|
Sagebrush posted:i'm a big fan of the "standard" on small wall-wart power supplies where the wire with the stripe is positive except when it's negative almost exactly 50% of the time. My "favorite" "convention."
|
#
?
May 29, 2017 01:38
|
|
|
|
| # ? Jun 4, 2024 16:42 |
|
|
Sagebrush posted:i'm a big fan of the "standard" on small wall-wart power supplies where the wire with the stripe is positive except when it's negative almost exactly 50% of the time. I was told the stripe was a minus sign to indicate negative. I am glad I checked. I still check three times every time I use one. I have a pair of Bluetooth earphones that I had to replace the lithium ion battery in. They'd used black wire for positive and red for negative. Whyyyy.
|
|
#
?
May 29, 2017 02:05
|
|
|
Thru got the pre wired batteries with opposite colors for $0.010 less than the correct wire colors. Duh.
|
|
#
?
May 29, 2017 02:29
|
|
|
Stabby McDamage posted:I've tried ferrules before, and even with the fancy crimper, I got pull-out of the ferrule from the connector in enough cases that we just switched to using solid wire instead. Maybe I'm bad/dumb? I've done a few thousand ferrule crimps with both Weidmüller (D-crimp), Knipex (Square and Hex) and Knipex knockoff (square) tools using both the cheapest possible aliexpress ferrules and some actually pretty nice ones and I've yet to have a single pull-out. That's with both fancy PTFE insulated tightly wound strands, general purpose Radox 125 and the very cheapest of silicone insulated cable straight from China. So unless your wire was really weird I can only guess that the wrong size ferrules were used with a crimper that won't crimp down (D-crimp can usually do 24 AWG in an 18 AWG ferrule if you're desperate; hex and square usually won't compress as much. Though I prefer square crimp generally since they work best with push-in terminal blocks). The only case I've had pull-out was crimping slightly large ferrules directly to component legs, which isn't really what they're meant to do but even that worked out once I switched to 26 AWG ferrules.
|
|
#
?
May 29, 2017 16:28
|
|
|
longview posted:I've done a few thousand ferrule crimps with both Weidmüller (D-crimp), Knipex (Square and Hex) and Knipex knockoff (square) tools using both the cheapest possible aliexpress ferrules and some actually pretty nice ones and I've yet to have a single pull-out. So this is basically my crimper, and I was using cheap stranded wire at something like 22AWG with cheap ferrules listed at 22AWG. That crimper puts three square crimps into the ferrule. Anything fishy about that?
|
|
#
?
May 29, 2017 16:53
|
|
|
Now I'm driving some steppers. I was going to buy stepper drivers, but they're like $5/channel, and I want six channels. The quad-stepper driver chips on the GM dash are apparently not available for some reason? Whatever, I've got about two dozen ATMEGA328PUs, so I'm rolling my own, using stuff I've got in my parts box. These motors are 12V bipolar steppers, drawing paltry current per winding. They seem to miss more steps than I'd like when driving them directly with 5v (using some 7406 hex buffers, so the power is there), so I'd like to use some 4049 CMOS Hex Inverters. Does this pullup arrangement work? I'm just doing standard pulse control, no half-stepping or anything. When I want to send a pulse, I'll set the IN_A or IN_B to low, have it hi-Z otherwise.
|
|
#
?
May 29, 2017 17:44
|
|
|
Stabby McDamage posted:So this is basically my crimper, and I was using cheap stranded wire at something like 22AWG with cheap ferrules listed at 22AWG. That crimper puts three square crimps into the ferrule. Anything fishy about that? That's one of the Knipex knockoffs, but it seems to have a non trademark infringing colour scheme unlike mine. From the factory the default crimp strength on mine was way too low (though they did stick it just felt far too weak), I adjusted it up about 3 notches or so.
|
|
#
?
