Image limit reached >>2996880>I'm new to electronics. Where to get started?It is an art/science of applying principles to requirements.Find problem, learn principles, design and verify solution, build, test, post results, repeat.Read the datasheet.>Comprehensive list of electronics resources:https://github.com/kitspace/awesome-electronics>Project ideas:https://hackaday.iohttps://instructables.com/tag/type-id/category-technology/https://adafruit.comhttps://makezine.com/category/electronics/>Books:https://libgen.is/>Principles (by increasing skill level):Mims III, Getting Started in ElectronicsGeier, How to Diagnose & Fix Everything ElectronicKybett & Boysen, All New Electronics Self-Teaching GuideScherz & Monk, Practical Electronics for InventorsHorowitz and Hill, The Art of Electronics>Recommended software tools:KiCAD 6+CircuitmakerLogisim Evolution>Recommended Components/equipment:OctopartLCSCeBay/AliExpress sellers, for component assortments/sample kits (caveat emptor)Local independent electronics distributorsladyada.net/library/procure/hobbyist.html>Most relevant YouTube channels:EEVblogW2AEWMoritz Klein>microcontroller specific problems?>>>/diy/mcg>I have junk, what do?Can it run Doom?>consumer product support or PC building?>>>/g/>household/premises wiring?More rules-driven than engineering, try /qtddtot/ or sparky general first>antigravity and/or overunity?The doctor prescribes .45 ACP brand aspirin STAT.
>>3002473I picked up a Sony KV-9400 TV. Picture from Reddit but mine is exactly the same. I'd like to use it for old computers and video games but the picture quality has issues. It does not have a plain RF jack or composite inputs. I have to use a UHF/VHF matching transformer and then an RF modulator for the original device. Something along the line is causing interference as I get a bit of snow in the picture. How can I improve the quality? Is there a way to ditch the matching transformer? Or shield everything? Or add composite inputs directly to the TV (not an RCA to coax converter box)?
>>3002460WINRAR!Luckiest fucker around award. Well... other than having to try so many. But this might be IT, the ONE single fault. No Vcc on the IGBT driver - > no output.16V 220uF. Dead short.
>>3002479Awww dang. I hardly ever buy electrolytics because they can suck my dick. But what to put in there now. Got some fresh 1000uF 16V, shouldnt hurt on just Vcc to GND. Got a bunch of used ones. Could fuck alot of films on there.Since we can't make 16V films, why dont we just put a kV film on a transformer. Im sort of a genius you know.
>>3002481I'm not convinced it's needed. Don't ask me why it died. But there's nothing nearby that warrants this. A 5k R going to a switch or knob on control panel. S8050 collector, emitter is on a 1.3k R. The one 4093 is already closer to the large cap next to the reg and frankly idk why anyone would put the cap there, it's not bypassing shit.
>>3002476I’d try shielding and clip-on chokes, but if anything I’d want to go for VGA. The TV is converting a composite signal into RGB components and audio and sync signals anyhow so there’s no real reason to settle for an inferior composite input. But adding component or VGA inputs is likely more invasive than adding composite input.
>>3002501Wait, TVs from 45 years ago can have RGB input?
>>3002505I don’t see why such a mod wouldn’t be possible. The RF input gets demodulated into composite video, which gets split into amplitude and a chrominance signal, which then gets demodulated into red green and blue signals, which then get amplified and sent to the electron tubes. The more superfluous steps you can eliminate, the better the fidelity.SCART is an RGB video input and was in use since 1978.
>>3002490I sort of rigged it. I'm relying on my belief that the thing just clocks through and has little faukt and error checking. I disconnected the EMI module and connected only the auxiliary transformer. The power section isn't getting mains, the pulse transformer stage is still on the bench, I am just hoping the machine cant tell and I will get gate drive for every power transistor. Also curious to poke around a little once it's going.Fingers crossed for the smoke to stay in and signals to be there.Swapped a new 1000uF 16V cap in there as it's just bulk capacitance, unlikely to cause issues.
>>3002524It came on, some relays click on startup, idk what LEDs on logic board mean, some green ones one red one on others off, I'm sure some are clearly there to indicate gate drive. Warning light on front panel is on. The 555 on repaired board is ticking away happily at the set AC frequency and the AC DC relay clicks with the switch. Odly the stick TIG switch doesnt overwrite it as I've expected. Adjustments to currents and all that are reproduced by the segment display. No gate drive anywhere but I think in case of the pulse transformer it might be that the feedback loop is just maxed out or it comes on on demand only. Logic board is doing suprisingly little. Suprising because the general strategy for almost everything seems to be to have a timer running and the timer gets connected on demand instead of starting to run on demand. I would have expected a generator for the switching frequencies but perhaps its all current loops with no fixed frequency. I'll rig a switch to fake the torch but I suspect the boars does a bit more error handling than I hoped.
>>3002515I did some checking on this TV. Schematics are available. Picture related. It is a hot chassis. Everything is AC. This guy would have to use some kind of optical isolator to add any kind of new input to the system or the grounds on that input would be live 120 volts AC. Heck, they say to use an isolation transformer to even work on the thing while it is plugged in.It also uses some tiny jungle chips and I can't find a datasheets for most of them. CX556 has a tiny, shitty pinout available. It appears to be an NEC uPC1352C chip. Pins 26-28 are RGB out but there is no RGB in. It does have luma (5), chroma (13), and blanking (24) inputs though. And 'Demo input' (25), whatever that is. The only RGB mods I have done always tap into the RGB inputs on the jungle chip.This would be a hard row to hoe for someone that hasn't do this kind of stuff before and he's just asking how to get rid of snow over RF. Heck, how would he even get RGB out of old consoles and computers? AFAIK things like the 8-bit consoles were all RF or composite. 8-bit computers like Atari's and Commodore's only did Chroma/Luma at best.
>>3002530Yeah I never said it would be easy. As for the RGB source, VGA. You’d also need to handle the sync signals.
>>3002529The temperature switch from the power board fixed the warning light, duh.Found a 555 that reacts to the pulse frequency. The pulse width pot is routed close to it too. Just like the up and down slope settings. I know there's analog switches. I have a theory where the analog switches multiplex through the adjustment pots accordingly to the phase of the process the machine is in. This would make sense, slopes aren't pulses so they probably have a S-trigger that is equally multiplexed in, so they can make slopes into pulses but I can see why one would one source for all the envelope timing stuff and just swap out the resistor for the RC element and maybe slam the reset on the 555.The machine now wants to gate drive the first pulse transformer stage. The green LEDs faintly agree. But I can't make much sense of the driver. The cap close. to the two TO-220, IRF640N is charged to 15V. The trabsformer puts out a signal for the base of those TO-92 half bridges. The base drive signal and only that appears on the output, the small caps feeding the half bridges are not getting charged. I suspect the voltage for the gate drive does somehow come from the board with the transistors on it. On the power board the gates are merely pulled to source via resistors. Idk, more tomorrow.
>>3002532Computer VGA is not the same RGB as TV RGB. RGBhv versus RGBs. You've need a converter even then. Besides, who would want to output VGA to a 9 inch screen when CRT computer monitors are easier to acquire than CRT TVs? 9 inch was fine for 8-bit machines but I wouldn't want to use it for something natively outputting 640x480.
>>3002534Errr maybe not tomorrow.Seems to work. Shorting the switch input gave HF start, lots of relay clicky action, gas valve and all that.Throwing it back together completely.
>>3002540>Computer VGA is not the same RGB as TV RGBthere's also the non-trivial problem of HSYNC frequencycomputers are around 31Khz but TVs are 15Khzso you get this rolling double image shituation
>>3002476>matching transformerAlso called a Balun. Do you even need one of those when you aren't using an antenna? I don't know. If you can eliminate that you might get better picture quality.
>>3002603A balun is not a matching transformer. A balun converts a balanced to an unbalanced signal or vica versa. A matchibg transformer can literally be an unun.
>>3002603Your pic and my pic make a good team.
>>3002608For TVs a balun is a matching transformer. Google it. Shit man, Wikipedia even redirects matching transformer to Balun. Am a saying all matching transformers are baluns and all baluns are matching transformers? No. In this specific case though, it is correct. Now, do you still need one to connect an old TV and an Atari 2600? Fuck if I know.>>3002611I actually had one of those growing up. Probably still on one of my TVs somewhere.
>>3002619>Now, do you still need one to connect an old TV and an Atari 2600? Fuck if I know.If the TV already has an F-type coaxial connector input then you just need to adapt RCA to F-type. If the TV only has antenna terminals, you need the matching transformer and female RCA to male F-type adapter.
>>3002621This is assuming the Atari is outputting a RF signal and hasn't been modded for composite output.
