|
Foxfire_ posted:What function are you trying to accomplish with the grounding? ESD protection mostly, less than six months ago I had a thunderstorm take out specifically and only a port on my router, cuz (as far as I can tell) it found a path through the delicate SFP module (the thing that died), into the shielded ethernet, through the modem, through the coax shield to the outside. I figure the router has a prominent grounding lug for exactly that reason.
|
# ? Oct 23, 2023 19:51 |
|
|
# ? May 31, 2024 19:23 |
|
Is there a more clever way to hook up a comparator to the output of an optocoupler? The most straightforward seems to be, pull up resistor, the photodiode pulls to ground, feeds into one of the inputs of the comparator, and then the other end (swap +/- at your preference) to some kind of bias resistor divider, possibly with a trimmer. Just seems kind of a lot of parts. I feel like there must be a simpler version. In this case the input to the optocoupler is a sine wave, so the current varies from basically max for the coupler all the way down to zero essentially, trying to do zero-crossing detection. Otherwise it would indeed be simpler to ensure everything is fully saturated all the time.
|
# ? Oct 24, 2023 01:26 |
|
Draw a schematic
|
# ? Oct 24, 2023 01:36 |
|
Shame Boy posted:ESD protection mostly, less than six months ago I had a thunderstorm take out specifically and only a port on my router, cuz (as far as I can tell) it found a path through the delicate SFP module (the thing that died), into the shielded ethernet, through the modem, through the coax shield to the outside. I figure the router has a prominent grounding lug for exactly that reason. Current coming in a signal wire, across some internal protection device to the chassis, out the ground lug, through a wire pigtail, then through stuff to go back to where it comes from is going to be a big loop area and fast current isn't going to want to do it vs jumping somewhere via capacitive coupling Protection against things like lightning induced common mode voltages mostly have to be designed into the circuit, they're hard to bolt on after the fact unless you can stick something in front that has the protection you wish they put into the main device. (By ESD, I meant more like a connection meant to provide a path for charge to trickle through to stop leakage currents from slowly charging the case up to voltages that can spark. There you're looking at slow DC currents so you don't really care about the wire inductance.)
|
# ? Oct 24, 2023 01:40 |
|
Shame Boy posted:ESD protection mostly, less than six months ago I had a thunderstorm take out specifically and only a port on my router, cuz (as far as I can tell) it found a path through the delicate SFP module (the thing that died), into the shielded ethernet, through the modem, through the coax shield to the outside. I figure the router has a prominent grounding lug for exactly that reason. Thats pretty neat. Run two wires from the ground lug of the plug. One to the case and one to the the coax connector. Coax grounding blocks are dirt cheap. IDK much about EMI/Lightning but I think you're not supposed to daily chain because of ground loops or something so run them both to the same place. Someone smarter than me (most people ITT) can correct that if wrong.
|
# ? Oct 24, 2023 01:42 |
|
Foxfire_ posted:Transients are harder to protect against. I wouldn't really expect a case grounding lug was meant for that, you'd normally do that right at the entry with the signals being protected passing inline through the protection. I will also add that it's a prominent ground lug which they even went to the trouble of giving a wingnut for easy use and a nice little stencil to ensure you know what the gently caress it is, so even if it's ultimately not going to help with transients it still feels wrong leaving it unconnected
|
# ? Oct 24, 2023 02:23 |
|
I realize I'm just doing vibes-based electronics here but it seems like someone went through a lot of trouble to design in a nice little easy to use ground point that's just begging to be properly wired up.
|
# ? Oct 24, 2023 02:26 |
|
Shame Boy posted:I need to ground my network stack in my closet (my switch has a dedicated grounding lug for it since otherwise everything uses isolated power brick supplies) and I'm trying to figure out a good way to do that. The difficulty is there's only one wall outlet in there, and it's being taken up by the UPS power cable. I guess I could plug something like this into the UPS: Could you run an outlet strip off your UPS outlet and plug both the grounding thingy and whatever else needs UPS protection into it?
|
# ? Oct 24, 2023 03:07 |
|
Update on my servos from a while back - either I wrecked the servo that was twitching, or it was bad in the first place, because after changing the signal frequency it still only twitched, but switching the same wiring and signal over to the *other* supposed-to-be-identical cheap servo has that one responding as expected now.