May 29, 2017 19:47
|
|
|
ate all the Oreos posted:I've got kinda a weird issue about where to get odd parts. Specifically, my friend designed this peripheral she'd like to use with photoshop, basically just six mouse wheels in a row in a box, so she can map each one to a different keyboard shortcut (with the keyboard shortcuts controlling values like brush size, lightness, general photoshop stuff y'know). I also want the mouse wheels to be transleucent so I can light them with RGB LED's to indicate function or state or just make it look cool.... 1) it is very easy to get an Arduino to work as a keyboard input device. https://www.arduino.cc/en/Reference/MouseKeyboard trying to do this as a "mouse" might be more difficult, unless you meant you'd just send a different letter code combination for each of the six wheels' [up] and [down] signals. 2) mousewheels are cheap plastic wheels that wear out, because it's a microswitch that rubs around on a bunch of little plastic teeth. The teeth eventually get beat down and don't feel crisp anymore. When it gets old and dies you'll be doing this search all over again. If you use a 200-step stepper motor as a rotary encoder, it will NEVER wear out in this use. It will last virtually forever, since the "notches" that you feel are magnetic and frictionless. Plus, it is a 5mm or 6mm metal shaft, and has ball-bearings in it. And the motor is sealed, so it's never gong to get dirt or dust in it. When used as an encoder, a 200-step motor gives you 50 "steps" per rotation. And there is guides online for how to do that also: https://wemakethings.net/2013/02/18/steppers-as-rotary-encoders/ NEMA-17 motors are kinda big for this, but there is NEMA-11 and NEMA-8 sizes that are smaller and still have 200 steps. You could also use NEMA-17-size pancake motors and place them with six knobs facing up.
|
|
#
?
May 29, 2017 20:58
|
|
|
longview posted:That's one of the Knipex knockoffs, but it seems to have a non trademark infringing colour scheme unlike mine. After leaving the hardware store with no appropriate wire, I started thinking about twisting my own solid red black wire with some sort of hacked up machine. This means I need to step back and try to make ferrules work again, because that's bonkers.
|
|
#
?
May 29, 2017 20:59
|
|
|
edmund745 posted:Two suggestions, that you may or may not be aware of: or you could just buy a real rotary encoder for like a dollar fifty if you're gonna go that way https://www.amazon.com/CTYRZCH-KY-040-Encoder-Development-Arduino/dp/B015QSIRPK/ also some mousewheels (at least the ones in the Logitech mice I've taken apart) have a spoked slotwheel in the center and use optical sensing so they should never wear out. and finally, you can make an Arduino with an ATMega32U4 (like the Leonardo) behave as a joystick as well as a mouse or keyboard, with up to (I believe) 8 virtual analog axes and 32 virtual buttons.
|
|
#
?
May 29, 2017 21:05
|
|
|
ate all the Oreos posted:I feel I'm not qualified to answer about the peltier stuff since I've melted my last two peltier junctions I played with but as for thermistors, generally people just tape them to the surface using some kapton tape. You could also use thermal adhesive like this stuff: https://www.amazon.com/Arctic-Alumina-Thermal-Adhesive-5g/dp/B0009IQ1BU
|
|
#
?
May 30, 2017 06:32
|
|
|
Sagebrush posted:also some mousewheels (at least the ones in the Logitech mice I've taken apart) have a spoked slotwheel in the center and use optical sensing so they should never wear out. Yeah the nice ones do, I am not buying the nice ones because they're expensive. However, I figured out that you can buy the cheap poo poo mice and swap out just the rotary encoder for a (still mechanical, but) nicer rotary encoder and they're all standard dimensions so it seems to fit and work fine. So far I have those CYBER SNIPA ones, which have the full harness and wheel assembly usable with swapped-in encoder, one of those cheapo 99 cent mice from China that hasn't arrived yet, and a couple of replacement wheels for various RAZER gamer wank mice products. The replacement wheels are much nicer feeling than the other ones, but I'd have to build them some kind of bearing since they don't fit in the harness for the other wheel... Sagebrush posted:and finally, you can make an Arduino with an ATMega32U4 (like the Leonardo) behave as a joystick as well as a mouse or keyboard, with up to (I believe) 8 virtual analog axes and 32 virtual buttons. I put an ATMEGA32U4 right on the board and I'll need to try out different arrangements for control, I was considering making it a joystick like that (maybe in one mode?) but my friend is explicitly looking to use it with photoshop so I think it needs to output keyboard shortcuts for that unless I want to write a plugin...
|
|
#
?