>>3002619You're not getting my point.You explained that 'a matching transformer is also called a balun'.This is incorrect. A balun can be 1:1 and in this case will not provide impedance transformation. Equally a matching transformer can be a simple unun and in this case not change symmetric to unbalanced. Baluns can be matching transformers, balun and impedance transformation. But this doesn't mean that all baluns are transformers or that all matching transformers are baluns.
>>3002619>Am a saying all matching transformers are baluns and all baluns are matching transformers? No. In this specific case though, it is correct.>>3002646>But this doesn't mean that all baluns are transformers or that all matching transformers are baluns.>You're not getting my point.And you aren't reading my post. I posed it again to give you a second chance. Don't make me have to post it a third time.
>>3002655>second chanceKEKImbecile. You're correct that I did not read it as you even started it with repeating the same statement. After you had a chance to realise your error. Meanwhile your argument boils down to 'links between wiki entries'.And now you go on about 'chances' because for some reason you keep insisting that either you didnt say or mean what you did or you were correct and matching transformers are baluns or vice versa.Retard.
>>3002657>You're correctApology accepted.>>3002476>ditch the matching transformerNah, seems you are stuck with it.>add composite inputsNot without seriously modding the thing. Which you should not attempt with a hot chassis. Maybe circle back on it if you ever get into modding and get more experience under your belt.You could try broadcasting your own TV channel. Basically, all you need is something with RF out (like a console), an RF amplifier, and an antenna. Console RF out > amp > antenna. Turn the TV to the channel the console is outputting to and it should work. This requires no modding of anything, lets you output to multiple analog TVs, and won't get your electrocuted (RF burns are a different thing).
>>3002530I’m not sure what’s going on with the mod/ext switch, looks like there might be some sort of balun already in there, but I can’t tell if it does any impedance matching. That UHF input is definitely 300 ohms balanced, but the VHF input to the demodulator chip looks to be coaxial. There’s three VHF wires going to connectors, but I only see two screw terminals for VHF in the image >>3002476? The switch seems to be designed to let the user connect a rod antenna, if it’s a quarter wave antenna it would be an unbalanced 36.5 ohms, since it’s not 75 and has no shield you’d probably want to go for a 300 ohm input instead.You’d probably want to go through that transformer either way to isolate yourself from any hot chassis, but idk about that stuff.Also one of those balun windings looks shorted.>>3002555Oh, and you can’t change that in the computer video settings? It’s still an option to use an HDMI to RGB-component or SCART converter, with the right level shifting or amplification or whatever to tap in after that big IC in the top right corner, if you wanted to maximise video quality. But with that and the high voltages, it’s an advanced project I wouldn’t recommend to anyone who hasn’t done simpler CRT mods.>>3002657>>3002674Quit polluting this space with your failure to communicate. It’s obvious that the first anon meant:“All external television baluns (the only type of balun relevant to this discussion) are 300 to 75 ohm matching transformers, and so a person looking into TV matching transformers should use the more common name of TV balun”.
>>3002683If you would like better quality images of the schematics you can get them here:https://archive.org/details/manual_KV9400_SM_SONY_EN/Due to resolution and file size restrictions on this board I had to halve the image size. Nice to see hand drawn PCB layouts every once in a while, eh?Here are some videos of the TV opened up and in action:https://youtu.be/KwrXDv9wYZ4https://youtu.be/mdBcdy3S_T4?t=245https://youtu.be/LXaYUPCCn0o?t=45In the first two videos is does appear that they are using baluns on the VHF terminals. If one is build-in, would using a second one the cause the snow problem >>3002476 is having? Shame I can't find an operators manual for it as that might cover that.
>>3002700>If one is build-in, would using a second one the cause the snow problem >>3002476 is having?I suspect not. I'd try both VHF and UHF since there's definitely a difference in circuitry between those, but those balun cores shouldn't make a significant difference to quality.>VHF: 75 ohm unbalancedOdd. If there was a shield connection to connect a coax up to the internal unbalanced VHF input I'd definitely try that, but there isn't.
>>3002700>it does appearyou mean at 8:30 in the first video? The fun bit is that I expect the device to flip it the other way round again almost first thing after the input. What's the TV gonna do with a balanced signal?So yeah at 8:30 that guy uses one almost certainly, unless hes up to really weird shit, he's using an unsymmetric like of god knows what impedance to connect to the symmetric piece of line and inputs after all. So unless weird quirky eqceptions apply here that's definately the case. Sorry, know all about transmission lines and HF but nill about TVs other than cathode ray, funny magnet coils, reset on sync and advance by 1 scanline.
are rolling op amps a meme? shit all sounds the same to me
>>3002789I have no idea what 'rolling' here means but judging by the obvious audiophile context:Yes. Most likely.No field is more of a scam or delusion than audiophile BS.If you want good audio designs stick to the basics and if the water is murky, check if it's something that is relevant to HF and radio communication.You see, if systems that are sensitive to - 130 dB, are governed by very tight channel separation rules and so on work without salting your transistors under the full moon - then it's probably not needed. Meanwhile established and well proven methods to make amplifiers linear asf exist. Idk go digital and predistort your signal or go analog and severely degrade your active element.>inb4 but digital some HF component blablaBesides humans being unable to perceive frequencies much higher than audio it's not even reasonable those signals are being reproduced, perhaps wont even ever reach a speaker because of intentional and unintentional LP filtering and the speaker - who knew - is a LP filter on its own again, just like an ear.Sorry. But audiophile ideas are exhausting.
>>3002792rolling means swappable in this context. I'm curious if switching out the op amps from a NE5532 will do anything whatsoever
I want someone to point me at a video, book, or any other source that explains how to build a mist generator with those ultrasonic speakers.I don't need a source that tells me>here are the components>solder them and violin, mistI need to learn the theory.From what I see you need to do something like impedance matching in order to drive the speaker efficienty so as not to waste energy.Is there any source on that?
>>3002789>>3002800There definitely are some differences between op-amps when they're pushed to their extremes (mainly slew-rate and hard-clipping) but generally the whole point of an op-amp is that it's acting like an ideal component. Another difference you may notice is noise floor, if you're amplifying by a lot of dB or otherwise have a very quiet signal, bad PSRR and CMRR figures can make noise audible, especially if you're trying to maintain 24-32 bits of dynamic range. Square root hertz noise may matter here too. Standard listening doesn't get anywhere near 32 bits, but some fancy microphones record at 32 bits to avoid any clipping distortion, allowing for editing of levels in post. You should primarily aim for low power rail noise and common-mode noise in the first place, but the ripple rejection of your op-amps still make for an important step in your noise budget.I usually use socketed op-amps in my circuits because I'm often pushing them to their limits, with high-ish gains and frequencies where oscillation is a possibility. Also where blowing shit up is a possibility, and I'd prefer to easily be able to swap out my silicon chips. I'm not making hifi circuits.>>3002805Ultimately they're capacitive piezo elements, so you might have better luck searching how ultrasonic cutters or cleaners or antifouling systems are driven. From what I understand you can either pick a fixed frequency, or make a circuit that finds its own resonant frequency. Some piezo elements have their own feedback pad that make this easier, though I've never seen that on a mister. You couldn't make an LC tank circuit with the piezo element as the resonance frequencies would differ, but when driven at its mechanical resonance point I imagine the power factor changes enough that you could use current feedback from a square-wave voltage driver. Something like pic related. Modelling the piezo disc in spice seems like a massive pain so I'd just start breadboarding.
>>3002811thanks, good to know
>>3002811>slew rate>clippingin audio designs>noise floorwith your ears, even after amplificationI will gladly sell you resistors that violate the sqrt(4kbTRdf) for some audio designs KEK>CMRR makes noise audibleyou mean passes existing noise through.>microphones record at a resolutionthey just convert physical phenomena, a digitizer that may be downstream would have finite resolution and so on. Audiophiles have alot in common with amateur radio.>not making HiFi circuitsthank god but still: None of this applies to human audible perception or at all to a sensible design for audio frequencies.>Piezo transducer frequency.They will faithfully reproduce any oscillation within their specified limits. Usually (ultrasonic cleaners and such) one gets to see suprisingly high voltage drivers operating at fixed frequency which very likely happens to be equal to the datasheet resonant frequency of the element.
>>3002534Sort of rigged it back up together after believing I had identified and rectified the one single issue. I have added probes to the output of the first power stage, so between the rails of the split power supply that is the output.>picrel I am probing like thisI am getting normal operation from seemingly everything. LEDs on the gate driver on one side show it's active. This means the thing is operating as a two switch forward. I am getting less than a Volt on the output. I believe if everything works I should register idle voltage, which is quite high, there.Obvious issues?