|
# ? Oct 24, 2023 04:14 |
|
Bad Munki posted:The days of plumbing being a reliable source of ground are long gone. Yep. PEX is here to stay. You're only allowed to ground to metal water pipes now within 6 feet of the service entrance where those 6 feet can be visually verified. Also, depending on how your electric panel is grounded, couldn't you be making a ground loop? kid sinister fucked around with this message at 04:22 on Oct 24, 2023 |
# ? Oct 24, 2023 04:19 |
|
Rescue Toaster posted:Is there a more clever way to hook up a comparator to the output of an optocoupler? Could you replace the comparator entirely with a schmitt trigger inverter (like 1 unit from a 74hc14)? I'm having trouble picturing an application (this is in no way me claiming it doesn't exist) where you would need a very precise threshold after you've distorted the waveform with an opto. If you need to use the comparator, my one tip is that it's often good to have some extra hysteresis to make it less sensitive to noise during the relatively slow transition from fully-off to fully-on and back (the comparator will have a few mV of hysteresis already built-in). If your voltage divider is on the non-inverting comparator input, an extra resistor from the comparator output back to the divider gives you this, with the size of the resistor vs the equivalent resistance of the divider setting the amount of hysteresis.
|
# ? Oct 24, 2023 04:22 |
|
So for a zero-crossing detector for 120v (or 240v) AC, you have some large voltage drop resistors, but as a consequence, in order to not overload the LEDs at the 170v (or 340v) peaks, the LED current drops very low as you get close to the zero crossing point. In order to narrow the off time, you want to make the circuit as sensitive as possible. So ideally you can detect even a very small amount of current in the photodiode receiver side. I have an extra channel of a 74x14 that I'm going to hook up too, this is just a prototype board. But I thought I would populate some kind of comparator circuit too, so I can observe them both on a scope. My thinking now is basically a pull up on the photodiode, feeding one input of the comparator, with a positive feedback for hysteresis, and bias provided by a trimmer across the rails and feeding the other input. Biased close to the positive rail, even a small amount of photodiode current will cause the comparator to trip then. I just have to balance pull up strength vs switching speed vs the positive feedback input impedance. I just thought maybe there was a more clever way to wire the photodiode and comparator together. In software you can do tricks like counting the 'dead time' and interpolating where the zero-crossing likely is during that time. Though for other reasons ideally I would like to get a hardware interrupt to fire as close as possible to the zero-crossing.
|
# ? Oct 24, 2023 16:45 |
|
I know you only asked about comparators but another thing maybe worth trying if you're concerned with small currents is a 2n3904 or similar with its base on the opto like this: That way you get a ~100x multiplier on the current up front. That's the difference between a 50mV swing and a rail-to-rail swing for a given resistor value. Don't want the extra package on the board? You can buy "darlington-output" optocouplers that have the extra transistor built-in.
|
# ? Oct 24, 2023 17:29 |
|
We're so surrounded by super high end FET poo poo it's easy to forget the simple current amplification of a humble BJT. Good idea, thank you. I'm using a particular bidirectional (two LEDs) opto instead of a bridge, and it didn't have an integrated darlington version, but I have plenty of room for a SOT-23 bjt.
|
# ? Oct 24, 2023 19:09 |
Stack Machine posted:I know you only asked about comparators but another thing maybe worth trying if you're concerned with small currents is a 2n3904 or similar with its base on the opto like this: Need to be careful here. Once you include the few pF of capacitance across the isolation, you'll find that it's sensitive to common mode transients as well. There are some isolators which have faraday screens over the phototransistor which are suitable for this.
|
|
# ? Oct 25, 2023 12:17 |
|
ANIME AKBAR posted:Need to be careful here. Once you include the few pF of capacitance across the isolation, you'll find that it's sensitive to common mode transients as well. This is a good point. If a common mode transient goes positive, the base-collector junction of the NPN will forward bias and the decoupling network on VDD should provide good protection (but it's a good idea to have a ~1nF VDD-ground cap near the opto since these transients are fast and PCB traces have inductance and resistance). If you're looking for the best protection possible, you can help this out even more by putting a 1k resistor between the opto and the base of the NPN. Normally it won't do much since the current is small, but if there's a sudden transient it'll dissipate energy that would otherwise go into the NPN's base. Negative-going transients are more of a concern: the base of the NPN is in danger of reverse breakdown. A reverse-biased diode from the NPN base to ground would keep it safe. The VDD side, like before, will be protected by its decoupling network.