May 30, 2017 14:33
|
|
|
Quick question about decoupling caps. I understand that for some situations you want a couple of caps of different sizes (this one datasheet I'm looking at says 100nF and 10nF) and I get why you need that and all that, but I'm wondering - should I make sure that the different size caps have different dielectrics? I assume I should try and make both as low ESR as possible, but should I try to make sure that the smaller one has less ESR than the larger one explicitly? Sorta related question: are the dielectric type codes (like X7R etc) any good indication of ESR properties? I read the wikipedia article explaining them and they seem mostly unrelated to ESR, but then I read a few datasheets saying "low-ESR X7R dielectric" which kinda sounds like they're relating the two?
|
|
#
?
May 31, 2017 14:03
|
|
|
ate all the Oreos posted:should I try to make sure that the smaller one has less ESR than the larger one explicitly? quote:Sorta related question: are the dielectric type codes (like X7R etc) any good indication of ESR properties? I read the wikipedia article explaining them and they seem mostly unrelated to ESR, but then I read a few datasheets saying "low-ESR X7R dielectric" which kinda sounds like they're relating the two?
|
|
#
?
May 31, 2017 15:11
|
|
|
You also need to check the capacitance-voltage chart for anything bigger than about 1µF for ceramics. Samsung and Panasonic are usually good about providing these charts for basically every single capacitor they make. X7R/X5R are often better at maintaining useful capacitance across voltage than other types but that designation only covers tempco and has no formal relation to the C-V characteristics. See for example Samsung 10µF/50V 1210 capacitors; IIRC those are around 1µF when biased to 24V.
|
|
#
?
May 31, 2017 16:52
|
|
|
Ah ok, thanks for clearing that up. I think this is the first project I've done where chips explicitly called for multiple decoupling caps and I wanted to make sure I got it right 
|
|
#
?
May 31, 2017 19:36
|
|
|
Thanks for the tips about soldering onto kapton, hopefully I don't botch it when the time comes to assemble this thing 
|
|
#
?
May 31, 2017 19:49
|
|
|
ate all the Oreos posted:I assume I should try and make both as low ESR as possible Sometimes high ESR is desired for stability reasons.
|
|
#
?
May 31, 2017 20:35
|
|
|
To further expand on that you can have problems with voltage regulators being unstable when paired with too low an ESR on the output (usually older LDOs, but the data sheet should always be checked). Newer regulators designed for digital use are usually designed with a maximum ESR instead of a minimum (e.g. DRAM termination regulators). Another concern for large power planes is resonances caused by interactions between different capacitor types, the details take a while to explain though (among other places it's covered in Right The First Time by Lee Ritchey). It's mostly an issue for powering high speed CMOS logic, newer ICs are actually getting a little better at it (see for example the Zynq-7000 series PCB design guide; you only need like 10 fairly high value ceramics since the rest is built into the package). Here's a pretty decent tool for calculating effective plane impedance based on what capacitors you put in: https://www.altera.com/support/support-resources/support-centers/signal-power-integrity/power-distribution-network.html Use the device agnostic tool, it's good practice to always use a tool like that on all power rails to make sure you know what frequency range decoupling is effective over.
|
|
#
?
Jun 1, 2017 17:00
|
|
|
Speaking of Magic Blue Smoke, I just blew up a cheapo LED strip controller real good. Like a flash and a fire good! Unsurprisingly, it's the MOSFET that poo poo the bed, apparently Vishay Si2302DS or knockoff. The resistor seems to be dead too, but overall it's probably salvageable, if totally not worth it. It's actually a great excuse to try to hack together a wifi controller with an ESP-01 and giant FQP30N06 MOSFETS I have laying around. Dunno if this could've burned the house down, I certainly thought so when I saw the bright flash in my peripheral vision It would probably been my fault too as I noticed that it was getting hotter after hooked it up to a more powerful supply, but didn't move it from the shelf where it was stuck quite tightly and probably wasn't getting much airflow.
|
|
#
?