>>3002838So I checked gates on the first stage next. Seems to be okay. It looks like the operating frequency is set at 25 khz. Pulses are ~1us wide but there isn't a load and I turned the current setpoint way down.I'm getting noise on the gates of transistors that should be off at 500 khz, 10 Vpp. Idk but I'm not sure if 5V is enough to turn them on. I should try to put it in stick welding mode next that switches off the HF start which I suspect.I dont have gas connected at the time so I shouldn't be making arcs. Does anyone know if it's stupid to choose a resistor for the set amps and see how that goes? IIRC the machine will start even with a dead short.My issue continues to be the fact that I get hardly anything on the rectifier while the transformer should be running and the output is open.
>>3002846More suprises:Yes it was the HF start messing with the gate.Oddly in stick mode it uses both primaries on the transformer. This makes no sense unless it's a function promising easier stick start or something, because the second primary is only needed to provide AC on the output.I had to revise the schematic I realised the one I drew earlier made no sense at the output. Here is the new one. Negative output is on the transformer center tap. So the final stage is much like a push pull but in reverse.
>>3002851Check the smol caps near the pair of transformers in >>3002534
>>3002851This makes me think: I might have the primaries hooked up in the wrong fashion. This would be a PITA to change again. I have no definitive knowledge about the function of the gate drivers so they might not be synced or antiparallel in the way I suspect. So the fuck all I'm seeing on the rectifier might just be some residual cross conduction from one side turning on before the other turned off. But for the type of converter both switches need to be on. Begs the question how to measure that. AC coupling and the full voltage appears on the scopes input capacitance. Run it with the gates disconnected and hope the pulldown is enough to handle noise, disconnect gates and short to source, disconnect mains (PITA), disconnect mains relay on the board, add disc caps to probes...It would suck so bad if the primaries are crossed over.
>>3002855I think I buzzed everything out when reversing the gate drivers. I will do but I think it would have shown if they are shorted.
>>3002858I disconnected the relay and checked the DC link to chassic GND. The relay switches only L or N. No matter which way I turn the plug I am getting half of mains. After disconnecting it drops back to zero almost instantly, if the caps had charged it would be significantly slower. Why would that be? Shouldnt it read zero or close to zero if only N was connected and mains if L was connected?
>>3002860Gates are doing fine. It came to. me that I tested them all in DC TIG mode first, where it uses only one primary and the transistors that I expected to be a pair were being driven. They are connected by a primary. Bad news I guess. More issues.
>>3002859>shortedDid you test ESR?
>>3002865No I have no means to other than generator + scope.The driver seems to work well. More head scratching. This is not what I saw on the gates. The gates looked like this but capacitively coupled. Please not leaky gates.>PicrelWhat I got on the gate earlier. Settings were different but really looks like the same capacitively coupled.What I get on the driver.Capacitive divider to generate reference.
>>3002868picrelDouble checked. Then realised: Those are not IGBTs. I thought I had checked the datasheet. All 4 BJTs seem fine so far. BJT explains why the voltage after the base resistor is clamped to one diode.They are marked FGA60N60 but maybe they are BJTs in a fake package? With the driver disconnected they all have one diode drop between gate / base and source / emitter. Can't be that all 4 IGBTs failed in an equal manner, can it?
>>3002870Absolute shitshow.Testing across primaries.Getting about 300 Vpp 25 kHz across ome primary. Roughly where it should be figuring in the drop of the rectifier and the transistors themselves. So this means whatever is up with the gate voltage situation the transistors are at least somewhat okay. But the waveform is fucked up. Like saturation. Well it's got no load om the output. So under those conditions mhmkay. The other primary has hardly anything, could be capacitively coupled from the first. Especially since the waveform looks similar.
>>3002873I believe to have narrowed it down:Only transistors 1&4 or 2&3 can be driven at any one time. Unplug the other two - > square wave. Removal of the gate resistor is sufficient. Else the fucked up signal. This is without mains power (relay deactivated). For this T1 & T2 and T3 & T4 connect to one coil. T1 and T3 are LS and connect source to GND and ofc T2 and T4 are HS and connect drain to V+.It is not the load. Every gate looks like 1kR to the driver. Loading a single output with 220R and the driver remains happy.I know the transformer has no short on the primaries.It is almost as if this depends on the states of the transistors. I doubt that one transistor manages to pull the others gate somehow. I'd rather guess it has to do with a transistor pulling the other transistors source potential low and the driver not liking that somehow?
>>3002865In circuit all 10 uF caps read 10uF+
OP here. So, I moved the TV to show my setup where I have some better light and now it is not doing it. I put it back to its original spot and it is still not doing it. Weird. Shame about not being able to add connectors to it.
>>3002892If you plug the TV into an isolation transformer you could add composite & audio input without issue. Otherwise you have to use high speed optoisolators and 1:1 audio transformer(s).
>>3002897I'm not allowed to comment on those matters ITT so I'll just remark that.
>>3002902Picrel, not >>3002603
>>3002819>in audio designsSome amplifier topologies (especially diy ones) use the op-amp to drive a class-AB stage directly, so the slew rate matters just because you’re driving high voltage amplitudes into your speakers. Such a topology is also subject to clipping, most other audio circuits are probably designed to avoid clipping at maximum input volume, but I can see it being an issue for low voltage circuits (e.g. 5V single rail for a USB DAC).>you mean passes existing noise throughYeah, as I said it’s a component of your noise budget. Your power rails and input circuitry can only be so quiet. Usually just having linear regulators on your op-amp power rails is enough for audio-grade PSRR, CMRR might be handy if you’ve got an unshielded ungrounded system but I’ve never run into common mode audio noise myself.>they just convert physical phenomena, a digitizer that may be downstream would have finite resolution and so onDon’t see how that invalidates anything I’ve said. Even if a human eye cannot tell 16K video from 6K video, having a 16K camera allows you to digitally zoom around your image in post. Exact same reasoning behind using a 32-bit microphone. Just because a sound doesn’t fit within the comfortable range of human hearing doesn’t mean there isn’t reason to record it.As for the piezo element, driving outside of resonance can definitely produce a usable output, but you’ll be wasting a decent amount of power in your driving circuit, and have to use a pretty high voltage. Capacitive loads are a pain. Well if the Q factor of the mechanical resonance is low enough it won’t really matter how you drive it.>>3002823How is that 60Vp-p max into the drivers, when you’re outputting up to 60Vp-p before the 62.5:1 transformer?
>>3002892Some form of TV interference. Whip up a little wire antenna on the end of a length of coaxial or twisted pair line going to one of the antenna plugs and wave it around your house to see where the noise is coming from. Maybe you can kill the noise with ferrite beads on the TV or offending appliance.
>>3002935>How is that 60Vp-p max into the drivers, when you’re outputting up to 60Vp-p before the 62.5:1 transformer?It's an impedance matching transformer designed for maximum power transfer based on impedance ratio instead of stepping up/down with a turns ratio.
>>3002941That's not how impedance matching works. If it's not a 1:1 transformer, there's also a voltage change. If it is a 1:1 transformer, there's no impedance change in the first place.
>>3002946Why don't you lay out the entire theory then, Professor Reverend Dr. Magneto. I'll check in every hour to let you know how wrong you are. lmao
>>3002950Ok assume we have a sine wave coming out of a RF amplifier's 50 ohm output, with a power output of 2W. That's 10Vrms, 0.2Arms.Then we want to change that to 300 ohms, with a transformer or an LC matching network or whatever. 2W at 300 ohms is 24.5Vrms and 0.0816A. The voltage ratio is 2.45, the impedance ratio is 2.45^2 = 6.0.In either case I'm assuming that the transformer or matching network will have no resistive losses, in reality you'd aim to minimise any such losses as is reasonable. A lossy impedance matching network (a resistive divider) could also exchange impedance from 50 to 300, but transferred power would drop substantially. Both in the case of a piezo driver and a radio amplifier, such an approach would be undesirable, because as you said, we want to design such a system for "maximum power transfer".
>>3002952How does core size/material & permeability at variable audio frequency range up to 40kHz factor into power transfer?
>>3002954I haven’t looked into it too much. Lower frequencies seem to need higher inductances to carry the same power, so I’d assume there’s a rule of thumb for signal impedance with respect to winding reactance. You’d select this for the lowest audio frequency you care to accurately recreate. Common isolation line-level audio transformers (e.g. in microphones and ground loop isolators) don’t need to couple much power at all, and thats the only time I’ve dealt with audio in a transformer.So low frequencies at high powers seem to require high inductances, but they also most stretch your core towards saturation. I’ve tested a few different core materials for power electronics, and there’s definitely a trade-off between high field strength before saturating and high inductance. Air-gapped ferrite cores (like flyback transformers) have lower inductance for the same number of windings than an equivalently sized solid ferrite core, but more than make up for that with their maximum amp-turns. I found that out after doing some testing with a clip-on ferrite suppression sleeve, increasing the number of paper shims between the open end. 16 turns around the core with no shims had an inductance of 640uH and a maximum of 18amp-turns, while 6 paper shims dropped the inductance by 4x to 160uH while the amp-turns increased by almost 6x to 105. Because energy stored is a function of the amps squared, it could then store almost 8.8 times as much energy before saturating.There’s also a lot of different types of ferrites. The general purpose toroids I’ve found are high-loss at 100MHz and easy to saturate. But balun cores seem to be lower loss and higher inductances in general. And thread-on ferrite sleeves were high inductance but always high-loss, with a measurement range from 15kHz to 600kHz.Powdered iron cores are somewhat high loss, but can take a lot of amp turns.