|
# ? Oct 25, 2023 14:14 |
I’ve got a couple of this light projector thing I got at Fleet Farm on Halloween clearance: It shines light and spins, I love it: It’s fixed orange, though. I’d like to put an LED that’ll change color in there, but that’s maybe later. In the immediate term, I just want to swap the LED in one of them out for, like, purple or green or something. Without having any specs on the LED it came with, what’s the best way to spec an approximately okay replacement? I could pop the existing one off and measure the voltage provided I guess, and beyond that…?
|
|
# ? Oct 26, 2023 04:01 |
|
Physically measure the dimensions and then also look at what colour it is. The colour will dictate the voltage drop (measuring it with a multimeter while the LED is on won't hurt either) and then look up those specs on Digikey Orange LEDs aren't really common, but it might be, like, a phosphor coated white LED at 2000K. It'll be 2.7-3.2V if so
|
# ? Oct 26, 2023 04:20 |
Yeah, I did notice it’s definitely not amber, it’s a really serious orange, I was assuming there was probably something else going on. I’ll crack it open tomorrow.
|
|
# ? Oct 26, 2023 04:26 |
|
First time designing my own PCB and I'm hoping I can get someone to take a look at it and tell me if I'm making any major mistakes. The project itself is a button box for a flight sim along with an OLED display. It's driven by a Teensy 4.0. I've got a working prototype that is a mess of wires, and I got it in my head that I should make it into a PCB instead. I'm planning on using the Teensy to power everything off 5v USB. Switches are done as a matrix, and the 4-way switch is hooked up to a single analog input using the circuit I copied from a part from a joystick I bought. The display and dials will be hooked up via JST connectors so I can have more freedom to design the case. Button positions are meant to mimic this thing Here's the link to the whole album https://imgur.com/a/y01bGBe I guess first thing I should do is remove the middle pins off the teensy footprint since I'm not actually using any of them. Next steps will be some CAD work on the case to get the proper mounting distance from the front panel for the switches, and maybe adding a couple standoffs.
|
# ? Oct 29, 2023 07:39 |
|
The basics: - Mounting holes - Bigger radius on rounded corners - Some silkscreen with your name, the project name, revision, and date Less basic: - Your red signal wires could stand to be thicker. You have the space, let them be chunky. 0.25mm or 0.4 maybe. - Your blue ground plane is all broken up. Imagine that for each red trace, the signal wants to return unbroken along the blue plane. There are a few places like to the right of the RCL silkscreen where if you just extend the distance between the two vias, then the pour on both sides will have space to connect to each other. Similarly, I'd probably stitch the pours near SW19 silkscreen with a top layer red trace going between them. - For areas where you have lots of space, like SW5 and 6, push the passives a little further away. Give yourself some room to solder. - What's your hole size on a lot of these? What does your boardhouse allow?
|
# ? Oct 29, 2023 17:42 |
|
Anyone ever heard of 'encroached' vias? https://ww1.microchip.com/downloads...es-00001843.pdf The idea I guess is on the bottom side of your QFN EPAD vias, you bring the mask super close to the hole without tenting over it. That way flux gasses and poo poo can escape, but there's not enough surface to wick significant solder out on the bottom. At least for the guys like jlcpcb I think it's cheap enough to just get epoxy plugged & plated vias rather than screwing around with this, and they seem to want at least 0.1mm-0.15mm copper around via holes, not sure if that's a hole/soldermask registration thing or what. My only other thought was just do tented vias as normal, and poke them open with a needle or something by hand if you're doing your own QFNs, but again the cost of epoxy plugging is cheap enough to probably make that not worth the time/effort.