Jun 1, 2017 20:36
|
|
|
When a stepper motor has a voltage rating, what does that actually mean? Is that just the voltage that would lead to a safe current, assuming an infinite current source and that the coil impedance is the only restriction? Because I have found these neat little skinny steppers that I kind of want to install on my 3D printer's extruders to save on weight and space, but they're rated for 2.9v and the printer runs on 24v. But the printer's stepper drivers are also current-controlled, so I figure I can just alter the firmware to supply only the 700mA per phase that the motors specify for full torque, since that's the real issue at hand. Right? It's not like I have to worry about arc flashover or popping a memory cell or something. e: oooor I guess I could just divide the specified voltage by the specified coil resistance and discover that it is exactly 0.7A. Derp. So, no problem running these on "24v" since the drivers should adjust that dynamically to maintain the required current? Sagebrush fucked around with this message at 21:56 on Jun 1, 2017 |
|
#
?
Jun 1, 2017 21:52
|
|
|
Sagebrush posted:When a stepper motor has a voltage rating, what does that actually mean? Is that just the voltage that would lead to a safe current, assuming an infinite current source and that the coil impedance is the only restriction? It's also the voltage which won't blow through the insulation of various parts (varnish on the windings etc)
|
|
#
?
Jun 1, 2017 22:09
|
|
|
Cumslut1895 posted:It's also the voltage which won't blow through the insulation of various parts (varnish on the windings etc) I know what you're saying is usually the case but now I'm just thinking about how much I'd like to see a varnish that can't withstand three measly volts Like maybe if it was only a few atoms thick?
|
|
#
?
Jun 1, 2017 23:07
|
|
|
ate all the Oreos posted:I know what you're saying is usually the case but now I'm just thinking about how much I'd like to see a varnish that can't withstand three measly volts Yeah at that voltage the only thing you're likely to damage semiconductor traces
|
|
#
?
Jun 1, 2017 23:16
|
|
|
Sagebrush posted:When a stepper motor has a voltage rating, what does that actually mean? Is that just the voltage that would lead to a safe current, assuming an infinite current source and that the coil impedance is the only restriction? So your stepper is rated 2.9V@.7A that's amm, 2W? At 24V, that's 85mA.
|
|
#
?
Jun 1, 2017 23:41
|
|
|
Sagebrush posted:When a stepper motor has a voltage rating, what does that actually mean? Is that just the voltage that would lead to a safe current, assuming an infinite current source and that the coil impedance is the only restriction? I've read from a few sources that you can ignore the voltage rating if using a current-limiting driver that's set properly, for exactly the reasons you suppose. That said, the sources were 3D printing people who are sort of notorious for blind engineering according to an instructable or whatever, so maybe there's some unknown issue with doing so, but I've never had a problem abusing steppers myself.
|
|
#
?
Jun 2, 2017 00:43
|
|
|
Stabby McDamage posted:I've read from a few sources that you can ignore the voltage rating if using a current-limiting driver that's set properly, for exactly the reasons you suppose. That said, the sources were 3D printing people who are sort of notorious for blind engineering according to an instructable or whatever, so maybe there's some unknown issue with doing so, but I've never had a problem abusing steppers myself. I've been running a set of 6v motors on 12v with a current limiting driver for the last few years, with no apparent problems. It is, of course, on a 3d printer.
|
|
#
?
Jun 2, 2017 00:50
|
|
|
silence_kit posted:Sometimes high ESR is desired for stability reasons. Yeah I've run into several LDOs which claimed to be unconditionally stable with low ESR ceramic capacitors and turned out to be bullshit. So I usually throw a 47uF tantalum/electrolytic cap on each rail to keep things stable.
|
|
#
?
Jun 2, 2017 14:49
|
|
|
Aurium posted:I've been running a set of 6v motors on 12v with a current limiting driver for the last few years, with no apparent problems. Run them on 30 volts. They'll work even better (no foolin). Sagebrush: the steppers will work fine. Run those little 2.9v thingies at 24v.
|
|
#
?