>>3002936Yeah, this place is a mess. Built in the 1950s. One room in my house has all sorts of infrared interference so you can't use IR remotes in it*. Only the kitchen and bathroom have grounded outlets. The fucking main breaker in is a car port and doesn't have a cover. And it hums sometimes. Most of the wiring is picture related and the insulation turns to dust if you move it. The entry point for the power lines is on the literal other side of the house from it. Which is also where the fucking grounding rod is. They blew in about 4 feet of cellulose insulation in the attic at some point so I can't even tell if the wires are run up there or not. Oh, and the breaker box has not one but two breakers that I have no idea what they go to. I flipped them off and not a single outlet, light or appliance in the place stopped working. I suspect one is for an attic fan we had removed, why it would have its own circuit when it was a 3 amp motor is beyond me, but I have no idea about the other.*I even unplugged everything in the place, removed all the LED light bulbs, and closed the curtains and it is still blasted with IR.>>3002897Well, an isolation transformer is out. Pricey. Plus the 'cheap' ones seem to not actually isolate anything? Fucking China.The optoisolators looks interesting. All of this is over my head but I can solder and follow instructions without setting anything on fire too much. Do you know someplace I clean learn how to make a PCB that could handle that? Composite I guess. I could tie that into the PCB somewhere upstream of the RF modular right? >>3002530 said the jungle chip could handle S-video but I don't have much that supports it. Plus I actually have a Commodore 1702 which supports it natively.
>>3002965>Composite I guess.I was going over the schematics and, if I am reading this properly (please remember, I am an idiot), it goes from the RF chip (IC201) to the jungle chip (IC301*) on the chroma and luma pins. Does that mean I could tie into the circuit somewhere between pin 19 on the IC201 and the Video Amp Q202 with a composite video signal and the jungle chip would handle the rest? With an optoisolator so I don't kill myself.For audio, could I do the same by putting the signal in between pin 21 of the IC201 and pin 8,9 on IC202? Again, with an optoisolator,*UPC1352C https://datasheet4u.com/download/511124/UPC1352C.html
>>3002970Doesn't look like there's distinct chroma and luma at all, just a single trace carrying composite video out of IC201's pin 19.I'd consider desoldering Q202 and use its output pin to wire up your video output. And probably resolder Q202 onto your custom board. Assuming Q202 is easy to work with.I have no idea what optocouplers are available for isolating a video signal, I'd be tempted to just try some sort of a simple isolation transformer, but that's just a naive guess. No clue how DC bias or linearity compares in either case.For audio, there's already an isolation transformer T631 between the audio signal and the speaker (and the earphone jack), just use that. But it appears to be mono, so you may want to upgrade that anyhow, or just not use its internal audio capability in the first place.
>>3002873So I spent the rest of the time reflecting on why the gate driver can't switch on two switches on the same coil. To no avail.>PicrelIf I get to assume a mix of AC and DC conditions and ideal transistors I arrive at this. L and C could be shorts and switches closed. I believe it would be the same if either L or C wasn't shorted.The drivers are isolated and do not care.I have also found out why my diode test reads 0.5V from gate to source on every transistor. It uses 500 uA and drops 0.5V on the 1kR pulldown, tested with anothwr 1kR. Tested the zener diodes individually and they're good.With IGBTs isolating the gate signal I can only see three possibilities:- I's got to do with the source connections... but drivers are isolated.- It causes shorts that nuke the whole supply and trip up the gate driver... but everything else ses to keep running and fuses stay put. - The controller simply ramps up to 0.5 duty cycle when the mechanism is defated and very little if it's haply with what it's getting. It wouldn't explain why the same is observed without the relay for mains. Also yesterday the scope showed the driver totally crap out intermittendly. Could be the controllers work but not likely. I have a few more ideas what to do:Bridge drain source on one device and see if it does the same - aka if it's the device state or the driver connection.Hook one gate up to battery power and see if it behaves the same, same idea basically but also to check how much current it takes to keep the gate high.Bridge two transistors and buzz out shorts.Find a point to test upstream and see what the controller is doing and whether it's changing it's behaviour.
>>3002982sorry for all the obvious mania,it is working and has been since the fix on the one dead board. I am getting 150V DC on the DC link when idling. It must have been a shoddy solder joint when I added hookup wire for probing.This also implies the ugly gate signal >>3002870 is ugly but working. The appearance is probably due to the superposition of a very brief gate drive pulse (doesn't even reach full voltage), presumably the pulse from the other side coupling in and lastly some of the tansformer waveform coupling in.It's just not all bice and super clean but I have a good bit respect for chang working under conditions where his boss doesn't let him use any ICs but 'only the logic gates, OPs and passives we have in the bins'. The whole thing appears to be a very good exercise in economics and punishing lower caste guys like those in manufacturing and assembly. It's like someone purposefully went against DFAsm and DFManf.I will investigate a few kore thingsmI want ro completely reverse the gate drivers and a few more things, there is alot of cool ideas and circuits in there that could have only come from looking at others designs alot or books.
>>3002989REEEEE
>>3002990It's definately working. Counting it as a win! I finally have the confidence in it working that I needed to put it back together completely. And arc flash already!Because I'm lazy I rigged 2S5P 3.3R 10W so that gives 6.6/5 or 1.32R. The lowest programmable Amps are 5 according to manual, 7 according to what the machine reads out and god knows how much in reality. Didn't want to risk my 10A unfused meter. So that's 65W if 7A is right and I'm only doing short cycles.>TestIn stick mode at the output under no load It read 75V. Across the resistors it dropped 8.5-9V this gives the 7A as the display reads out. Good.Massive arcs when disconnecting. I think the controller does something to maintain arcs even for people too dumb to stick weld. Still seeing the arc.
>>3003002I have fucked up the polarity.>take it to the shop>hook everything up>try TIG>tungsten almost burns up and ruins a glas lens>feels like wrong polarity>check in the pictures I took>it's clear. Connections on the output driver board are reversed>this should have become clear when my meter explicitly said it's negative polarityTFW
>>3003023THERE WE GO!finallyEverything works again. IDK why I need 4 TIGs but meh.
>>3002970Build a selector switch with a THS7314D, ADC, and fiber transmitter inside that connects to the TV over single-mode fiber optic cable. Then put a DAC and fiber receiver in the TV and inject that into the jungle chip. lmao
>>3003042>THS7314DTHS7313 adds i2c control and 2 selectable inputs.
>>3003030Congrats. I’d like a welder myself, what’s the chance I can just buy a broken one and fix it? What kind of welders are easy to fix?
>>3003074>What kind of welders are easy to fix?Arc welders.
>>3003074The way I see it chances are very high. The impression this one left on me was: Skookum as frigg.Electrically it's a very rovust design and not much can go wrong with it. This is backed up by me running it under all sorts of invalid states (defective output driver, random gate drivers disconnected, output driver crossed over etc.). It is also corroborated by the fact that the only thing that was wrong with it was one electrolytic capacitor. Not a design or manufacturing error on the welder companies side, if anything a procurement issue but in any case an issue that befalls almost all older electronics. The expensive bits are those that switch the power, contain lots of copper or ferrite and those appear indestructible. Worst you are replacing diodes or transistors at 5 bucks a pop.This exact model was a white label product. I suspect any chinese welder from the era that has the same capability is the same design.Other welders I've been into seemed equally robust and easy to maintain or repair. Just keep in mind the smaller the thing gets for the same power the more voodo and MCU you can expect to find.I'd shoot my shot.
>>3003080>it was one electrolytic capacitor.Which one?
>>3003081>Which onesee thread.>probably C69 or so, hope that helps>picrelI wanted to explode it proper and it wouldn't even gibe me the satisfaction. Just a little pop.
>>3003079Isn’t every electric welding machine an arc welder, bar spot-welders? SMAW, GMAW, etc.I’d prefer being able to do TIG and MIG, and AC for aluminium, and don’t really have room to stow 3 different welders. The question was more about what topology (e.g. inverter, number/type of transistors), brand, and year I should be on the lookout for.>>3003080What year is your welder from, early 2000s? I do find that level of discrete logic very appealing from the perspective of reverse engineering and troubleshooting.
>>3003096I think I said it last thread but date code is also all over the pictures ITT it's 2011.I do not thibk you'll find a machine that TIGs and MIGs. Back in the day some machines were modular, perhaps that's possible. Like you had the main transformer / inverter and a seperate unit that does the feeding.
