFw Or am I missing something obvious. It looks an interesting alternative to the video and comes with a ready interface. Though I would like longer intervals than 1hz. EDIT: I think I found the ovious, no way to set a value to the lower value of the pulse, ideally it would be a percentage of the bigger value. Guess it's better to build the circuit, then you can get pulses slower than one second too. His Divine Shadow fucked around with this message at 19:44 on Oct 31, 2019 |
|
#
?
Oct 31, 2019 14:41
|
|
|
we have a mega watts/mega savings power supply winner, i think: the HP DPS-1200FB, an inexplicably-cheap server power supply that's apparently a darling of everyone from RC nuts making burly-rear end lipo chargers to bitcoin weirdos giving themselves server heat-induced brain damage. they supply 75A @ 12V and you can get em for like $30 on ebay. I want 24V at least, but people commonly buy a couple and connect em in series to get a more useful voltage, which I believe would work fine with my induction heater application as long as one supply doesn't crap out mid-heat?? on that note: very bad things happen to ZVS induction heaters if the voltage drops beneath 11V so 1/2 supplies failing could put me near that territory, and I've always heard bad things about using multiple power supplies for a single power-hungry device, are there any obvious pitfalls or shortcomings to this 2x1200FB power scheme Ambrose Burnside fucked around with this message at 19:04 on Oct 31, 2019 |
|
#
?
Oct 31, 2019 18:36
|
|
|
I built a 24v supply from series-ed Delta supplies. It will supply 24v @ 47A. I use it with my 308 Duo and can charge lipos at 30v / 30A now. owns bones. You have to do a bunch of ridiculous poo poo to isolate them from ground to put them in series. I completely disabled the grounding pin on both, cut grounding wires, isolation washers, and wrapped the chassis of both in Kapton tape.
|
|
#
?
Nov 1, 2019 13:45
|
|
|
Quick question I need to make a voltage divider to get 12V down to 6, to power a 1 watt bulb. What resistors should I use for this?
|
|
#
?
Nov 1, 2019 20:14
|
|
|
A 1W resistor. Or just use a second bulb.
|
|
#
?
Nov 1, 2019 20:25
|
|
|
To go into some more detail P = V * I 1W / 6V = 1/6 A current for your bulb. The resistor needs to drop 6V at (1/6)A R = V/I R = 6V / (1/6)A = 36 Ohms 36 ohm is not a standard value, and 1W is slightly uncommon, so you parallel a bunch of higher resistance to get the exact desired resistance. For example 36*5 = 180 So parallel 5x 180 Ohm resistors @ 1/4W should work great.
|
|
#
?
Nov 1, 2019 20:50
|
|
|
HorseLord posted:Quick question I got beaten, but to add consensus: Depends on the bulb. If the bulb uses a filament that acts like a resistor, that will be part of the calculation. In general, though, P = V^2/R, so you're looking at the neighborhood of 36 ohms for 6V to be 1W
|
|
#
?
Nov 1, 2019 21:00
|
|
|
Alternately, use a second 1W 6V bulb in series instead of a resistor.
|
|
#
?
Nov 1, 2019 21:20
|
|
|
Imo find a 12v 1w bulb
|
|
#
?
Nov 1, 2019 21:50
|
|
|
insta posted:I built a 24v supply from series-ed Delta supplies. It will supply 24v @ 47A. I use it with my 308 Duo and can charge lipos at 30v / 30A now. owns bones. 
|
|
#
?
Nov 1, 2019 22:35
|
|
|
Good name/quoted content synergy
|
|
#
?
Nov 1, 2019 22:41
|
|
|
I've done that out to 2400V and 1500W. It's fiiiiiine.
|
|
#
?
Nov 1, 2019 23:20
|
|
|
the fact its even possible at all to do that by cutting a wire means the converter topology was already an isolated one that they just carried the wall ground through on its fine
|
|
#
?
Nov 1, 2019 23:42
|
|
|
I got chased out of the Arduino thread for saying a couple of kW wasn't that much, you guys are the best kind of crazy
|
|
#
?
Nov 2, 2019 01:24
|
|
|
Splode posted:I got chased out of the Arduino thread for saying a couple of kW wasn't that much, you guys are the best kind of crazy that guy was talking about controlling 1500 amps with 25 arduinos uno
|
|
#
?
Nov 2, 2019 02:43
|
|
|
I still think that it could have been broken up into a enough independently powered sub modules to not be a fire hazard. It never would have worked though.
|
|
#
?
Nov 2, 2019 02:46
|
|
|
Sagebrush posted:25 arduinos uno I enjoy this plural
|
|
#
?
Nov 2, 2019 03:02
|
|
|
Sagebrush posted:that guy was talking about controlling 1500 amps with 25 arduinos uno The Arduino website clearly states quote:Arduino is an open-source electronics platform based on easy-to-use hardware and software. Arduino boards are able to read inputs - light on a sensor, a finger on a button, or a Twitter message - and turn it into an output - activating a motor, turning on an LED, managing 30 other arduinos to control 60,000 neopixels in a display you have no reason to create.
|
|
#
?
Nov 2, 2019 03:10
|
|
|
Foxfire_ posted:I enjoy this plural Arduino Venticinque?
|
|
#
?
Nov 2, 2019 04:58
|
|
|
Kilowatts and whole amps are a different world from what I'm used to in radiation detection. I live in a land of gigaohms and kilovolts and picoamps. I'm having trouble figuring out proper grounding practices though. Here's a picture of the charge-sensitive preamplifier I built:  The circuit board has a ground plane on the back side of it except near the high voltage stuff. Originally the board was grounded to the chassis in only one area, the black wire leading to the solder tab around the BNC connector on the right of the photo. This seemed like it would be fine because the resistance between the shield of the SHV connector on the left and any other grounded part in the circuit was the same resistance of like 0.1 ohms (which is like the resistance of my leads.) However if I plugged a cable into that SHV connector on the left, the cable seemed to act as an antenna even if it was terminated with a proportional counter. Like I was getting really bad 100 MHz noise with a 1-foot cable and really bad 40 MHz noise with a 3-foot cable. As a test I added the second chassis ground wire on the left, to the tab on the SHV connector. Everything is good now. Why was that second ground connection the difference between really bad performance and good performance? edit: this thing is sensitive to a fraction of a picocoulomb so even with a shielded proportional counter hooked up to it and the cover screwed on, I get spikes on my scope from static generated by moving around in my seat nearby, so I guess it's no wonder that weird stuff would cause it to get noise, but I figured the grounding situation was fine originally but I guess not??? edit 2:  I didn't take a picture of what it looked like before I added the second ground connection. That's a pulse from a proportional counter with a built-in alpha test source. Without the second ground connection the noise was so bad it would have made that pulse impossible to see, while with it the noise is more or less the same as the noise seen by the scope with nothing plugged into it. Why did that make such a difference???? BattleMaster fucked around with this message at 07:03 on Nov 2, 2019 |
|
#
?
Nov 2, 2019 06:40
|
|
|
|
| # ? May 9, 2024 15:13 |
|
|
Grounding stuff is hard enough to figure out normally, let alone when you're dealing with spooky poo poo like picoamps. My only suggestion is more blood sacrifices
|
|
#
?
Nov 2, 2019 07:03
|
|