Jun 3, 2017 01:44
|
|
|
Sagebrush posted:When a stepper motor has a voltage rating, what does that actually mean? Is that just the voltage that would lead to a safe current, assuming an infinite current source and that the coil impedance is the only restriction? Running higher voltage just means that the stepper driver will adjust to a lower PWM duty cycle for the same current. You could estimate the duty cycle given the specs: assuming you run them at their rated current 0.7A, then 2.9V / 24V = 12% duty cycle (stall). Your torque is directly related to your motor current, so higher voltages won't give you more of that. The advantage of running more voltage is that it allows higher top speed. The inductance of the motor coils resists changes in current, producing back emf. At top speed the back emf is balanced with the supplied motor voltage, so by using higher voltage the motor can spin faster. Depending on the application (if you don't care about top speed) it would also work fine at 12V, and would likely be quieter. peepsalot fucked around with this message at 02:30 on Jun 3, 2017 |
|
#
?
Jun 3, 2017 02:25
|
|
|
Is mouser / octopart the right place to be looking for RGB LEDs? The selection seems real anemic and overpriced (theres none of the nice addressable ones, I can only find like one 5mm one and even at volumes of 10,000 they cost 50 cents each, while on eBay I can get 100 anonymous ones for $3, etc) is there another site people buy them at where you can get not-complete-poo poo ones for less than that? edit: Eh I guess I'll just buy 50 WS2812's from ebay and filter out the bad ones ahead of time, but I'd still like to know if there's a secret good LED site I'm missing out on Shame Boy fucked around with this message at 14:25 on Jun 5, 2017 |
|
#
?
Jun 5, 2017 14:01
|
|
|
ate all the Oreos posted:Is mouser / octopart the right place to be looking for RGB LEDs? The selection seems real anemic and overpriced (theres none of the nice addressable ones, I can only find like one 5mm one and even at volumes of 10,000 they cost 50 cents each, while on eBay I can get 100 anonymous ones for $3, etc) is there another site people buy them at where you can get not-complete-poo poo ones for less than that? Adafruit sells them. If you need fewer than 50, their price shouldn't be insane. BTW, WS2812B's are vastly better. Don't be fooled, like I was, by the minor difference in part number. It's basically the second generation part. Heck, you may have better luck finding them, too.
|
|
#
?
Jun 5, 2017 14:35
|
|
|
KnifeWrench posted:Adafruit sells them. If you need fewer than 50, their price shouldn't be insane. BTW, WS2812B's are vastly better. Don't be fooled, like I was, by the minor difference in part number. It's basically the second generation part. Adafruit seems to sell them for much more than ebay, but I guess that should be expected. And yeah I'm going for the B's, already experimented with these a while back and found that out through research, but thanks! e: Weirdly, the only supplier of them on octopart is adafruit, which is... suspicious, to say the least. I mean where's adafruit getting them and why can't I buy from them? (I assume "shenzhen" and "because you don't live in shenzhen" but thought I'd ask anyway )
|
|
#
?
Jun 5, 2017 16:29
|
|
|
ate all the Oreos posted:Adafruit seems to sell them for much more than ebay, but I guess that should be expected. And yeah I'm going for the B's, already experimented with these a while back and found that out through research, but thanks! You can get them on Alibaba, but standard delays/conditions apply. Last time I went down this rabbit hole, it seemed like Adafruit had secured the import monopoly, to the extent that they've even rebranded the part. E.g. Digikey will sell them... via Adafruit. Who knows if it's a natural monopoly or something in writing.
|
|
#
?
Jun 5, 2017 16:37
|
|
|
Dehumanize yourself and face to chinese eBay
|
|
#
?