I've got a few of these I'd like to have around the house but I don't want to buy the power supply at $50 each. I'm a beginner in electronics. I have a bunch of those 12v power supplies that I think could work to power the LED panels.Is it as simple as connecting the wires and taping it off or is there another safer method?I have one 12v power supply that reads Output:24v = 0.75a. Does that mean it should only be used to power 2 of the LED panels?Here's the product I'm trying to save money on, might provide more information: https://www.bunnings.com.au/kaboodle-12v-15w-led-driver-and-distributor_p0176166
>>3003137Ok makes sense. I'm guessing there are existing products to add MIG feed that tap off the same pedal signal, if not it would be pretty feasible to DIY that sort of thing. I could feasibly handle the gas turning on and off externally to the welder also, allowing me to use a machine designed for stick or other fluxed welding, but getting a machine designed with TIG in mind seems like the way to go.>>3003138>I have one 12v power supply that reads Output:24v = 0.75aThat's not a 12V power supply, it's a 24V power supply, might want to check your brain. If the LED panels are sufficiently well matched or stable in series (they probably are), yes you could run two 12V panels in series off a 24V supply. But in this case, 0.75A @ 12V = 9W, and those LED panels want 9.2W. It might not be a problem, chances are those PSUs are just going to run hot, but that's not good for their longevity. It could possibly go into some sort of protection mode, turning on and off as the current or temperature gets too high. I'd want to use a dedicated LED power supply that goes into constant-current regulation if I was running it that close to its rated power output.Each LED panel theoretically draws 0.76666... amps, so you want to add up panels such that they're using most of each power supply's rating. Say if you had a 12V 2.5A supply, you could run three panels in parallel at 2.3A total, that's a safe margin of error. If you bought their 30W driver you'd be able to run three LED panels off it, instead of their 15W driver that you listed that would only allow for one.
>>3003138>Here's the product I'm trying to save money onyou can def get adequate supplies much cheaper current prices i'm seeing at thrift stores is around 2 canadian pesos
>>3003142>handling gas externallyFor TIG that's a common thing, valve on the torch. Works well but can make the torch bulky.For MIG? Why? For MIG I want the gas to come on when I press the button and bobs your uncle.>PedalAgain for MIG? No thanks, the things I work on and places I work in with a MIG are definately no pedal places and I dont see why I'd want a pedal for MIG. Hit the button and lay down weld.Besides that:Any recent TIG machine unless painfully cheap will have a gas valve, so provided you do something to the logic, or MIG in TIG mode and lockout the HF start or just hook the gas valve up directly to the button input, you already have gas control so I'd imagine you'd want to feed it through the feeder unit as you want one place where you hook your cable up and not sort of try to tee off to a bottle or something.>flux corenever done that probably ass just stick itWhat I really like about the box, now that I have seen it and understand the grand picture of operation: The thing is a 200A programmable current source. Probably wont go very high voltage at 200A but apparently goes close to 70 at low settings. There's a whole lot of things that thing can be useful for.I almost regret it wasn't shot and I would have had a reason to get the goodies desu. >>3003138Since it says "12V" and nothing else I wonder if it is somehow regulated. At 9W I could even see how they just put series resistors in that burn a little. If it had no current limiting I would expect it to explicitly state the current. But the math says you need 770 mA.What I would do is see if you can get something that is limited, or put something inline to limit the current to double that and run it very briefly to judge if it's obscenely bright, implying it has no current limiting. Even better if you make it adjustable.If it isn't very bright at 1A and above it's safe to assume it's somehow limiting current and you can hook any regulated 12V supply to it. No need to shell out 50 bucks.
>>3003202cont.If on the other hand you discover it is not current limited, unlike the label implies, then you have many options, DIY or buy, almost all of which come wa, below 50 bucks and the cost or effort really just boils down to efficiency and neatness. If ydgaf you can pull a 12V transformer and rectifier out of something and make a transistor current regulator, 2 resistors, a zener and a transistor get you there, maybe less but it's almost the same as just putting resistors in series.
>>3003202>For MIG I want the gas to come on when I press the button and bobs your uncleWell yeah, if I’m using a welder that’s just a power supply it won’t have a gas outlet or a wire feed, so both would need to be controlled externally. Whether by pedal or button for the TIG, or trigger for the MIG, I’d want it to be automatic. A valve on the handle for TIG seems like you’d waste more gas compared to automatic gas feed.A quick look locally and there’s mainly just some ancient stick machines, and some functional $150+ MIG machines. I should look for industrial auctions.
can i put a 12v dc motor speed controller inside a small airtight sealed box? it powers a 20w pump its for outside and won't run more than 12 hours max at a timewondering about the heat
>>3003218That should work. And you don't need to go deeper than trying it out. A larger box will have more surface area which helps.In the end it depends on controller efficiency (or losses), which I assume to be high and the ability of your box to get rid of the heat. 12 hours or indefinately should not make a difference and you should design for continuous operation anyways. For thermal models you can use common electronic concepts, it's all RC networks and a source, everything has a resistance (insulation) or conductivity for heat and a thermal capacity. The later can be ignored for a steady state case. Isothermal points (ambient) can be thought of as a potential source.Plastics are generally on the order of 0.1 - 0.5 W/m•K.Idk lets say your box is 2" and all around so you get 6x0.025^2 = 0.00375m^2 and for simplicity lets just ignore the air inside and the thermal junction between controller and air and air and box. Box is 1mm thick. That then means we get 0.1 - 0.5 W * 1000 / m^2 * K and taking into account the box geometry and again ignoring convection and radiation at the outside the box gets rid of 500 x 0.00375 W / K.Electronics are usually happy to at least 60 degrees C or so, ambient 20 C gives us 40 C of headroom at which point the box gets rid of 40 * 500 * 0.00375 W which is 75W!That's alot, alot more than I would have guessed. Someone double check please. Maybe I moved a decimal.But you get the idea.>Coaxial vs twisted pairWhen doing up the welder I was sort of delightes to see all those twisted pair wires. Amongst other things it just screams 'getting things done'. They just do anything they can get away with if it works and saves cost. This makes me wonder: How much worse is twisted pair (for asymmetric signals) generally when compared to coaxial? And I'm not talking differential. They literally solder wires to gate and source on one board, give them a twist and solder it to the GDT on another board. Is that even really asym?
>>3003226I found my mistake. I have dropped the range for plastics thermal conductivity half way, so you get a range of 15 - 75W which still seems high IMO but there's still the convection and conduction inside the box and outside the box which will lower the result. Still seems safe to me. Your speed controller must have terrible inefficiency to fuck this up. If you ca shape the box (say 3D printing) give it more surface area like hollow fins.
>>3003218If heat is a problem, use an aluminium box and thermally bond the speed controller’s transistors to the inside of the box. Electrical isolation is a good idea. Add a thermal fuse or thermal breaker if you’re worried.I doubt you’ll need fins or forced airflow for a speed controller (probably less than 25W to dissipate), if you need more dissipation you can thermally bond a finned heat-sink onto the outside of that waterproof aluminium box. A compromise is to bond the heat-sink directly to the speed controller, then cut a heat-sink-sized hole in a plastic enclosure and fill the gaps with epoxy or maybe a polyurethane glue. Ideally there’s a flange of the heat-sink sitting against the plastic, such that it isn’t just a glued butt-joint. Easier to get a good fit around that heat-sink with a 3D printed case.Cheapest option might be to use thermally bonded sheet aluminium as the lid to a plastic (printed) box, with a gasket squished in between.
>>3003226>>3003246thanks bros very cool info gives me a few paths to make it work without worries
>>3003042>>3003044Thanks for the info but I don't know enough about this stuff to build my own circuit. I found this one online that is video only. Would this work?
>>3003288That’s basically the recommended schematic from the datasheet of that opto. As for the circuit layout, it’s 17MHz bandwidth so you don’t have to worry about transmission line effects, I’d just ensure you have a solid ground plane and return paths for all signals. Personally I’d add a ground plane to the other side of that YouTube layout too, with stitching vias around crossing traces at minimum. I also prefer the TO92-wide footprint and horizontal resistors.Maybe there’s more details for recommended PCB layout in an appnote, or in an existing product.
>>3003288>>3003296
>>3003296I got it from here:https://antiqueradios.com/forums/viewtopic.php?t=430210There are a couple of PDFs and gerber files. I was just gonna make something using a prototyping board.
>>3003341I wouldn’t use prototyping board to make a high frequency analog video circuit. Let alone a circuit needing high voltage isolation. I’d sooner use manhattan style.Though gluing a piece of strip-board (especially the breadboard patterned stuff) atop a piece of copper-clad FR4 as a ground plane is pretty convenient. You just have to solder components flat atop the stripboard instead of through the holes. In this case, to break up the two ground planes, you’d want to cut or etch a gullet between the two sides of the base copper-clad, and have seperate stripboard pieces on either side.