|
# ? Oct 30, 2023 02:29 |
Yeah encroached vias are a thing, but lol at microchip suggesting they're better then filled/overplated vias. One thing to be mindful of is that encroached vias will thieve some solder (supposedly much less than normal vias though), so their recommended stencil openings are probably a bit bigger to compensate. If you use properly filled vias then their suggested stencil design might have too much solder. Edit: oh and another caution about encroached vias. Some fabhouses will ignore this sort of stuff when they run their DFM scripts. Same goes for tented vias. Trust but verify. ANIME AKBAR fucked around with this message at 13:00 on Oct 30, 2023 |
|
# ? Oct 30, 2023 12:56 |
|
I want some wake-up lighting in my bedroom. My idea is to install LED strips and diffusers around the top of 3 of the walls, besides the one I’m facing, to provide the right kind of light for waking (6500K???). Then I put a Nano or something between the strips and power, and have it turn them on at a preconfigured time. Anyone done this before? If so, how well does LED strip lighting work as wake-up lighting? Is it strong enough?
|
# ? Oct 31, 2023 19:21 |
|
It probably wouldn't be bright enough on its own, depending on how heavily you sleep. More lights is always the answer
|
# ? Oct 31, 2023 19:31 |
|
I have a single 5 meter strip of warm white (probably 3000K or 4000K) LED's set up almost exactly as you've described except it's wrapped around the trim of my closet facing parallel with the wall so that you can't see the diodes themselves. It is bright enough to wake me up without an alarm which is especially helpful in the dead of a Midwestern winter but it is smaller bedroom about 11ft by 11ft. I used an old TC420 LED controller since I already had it and didn't need anything with wifi, bluetooth, an app, or any of that jazz. Just set it to gradually turn on at 6:30, max brightness by 7, and I usually wake up around 6:50.
|
# ? Oct 31, 2023 21:11 |
|
ante posted:It probably wouldn't be bright enough on its own, depending on how heavily you sleep. Good point. I also have a habit of burrowing under covers if it’s cold, especially cuz I use a sleep mask and I like to cocoon. Maybe those would have to be REALLY bright lights? Wandering Orange posted:I have a single 5 meter strip of warm white (probably 3000K or 4000K) LED's set up almost exactly as you've described except it's wrapped around the trim of my closet facing parallel with the wall so that you can't see the diodes themselves. It is bright enough to wake me up without an alarm which is especially helpful in the dead of a Midwestern winter but it is smaller bedroom about 11ft by 11ft. I used an old TC420 LED controller since I already had it and didn't need anything with wifi, bluetooth, an app, or any of that jazz. Just set it to gradually turn on at 6:30, max brightness by 7, and I usually wake up around 6:50. That’s more or less what I was thinking, except either right-angle diffuser channels along the walls or along the window trims. I suppose this does all come down to how well I get woken up by an LED strip at that brightness. Maybe I’ll order a smaller sample and see if that gets me close!
|
# ? Oct 31, 2023 21:41 |
|
I have a remote control for my projector that broke. I could hear some stuff rattling around inside so I cracked it open and discovered a bunch of components had basically come loose from the PCB somehow. The resistor I'm sure I can solder back on, but there's a little thing next to it - that black piece with an "N" stamped on top - that I don't recognize or have an easy way of identifying. Is there any way to figure out what needs to be soldered back into that spot? The piece with the N on it is clearly busted, there are no contacts on it anymore. Just wondering if I can figure out how to fix this without spending a ton on a whole new remote.
|
# ? Oct 31, 2023 23:40 |
|
Pollyanna posted:Good point. I also have a habit of burrowing under covers if it’s cold, especially cuz I use a sleep mask and I like to cocoon. Maybe those would have to be REALLY bright lights? Yeah I think you would need a lot or very bright LEDs if you wear a sleep mask. I use blackout curtains and blacked out all my other electronics/clocks/etc. Maybe a nice 50W spotlight pointed away from the bed? For what it's worth, I found that being able to see the actual diode was much more obnoxious and alarming when waking up than when it was diffused along the wall and I had no direct line of sight. So the right-angle diffuser would need to be quite opaque. I tried some frosted window film on normal glass in a previous iteration which had the lights pointed more towards me and that was NOT diffused enough. Probably need some nice sign white acrylic or whatever.