Jun 5, 2017 17:03
|
|
|
Yeah just ebay that poo poo. I'd probably secure a real supplier for production, but for just dicking around, I've never had any issues. Would anyone mind explaining the difference between WS2812/WS2812B? I was actually just looking to buy a bunch (probably on a strip) myself. E: another thing. I had an ESP-01 laying around and after blowing up my LED controller (see previous post), I hacked together new IoT one with leftover FQP30N06s mosfets. They're way overkill though as I need like 3 amps drain current tops so I'm looking for a smaller replacement, but all I can see in compatible G/D/S throughhole packages is either very similar, or wimpy little TO-92 stuff that's no good. Any ideas? I suspect that I just suck at searching... mobby_6kl fucked around with this message at 20:17 on Jun 5, 2017 |
|
#
?
Jun 5, 2017 19:35
|
|
|
What kind of failure mode would cause an op-amp to only output 2-2.5v continuously? Supply voltage is 9.5v, + input is a 0-5v source and - input is feedback from the output. I had a bitch of a time diagnosing a broken control panel for our analog controlled lighting dimmers. The circuit is supposed to output 0-5v on 5 different wires corresponding to different zones of lights, controlled by slider pots that form a voltage divider between a +V rail and ground. The master fader makes a divider between the output of a 5v regulator and ground, with the output going to the + input of the op-amp. The output of the op-amp sets the voltage on the +V rail that the 5 other pots take, so that all 5 channels are scaled by the setting of the master fader by setting the upper limit on their voltage. What was happening was that the op-amp was outputting 2-2.5v based on the 0-5v swing of the master fader, instead of the expected 0-5v output that it should have. Granted this whole system is from 1990 so I'm not surprised the chip gave up the ghost, but I am curious for my own understanding how it failed. I ended up desoldering the chip and cannibalized an equivalent chip from a similar panel left over from one of our other venues, which fixed the problem entirely. But that one is from 1992 so I'm already ordering a new chip anyway. Having learned all this about how the circuit works, I'm starting to plan out a circuit with an arduino to control the op-amp output from a DMX input or over IP, with a highest takes priority arrangement so the manual fader can override the arduino output. edit: I put the bad chip on a breadboard and rebuilt that part of the circuit to probe it with my scope, definitely something funny going on here: Blue trace is the input, yellow is the op-amp's output.  Good chip  Bad chip. The flat spot on the input towards the end is my screwdriver slipping off the trim pot I was using instead of a fader pot. ickna fucked around with this message at 01:10 on Jun 6, 2017 |
|
#
?
Jun 5, 2017 19:45
|
|
|
mobby_6kl posted:Yeah just ebay that poo poo. I'd probably secure a real supplier for production, but for just dicking around, I've never had any issues. From what I've been reading the B model has different timing and different timing tolerances, here's the article I got that information from: https://cpldcpu.com/2014/01/14/light_ws2812-library-v2-0-part-i-understanding-the-ws2812/ I'm not sure if it's more stable or otherwise "better" but it seems to be fundamentally different, behavior-wise, so it could be
|
|
#
?
Jun 5, 2017 19:52
|
|
|
mobby_6kl posted:Yeah just ebay that poo poo. I'd probably secure a real supplier for production, but for just dicking around, I've never had any issues. WS2812 is a 6-pin configuration that takes a separate power supply for I/O and LED WS2812B is a 4-pin configuration that ties them together internally for convenience/easier routing WS2812 has timing windows on the 1-wire signal that end up being different durations for 1 vs 0 bits, neither of which aligns with the nominal duration. (Chalk it up to lovely documentation, but it matters) WS2812B has symmetrical timing windows that are easier to sketch out, plan for, and bitbang. If you're buying a strip and using a pre-fab controller or someone else's (i.e. Adafruit's) Arduino driver, the differences are academic, but WS2812B, being a newer product, is easier to find and easier to assume will be supported going forward.
|
|
#
?
Jun 5, 2017 21:40
|
|
|
|
| # ? Jun 4, 2024 16:42 |
|
|
I know a guy that built a business around them. Here have a Hackaday link: http://hackaday.com/2017/01/25/ws2812b-led-clones-work-better-than-originals/ Adafruit doesn't have an import monopoly or anything, they're just one of the only legit US companies selling grey market Chinese goods. The manufacturers are almost exclusively selling them through Alibaba, and Digikey isn't going to supply anything from there.
|
|
#
?
Jun 6, 2017 01:11
|
|