>>3003368There’s the lazy manhattan method I used.
>>3003372What would happen if you used a piece of copper foil tape on the bottom of the stripboard instead of copper clad?
>>3003376So long as you can solder all your ground wires to it and it isn’t making contact with the pads in the protoboard (don’t use the two-sided protoboard) it should be fine. I’d want to use really wide copper tape, to give good shielding to all the stuff above it.
>>3003368>Manhattan styleWow, I'd never even heard of doing it this way. Shows how much of an amateur I am. I've soldered a bunch and am quite decent at it. Replacing all the RAM and a few of the larger chips with sockets on a C64 really dialed that skill in. It has all been monkey see, monkey do type stuff though. Putting together kits, replacing components on an existing board, copying existing designs that kind of stuff. How about a little prototype board for the IC, some standoffs (I have some boards with larger holes in the corners for these), and then the FR4 under that? Would I need some layer of shielding between them?
>>3003391Oh, scratch that. You can put chips on these like a dead bug prototype. I can work with that.
>>3003391Protoboards have trash grounding and are hard to desolder/rework if you make a mistake. Deadbug, Manhattan, and point-to-point are definitely the way to go
>>3003395>Protoboards have trash grounding and are hard to desolder/rework if you make a mistake.I've never had a problem with desoldering them. I've even reused them before. I didn't know about the ground issues they have though. Having also had to fix grounding problems on a Vectrex, which are notorious for having a constant humming sound because of a ground loop and poorly insulated wires, I can see the concern.
>>3003391The problem with that route is you’d have a much larger gap between the ground plane and your signal and power traces, than between those traces. Capacitive coupling and crosstalk would be hard to mitigate. At least with a single-sided protoboard flush against the ground plane those distances would be closer. Honestly single sided protoboard is just better than the double-sided stuff. Easier to solder and desolder, with a wider variety of layouts. But it is easier to delaminate pads since it’s almost all FR2.>>3003392Yeah that’s probably the most robust method. I prefer sticking little pads of copper-clad atop the ground plane as opposed to cutting isolation slots around pads, but either works. But I would carve or etch an isolation slot at least 5mm wide. Less would be acceptable if there were lacquer or solder-mask atop it.
I'm looking for a PCI-e 1x connector, but all the cheap ones I see are THT. Isn't that really bad for high speed? Why would they even make THT versions of high speed connectors if that was a problem?
>>3003426as with all things RF, It Depends. you can design the connector and footprint to have the desired impedance at the operating frequency/ies
I was told that old coax cable is a cheap solution when you need shielded wires. Cut the connectors off, solder the shield to ground, and use the core as your wire. I can get pretty long lengths of the stuff for a buck at the local thrift stores.
>>3003559>shielded wire is good if you need shielded wiremakes u think
>>3003559>old coax cable great quality but there are drawbacks which make it nasty to work with- quite large, so ugly and hard to hide- stiff, difficult to bend around sharp corners- mono only; need 2 runs of wire for stereo, more for video- shield needs big-watt iron- stiff center wire is sharp when cut, will poke holes in your fingers
>>3003563
Here I am stressing about what to do with my layout that requires the + and - pins of my ethernet traces to the swapped in order to keep everything on one layer, then I learn that in ethernet polarity literally doesn't matter and I can just wire + to - and vice versa and it'll be fine
FUCK SOLDER BRIDGES (OR SOLDER NIGGERS)I HATE THEM WHEN THEY ARE UNDETECTABLE UNDER CONTINUITY MODEI wasted 2 hours debugging I2C problem on my level shifter
How the fuck do I route this USB-C connector? The data pin pairs are blocked by each other so they cannot neatly connect. Seems like I have to go all around the back of the device to connect them, doesn't that completely fuck the impedance and length matching requirements?
>>3003644Something like this maybe? This is fucking stupid
>>3003644Things like (fully functional) USB-C connectors are meant more for ≥ 8-layer boards.
>>3003661The number of layers wouldn't even matter, they are THT holes, they block all the way through
>>3003559Mr carlson was wrong about one thing… something like RG6 with steel center conductors (CCS) is not good at low frequencies due to attenuation and it’s a bitch to work with.Whether you’re using RG6 or the thinner RG59, try and get copper center conductors.The reason they can get away with CCS on tv cable and satellite cable is due to the skin effect at very high frequencies—the signal stays on the outside of the cable where the thin copper layer is.Generally, with modern drivers, we can get away with over 3 GHz on this cable tv coax now so there’s ways to use it for high-speed digital networking using the wire that was built into your home back in the 80s.There’s other, better coax you can get, like sirius/XM radio antenna wire that has excellent characteristics and you can find good audio coax with broken microphones—the low voltage signalling needed good shielding.
A Weller 3439-00-460-7198 is going up for auction in my area. I researched it online but couldn't find a ton. I think it is a 40 watt Weller EC1201A soldering pencil, which takes ET tips. Spare parts do appear to still be available. Assuming I can get it for super cheap (<$20) would this still be a decent pack away iron as a spare/travel? My current iron is an 10+ year old Hakko FX-888D and I think it works great. Is this Weller gonna be a serious downgrade?One thing I like about it is it has a ton of tips and desoldering attachments.>>3003665Who is Mr. Carlson?
>>3003644Man, just use an SMD socket instead. Pretty sure I’ve seen vertical ones, with the pads spread out in both directions.>>3003668Fat handle, ergonomics might be kinda strange. But I see no venting on the shank of the tip, so I assume it’s temperature regulated? A quick search suggests it’s curie-point controlled, so fixed temperature. Should be damn reliable, but those tips may not be designed for lead-free use. Naturally it will be less capable than a cartridge-tip iron or an iron wth greater power. 40W seems kinda low for that big DIP reflow tip.
>>3003683I was just going based on what I was able to find online, which is just about fuck all. Weller's naming conventions, constant model changes, and other bullshit has just about done my head in. The box says WTCPK and 24 volt, 48 watts (my bad, not 40). Outside of a single line in a very old Weller catalog and a bunch of old military TMs I can't find much else about it. The EC1201A comes from a video on YouTube were a guy replaced some parts on one and had links on where to get the parts. The tips in the box say PT but what is the difference with ET? Fuck if I know. The box says the tips are for 600 or 700 °F so I assume it can get at least into that area.
Anybody have this screen and was able to free the CS pin from the driver? The header does not come with a chip select meaning you cannot have this screen and other SPI items on the same line the way it stands. Pic-related is the pinout diagram I made. The 3v3? Label turned out not to be 3v3, U2 is the backlight power module. Screen (from Amazon):DollaTek 2.4-inch TFT SPI LCD module, 240×320, using the ILI9341.My fallback plan is to permanently use a Blue Pill to run it then send the Blue Pill whatever I want to show on the screen, via SPI for speed.
>>3003779Just put a transistor or logic gate on the CLK pin. Maybe MISO too, if you need to connect that in the first place.You could also use other pins on your MCU for another SPI peripheral, be that alternative hardware SPI pins, or software bit-banged SPI. If you planned on using another SPI line to talk between this MCU and something else, they you’d already need two SPI channels. Or are you otherwise pressed for GPIOs?
>>3003668> mr carlsonWe’re talking about https://www.youtube.com/watch?v=s5kU5SeaSQs
>replaced a cap to fix Ni-MH charger>2 dead batterieshmm that's weird>open it up again>see thisi had one fucking job, man...
>>3003946Reversed polarity on the center capacitor?
>>3003948my trusty 10y ENELOOP's, to cinders...
>>3003949I bough a nihm charger last week and 8 AAA nihms. Just got the step up converters i ordered today. Essentially I'll be running 3v3 MCUs with them. I also got battery holders. Gonna be holder (2 AAA) to stp up converter to MCU, with maybe a cap before the MCU but I don't have caps an resistors at the moment because a move. Got like 5 resistors at 1.5k ohm and a potentiometer. Maybe 2 caps at 10uF or something like that.
>>3003946whats wrong?
>>3003946>>3003949>>3003998Eneloops are the only normal batteries I will buy. It is a shame so many devices simply don't fucking work with them because they are 1.2 volts. Can't put them in my thermostat, a few remote controls, or my voltage detector pen.>>3004016He put a cap in backwards.
>>3004021The other advantage is that nimh or nicd seldom (almost never) leak. I’ve had 2 over the last 20 years or so that leaked a bit.On the other hand nearly *every* alkaline I’ve ever left too long in a device has leaked, and a lot of them required fixing the PCB because that salt creeps all over the place… it’s the perfect plausably deniable planned obsolescence scheme.