|
# ? Oct 31, 2023 23:52 |
|
Harvey Baldman posted:I have a remote control for my projector that broke. I could hear some stuff rattling around inside so I cracked it open and discovered a bunch of components had basically come loose from the PCB somehow. The L1( N )'s an inductor and the C3 is a capacitor. I'm guessing they're part of some kind of boost converter so you'd need to know the original value of it to make it work correctly. Can you get a side lit shot of that big IC on the board? Might be able to look up its datasheet if we can see its id code. Are there any components on the other side of the board?
|
# ? Nov 1, 2023 00:18 |
|
Hexyflexy posted:The L1( N )'s an inductor and the C3 is a capacitor. I'm guessing they're part of some kind of boost converter so you'd need to know the original value of it to make it work correctly. There aren't any components on the reverse side from what I can tell, though there's a membrane that lays over it and presumably presses down contacts on this thing: Here's that bigger IC: Pretty sure it says "AD009-01T" on the top line, and "4847.1ACP3" on the second line, to the best of my ability to make it out. That 2nd line comes up with nothing, but the first line seems to get me to here. The capacitor and inductor have no discernible markings beyond the "N" on the inductor. Also, FWIW there is this: I did search that out but didn't find a replacement remote option, and since the projector this is for is like ~$2,800 if I can fix the remote I'll feel like an absolute fuckin' hero. We have tried all kinds of alternative approaches like an IR blaster or apps on our phones but nothing seems to work with this particular projector other than this now-busted remote. That said, the pads on that "L1" spot on the PCB where I guess the inductor goes look pretty hosed up so I'm not sure I have anything to even solder to. Harvey Baldman fucked around with this message at 00:45 on Nov 1, 2023 |
# ? Nov 1, 2023 00:42 |
|
I think U2 is a buck regulator, so whatever voltage it's making is less than 3V. It looks like C2 is the input capacitor across the battery and C3 is the output capacitor and C3 is also in parallel with C1, which looks like the decoupling capacitor for the microcontroller. (I think, from the photo. Check the traces. I'm basing this off an assumption/guess from the photo that L1 is connected between a pin on U2 and a pin on C3, C2 is in parallel with the battery, and C1 is in parallel with C3. edit: If the battery's across C3 instead of C2 it's a boost, see final paragraph) It looks like that microcontroller will run down to 1.8V, so the output is maybe somewhere around there. If you asked me to design a remote like this, though, I'd run the LED directly from the battery so if that's what they're doing you can't just go find a 1.8V supply and bypass the buck regulator. What you could do, though, just to make sure it still works, is solder some short pigtail wires across C1 and power it up with an external 1.8V supply and a 3V supply (or the battery) where the battery goes. Without the inductor there, the buck converter will just be out of the picture. Then you can just check that it still works (remember, a cell phone camera can usually see an IR remote so you won't need to be near the projector). If that works for you, you can try to get the buck regulator working again or figure out an alternative power supply. The microcontroller will run at up to 3.6V according to its datasheet so it's possible that it would work just fine with the buck converter bypassed. If you get to the end of your rope and want to try one last thing, you can bypass it by connecting the battery positive lead to the non-ground side of C3/C1, running the microcontroller at 3V instead of 1.8 or whatever they designed it for. It shouldn't burn out the microcontroller but might eventually (or even immediately) damage something else. It will definitely kill battery faster if you do this. Edit: if C3's across the battery, which I'm warming up to looking at your photo again (but it's really hard to tell if the trace goes under the silkscreen on the left side of C3), then it's a boost, the microcontroller and IR LED are both running at 3V directly off of the battery, and C2 is the output of the boost and is higher than 3V. If this is the case, just look for what connects to the non-ground side of C2. I can't imagine what would do that unless there's a boost converter there for the not-installed laser pointer module. If that's the case you'd be lucky and can just scrape U2 and C2 and L1 off the board since they'd be doing nothing of value. Stack Machine fucked around with this message at 04:05 on Nov 1, 2023 |
# ? Nov 1, 2023 02:36 |
|
I think the switchmode power supply is the chip U2 with marking AX21. I'm not finding much online though. Edit: it might be this one: https://www.lcsc.com/product-detail/DC-DC-Converters_NATLINEAR-XT1861B332SR-G_C3007998.html From the datasheet for that part, it's suggesting a 10uH inductor, an input capacitor of 10uF and output capacitor of 22uF. The input capacitor will be sharing a net with the inductor, so I think C3 is Cin and should be 10uF. From what I can see, and the pinout for the SOT23-5L in the datasheet, I am pretty sure it matches your circuit To determine precisely what parts you need to buy, if you could measure the distance between the pads of C3 and L1 we should be able to help you pick some specific models. This is absolutely salvageable though. Splode fucked around with this message at 04:27 on Nov 1, 2023 |
# ? Nov 1, 2023 04:13 |
|
Splode posted:I think the switchmode power supply is the chip U2 with marking AX21. I'm not finding much online though. Annoying thing is I swear I've seen one of those mentioned on eevblog. E: Oh neat, I couldn't figure out which one of those alternate part names was the thing.