>>3003998Step up converters? I’d rather use step-down converters from 3 cells, a healthy NiMH should remain above 1.1V per cell for like 90% of its capacity. That gives you the option of using an LDO for low quiescent current, or a switcher for higher efficiency. Though I bet most 3.3V MCUs can run on 3.8V or so directly. Might need to bin them for manufacturing variance though.>>3004021I use alkalines in most places because they last years at a time. If it’s 20c per battery, and it uses 5 batteries in 10 years, that’s $1, cheaper than eneloops or even generic rechargeables. It’s only higher drain stuff that it’s worth using rechargeables for, especially things that last less than 6 months at a time. Don’t think I’ve got any of those at the moment. All my flashlights and handheld radios have lithium ion batteries inside.
>>3002851So the whole welder arc got me thinking.It's not only a very potent programmable current source.It's a nice AC source too.Idk how much a 3 kw isolation transformer goes for but it seems this topology wcould make a really nice one cheaply. SMPS are mostly about trading costly transformers for cheap silicon.So I wonder if anyone can, besides SMPS noise (but you get noise from line transformers too), argue against the idea. Why would one want a large 50 hz trabsformer for isolation over this thing that uses a smaller transformer thanks to the ooerating frequency? All this needs is changing the secondary turns and some output filtering and it would do for many things. SMPS development for example. Every time I work on SMPS I have to do all sorts of funny business to get somethibg like a test run on isolation and have to test it on the line because I don't have a large battery and inverter.
>>3004036> much a 3 kw isolation transformer goesParallel a few rewound microwave transformers.Cheap.
>>3004049MOTs suck ass desu but youre sort of right. Can even bash the magnetic shunt out.Then again it's long ago that I was a kid picking through the trash to find fun toys. And new microwaves have shitty SMPS for the same reasons really.I guess there is still weight savings to be had.
>>3004030> on 3.8VJust use lion then i would think.Those vape things use rechargeable li-ion as disposable.
>>3004052Well yeah Li-ion is better, so why use NiMH in the first place? The only benefit is not worrying about battery management, you can drain those things to however low you want and they’ll charge up fine afterwards.
>>3004053>why use NiMH in the first place?fewer fires in your garage
>>3004053> drainNiCd is the champion of drain recovery. Also recharge cycles.They’re shipped drained sometimes.Getting harder to find though, unfortunately. Still used with older industrial equipment.
>>3004068Never seen an RC NiMH fire? Those RC battery fireproof bags have existed since before lipos.>>3004091>recharge cyclesIf I cared about them, I’d just charge my li-ions to 4.0V instead of 4.2V. Or swap to LiFePO4. Still more energy dense either way. Even the lack of a requirement for undervoltage protection is trivialised by the cheap and available DW01 protection boards. Perhaps the resultant quiescent current is a problem in really long standby time applications? The ability to charge at low temperatures is also a reason to consider NiMH I think.I’d like to see more use of small LiFePO4 cells for electrical devices.
Now that you mention it maybe I should read battery voltage and mosfet open the supply line if it gets to say 2V coming out of the in series nihms. I don't have that fully thought out yet. The batteries should give me 6 to 8 hours runtime with a current draw of about 100 mA (wifi stuff with esp32). Maybe I'll at least install a piezzo buzzer which I have and ring an alarm when the adc reads less than 2 volts before the boost converter.
>BatteriesSo I asked this before but never got a definitive answer. When charging Li-Ion the usual CC-CV is advised. I suspect somethibg happens to the chemistry and it charges like a 'good old battery' first and through a very different effect later, where it continues to stash charge away at a decreasing rate without the voltage changing.This begs the question:Can a Li-Ion be safely used as a buffer if - the current into it is clamped to the recommended current for the CC phase- the voltage across it is clamped to the recommended termination voltage for the CC phase.I know this would forego a great deal of the capacity. But still. I think it would work, with the voltage clamped to 3.3V per cell and the current limited to say 500 mA I don't see how the cells would ever leave CC phase where 500 mA is permissible. I have seen a product that employed supercaps about the size of 18650s in a low current application and when I did the math it showed that even if not used to thwir full potential like above 18650s would still store far more energy in the application and leak less.
Is this just oxidation on the cap or is it a sign that its leaking? caps are about 50 years old
>>3004137That's corrosion. Could be from leaking electrolyte. It's old and cheap so just replace it.
>>3004123I’d say broader. The current should be at or below the recommended CC charge current, maybe higher peaks is fine if the rms current is low enough but I wouldn’t chance it. And the constant voltage could be anywhere within the cell’s usable range, though personally I’d stick to 3.4V to 4.0V per cell, holding a lithium ion at 4.2V long term is a recipe for reduced capacity. Going too low isn’t as bad, but you’d have bigger-all energy to work with. This makes it pretty feasible to use 3S as a 12.0V buffer, but the tricky part is cell balancing. If your cells never get to a “full charge”, most balancing circuits will never attempt to balance them. So you might need a custom balancing circuit.For a 3.3V buffer I’d use a LiFePO4 cell instead, since 3.3V is above its plateau voltage.>>3004137No other corrosion nearby? It’s a leaker.
I will be using this stamp-style SoM board on my own host PCB. Should I mill out the part of my PCB below the SoM, or is it safe to leave it intact and route my signals underneath it? The bottom of the SoM looks like this so there are no components in the way, and there is soldermask across the whole thing. However I'm worried about the soldermask rubbing off and causing shorts, and also the bottom might not be entirely even which could interfere with the assembly soldering.
>>3004175>Cell balancing.You slap parallel resistors. Said that before on /ohm/ and no matter hoe much screeching I'll say it again.>12V buffer.Yeah that is wgat I'm specifically interested in. So you're saying:3s 12V buffer is okay if the source is 12V and never more and the net current (from source minus to load) remains below what those cells would be charged at? And it has UVP of course.
>>3004268>worried about the soldermask rubbing offstick a bit of acetate sheet on the bottomif you dont have any acetate sheets (a real diy'er would) you can cut some off from a lot of packaging materials
Here's my pushbutton on/off circuit with smooth on/off fade. Do you think this would even work? I guess I have to test it in parts.
>>3004274If you don’t mind the higher quiescent current and have some way to shut them off when you lose power it’s probably feasible. It would be better if you put 3.7V zener diodes in series with each resistor, so they conduct much more at higher voltages.But I would want to use a constant current 12V source or otherwise have a way of limiting current, a normal 12V supply will try to push a bunch of amps into an 11V battery buffer.
>>3004303While C10 is ramping down again, Q2 will already be off. Unless you hold down the button long enough to see the ramp-down.
The wire of shame.Thanks to the anon who recommended fine gauge kynar wire though, it seems to work well for these purposes.You can see what happened, I accidentally deleted the GND in the schematic so the layout passed all the DRC checks despite it being FUCKED UP. I'm still mad as hell but it's not worth ordering a whole new set of boards just to fix this.
>>3004307Yeah I did mention that as a precondition. The source I'm working with is ideally suited IMO, it's not very stiff and even if one did nothing about it couldn't OC the batteries.>>3004351de nadalike the silver coating?
>>3004351That’s why you run ERC for your schematic. I never do, but maybe it’s something I should learn to use.
>>3004367I did see the ERC option, but there's a lot of noise generated by things like unspecified pin types for ICs which I would presumably have to set up or selectively ignore. You are right, that would have caught it, though.>>3004361>like the silver coating?it's good although I'm not sure of the best way to strip insulation off such tiny wire, the iron doesn't cause it to shrink back
>>3004469>doesn't shrink backthat's weird! Mine always does but maybe the rabbi didn't get to touch it.I find the effect is even more pronounced on PVC insulated.Anyways, what I use is the stripping tool in a wire wrap pen and I think it should be pretty simple to imitate it's working principle.>picrelI'ts just a piece of spring steel with a slot that narrows down to but never ubdercuts the conductor diameter. Pass it through, pull down and out. But to solder it I just clip and solder it. It's not like those wire gauges are for high current or anything.
>>3004351Just like I don’t trust V0.0 software, I don’t any board without some manual jumpers on there.
Does an xray vacuum tube need ANY metallic component? Xray tubes with only one electrode have been demonstrated. AC does a fine job of slamming electrons into the single electrode. So what about no electrode but 'some source of electrons' in it, perhaps just enough gas that it can be ionized in a field and the same alternating field could then slam electrons into the tube envelope?
Anybody have a study guide for the ISCET industrial CET journeyman? Or insight on what’s on it/how much of a pain it is?
>>3004317Thanks, I think I see what you mean. C8 might need to be quite a lot larger. But if I use a big electrolytic cap there, it will have lots of leakage current and quickly drain the battery. I'm this close to just installing one of these on the plus lead.
>>3004521What’s the quiescent current like with that boost converter with its ADJ pin pulled to GND? What about with its SHDN pin pulled low?