|
# ? Nov 1, 2023 04:30 |
|
Splode posted:I think the switchmode power supply is the chip U2 with marking AX21. I'm not finding much online though. If it's this, AX21 is the code for the 3.3V out version and you can bypass it by shorting the left pin of whats left of C2 to the left pin of C3. I have no idea what it's supposed to power but it's hard to damage anything you'd find in a remote by undervolting it. Maybe it drives the high side of the IR LED to keep its brightness consistent as the battery dies. Neat but I bet it'd work (at least with a fresh battery) without it while your parts are on the way. You should make sure the board matches the schematic before you try that, of course. C3 would be the capacitor on the left and C2 the one on the right.
|
# ? Nov 1, 2023 05:06 |
|
Ne Cede Malis posted:I would not use a screw terminal to terminate a stripped AC power cord. The risk of one of the wires slipping out is just too high. {edit} The industry term for such a cable is a "suicide cable" btw. Hey, so it doesn't go unsaid, thank you for the feedback -- definitely taking all the safety info into consideration, and it's good to know I'm not far off base in some of my assumptions. Work got in the way of any hobby stuff these past few weeks but I'm going to try to finish this in the upcoming week so appreciate the feedback.
|
# ? Nov 1, 2023 13:33 |
|
I was looking at this video of drag soldering: https://www.youtube.com/watch?v=Fwx_qzfwdgk I use the same solder, Kester 63/37 (mine is rosin core, not sure about his) and I set my TS101 just a little hotter than the melting point, but the flux in the core of the solder and any flux I apply with a pen pretty much instantly go up in smoke when I start soldering. How is this guy able to solder for so many passes without flashing off all that flux? Is there something magical about the quality of the flux being used? Is it the sheer volume of it, like by virtue of being a gel he can goop on a lot more of it in the first place?
|
# ? Nov 2, 2023 15:40 |
|
Yeah you say flux pen, is it one of those that's like a marker, that you press down and it dispenses the really thin flux? You definitely want a good gel flux in a syringe if you're going to be trying to drag solder. Also I don't know that soldering iron, I don't think there's any way to know for sure what the temp is without a thermocouple. I use chip-quik gel no-clean flux and a metcal iron with a hoof tip when dragging soic/tqfp/tssop parts and I'm pretty happy with it. I still get bridges on 0.5mm when dragging though. My chip-quik hangs around a while but still not as well as that video seems to show. I was thinking about trying amtech flux next, see if that's better.
|
# ? Nov 2, 2023 16:49 |
|
I like the flux pens
|
# ? Nov 2, 2023 18:42 |
|
|
# ? May 31, 2024 19:23 |
|
Did some testing with my variac today (still exposed and no enclosure finished). Tried to see what would happen to one of my LED christmas bulbs if I ran it on lower voltage. Well what happened was when it got down to around 120V the led exploded, well everythign was contained in the plastic bulb but I saw it, started to flicker and then pop! So nah couldn't dim them with voltage, which I figured. The strange bit is the variac hummed, hummed more as I lowered the voltage and it got quite hot actually near the start of the windings. Never done that before, never hummed. Let it cool a bit and I tried lowering the voltage real low, then it turned on again and I thought I saw a tiny puff of smoke. Then it behaved normally again. Haven't tested it again but I am wondering if there was some kind of short, perhaps a sliver of metal or something got on the exposed part of the windings? It no longer vibrates or gets hot. I haven't tested it with a load but I will, IIRC it did humm without a load too.
|
# ? Nov 14, 2023 18:21 |