>>3004521>>3004564C8 being really large could be a solution, electrolytic caps can have very low leakage currents. Especially so if they’re high voltage rated or low-leakage caps. You’d just need to ensure R10 and R13 are sized appropriately.Bumping R10 up to 10M is an option, that pushes the time constant from 10s to 100s even just with a 10uF cap. So you’re probably actually fine as-is. The C8/R13 time constant should be much quicker than the R16/C10 time constant added to the debouncing time constant, in case you do increase C8.
>>3002473How difficult would it be to DIY a smaller version of something like this, and what components would one need?https://portable-electric.com/voltstack-power-stations/voltstack-30k-level-3-mobile-e-charger/15 kWh 0.5C batteries sell for at least $100/kWh in my country. buying from china would cost ~$80/kWh. assuming a target charge of 30kW, you'd need $6k for the batteries (60 kWh) + some kind of battery/charge management system + cabling + encasing + some kind of cooling solution, right?
>>3004705>DIY something like thisPretty easy from a power standpoint. It's just power electronics, the most well understood, documented and easiest branch of electronics really. Build an inverter that suits your needs.If there is charging protocols you'll probably looking at a bit of coding or begging for closed propertiary stuff and might want to have feed forward, depending on your inverter design but I don't believe you have a reason to work with isolation. You're starting at the DC link.
>>3004711forgot:>"e-generator"-. -
>>3004711>Build an inverter that suits your needs.if you only care about DC, you'd need some kind of converter, right? in this case 4 x ~50V batteries would give you 200V, so you'd need something like an "up converter", right?I don't know shit about EVs, but AFAIK there are like 4-6 different types of DC plugs. I'd assume there are at least 3 main protocols for that.I just checked. for some reason, these cheap batteries can't be used in series, only in parallel, and I'd be scared of having a bunch of batteries pump 300A. I'd have to buy "high voltage" stacked batteries instead, which are a bit more expensive.Tbh I guess I could buy the batteries, but I have no clue how to do the rest. are there BMSs that can charge EVs directly? if not, how much do you think someone would charge for a job like this?I wish I had learned electronics decades ago...
>>3004717Don’t buy batteries with BMSs already in them. Those BMSs will have a maximum series stand-off voltage. Instead you want to be buying prismatic cells and an appropriate BMS. Should be significantly cheaper than buying pre-made batteries too.EV charging is converging on just one or two plugs and one or two protocols, I don’t think they’re that complicated.But electrically, you’re going to have to use a high-power DC-DC converter, one that provides the constant current charging curve for the EV. Because EVs only contain low-power AC charging circuitry; EVs offload fast DC charging power electronics to the charger. Best make that DC DC converter a bidirectional one so it can also be used to charge itself.Then you’ve got to sort out the liquid cooled charge cable.
>>3004717>you'd need some kind of converterYou showed us a specific product with specific capabilities.>if only DCif the only thing you cared about was DC being available at an output you would simply connect batteries straight through to an output and build an enclosure around it.If you wanted a current limit, constant voltage or selectable parameters then you'll need at least a kind of regulator and only 'lossless' switching regulators are reasonable here and will have the benefit of allowing you to get higher voltage from a lower voltage battery pack.4 x 50V in Series will give you 200V, depending on the state of charge. If you want higher voltages you'll need a converter topology that enables this.>cheap batteries can't be used in seriesI doubt it. I can not see why.>300A from batteriesWhy should it scare anyone if the system is layed out correctly. If done right every cell will be used within its specification. Nowadays you can get 300A from suprisingly little battery.>BMSBMS keep battery packs healthy. They don't really charge batteries. But it's probable that someone sells a module that has a DC input and generic EV charging output and it is likely that employing the module is preferable to own development.>how muchYou mean development of the complete system? Depends on the quality of the engineering. An indian on fiver will give you a complete set of production files and the thing will probably blow up immediately or be highly inefficient. >I have no clueI dont want to insult you but it became obvious from your second post.
>>3004729>Don’t buy batteries with BMSs already in them. Those BMSs will have a maximum series stand-off voltage. Instead you want to be buying prismatic cells and an appropriate BMSI see. I had forgot this little detail.>EV charging is converging on just one or two plugs and one or two protocols, I don’t think they’re that complicated.that's great to know>you’re going to have to use a high-power DC-DC converter, one that provides the constant current charging curve for the EV>Then you’ve got to sort out the liquid cooled charge cable.right. that sounds complicated.>>3004732>If you wanted a current limit, constant voltage or selectable parameters then you'll need at least a kind of regulator and only 'lossless' switching regulators are reasonable here and will have the benefit of allowing you to get higher voltage from a lower voltage battery pack.interesting. will research this. thanks>I doubt it. I can not see why.the other anon understood correctly: I was talking about commercial battery packs that come with BMSs, not just batteries.>I dont want to insult you but it became obvious from your second post.not an insult, I'm ignorant about the topic, though I have some basic knowledge and notions. I was hoping I could buy premade parts maybe. for whatever reason, there doesn't seem to be a big market for shit like this, even though it seems rather obvious for certain uses and contexts.funnily enough, some big companies like VW had robots like this half a decade ago but apparently the project died? https://www.youtube.com/watch?v=dftBK7ck650
>>2999611so i just pulled my diff eq book off the shelf and read through itit has a chapter on laplace xforms, unfortunately all equations discussed were real, and it didn't really address zeros and poles at allit's easy to see that zeros and poles completely define a rational polynomial function, but i still don't have a clear idea of how engineers think about this stuffanyway, at least i can jump back into my analog electronics study with a good idea of how to use laplace xforms to solve diff eqs and when they are useful(btw, they are most useful to turn linear differential equations into algebraic equations when the non-homogenous term contain discontinuities)
>>3004743>will research this.Look man, it's probably a too involved project to start out with. You'll end up frustrated and pouring out money.The conecepts scale very nice, so if you want to learn about this read up on switching power supplies or power elrcteonics in general but make sure you read high quality academic publications and follow the math.If you think you'll get value from this you can build small systems, like using 1 lithium cell and some incompatible load but make sure you have purpose like something you want to validate about your understabding.Once you want to scale it up like the trailer thing keep in mind you'll have to deal with thermals. But once you have understabding of basic electronics you really understand how to come up with thermal models (and good models for alot of other systems) too.
>>3004745desu in an engineering profession you often 'thibk of' poles and zeros as far as you find them using intuitive methods, you know 'what if DC what if AC' or 'I see a structure that is a LP/HP/BP...' for a qualitative idea and when you need stability you do the math anyways.I've given up on talking to people about how they or I 'understand' or 'think of' things. It's always just very complex, convoluted and I'd just say very very individual. I mean that's why we come up with 'ways of thinking about' in the first place. From time to time I still think about something in a new way and go 'uh-huh' but it's really mostly just redundancy.Idk draw a buck convertery see if you can see poles and zeros, then see if you can write down a transfer function and compensate the feedback.Yeah no laplace transform needed IK
>>3004745Probably best to just go through a book/webcourse on control theory. That’s where people talk about the impact of poles and zeroes on the behaviour of the circuit.
>>3004762well duh 'you got two on this side of the line and the fucker oscillates, other side and it dont... ' is what my prof always said
>>3004745You just intuit it after a while.Poles are sometimes just the number of capacitors in a filter.
I have a damaged wire in my old 360 controller that I want to fix and I isolated the issue to inside the red circle of pic related. Is the bulbous part circled in blue just a ferrite ring? And if so how important is it that it stays? Repairing this would be a LOT easier if the problem was literally anywhere other than where it is because it's close to the controller on one side and close to the ferrite ring on the other. If I have to remove it can I just use any old ferrite ring to replace it?TLDR: There is a damaged wire inside of the part circled in red. How would you go about repairing it? (Ideally without having to open the controller)
>>3004825its a ferriteits bot important but might lead to glitchesyou can replace it. It's even okay it ends up being held with shrink wrap or tape. >I have isolatedare you sure about that? >repairing would be alot easierare you sure about that? I see no differencde. Cut the cable ooen the controler, remove the damaged end, put the ferrite back on, strip back and connect the good end of the cable. Voila.>kinda weird shielded cable and ferrite
>>3002473pioneers old ohm logo kicks arse
>>3004832Those ferrites are moulded on, you can’t easily slide them off and on again. But you can just buy a clamp-on or thread-on ferrite sleeve.Reterminating that wire inside the controller is probably the way to go, but I’d want some way of adding a proper flexible strain relief. Maybe it would be easier to buy an off the shelf replacement cable with the strain relief built in.
>>3004825Get a play & charge cable or an extension cable and replace the whole thing. You should also buy a ratcheting JST crimp tool, pins, and housing (or re-use existing housing).
>>3004834>cant slideI never said you could. But the plastic can easily be removed. In addition to that if yiu cut it there you can actually slide it off just like steipping cable. To get it back on you'll again need to remove plastic / PVC from the bore. Or just toss it and get a clamp on one if you find out it was needed.>strain reliefDoes the controller housing provide none?>flexible you dont want your strain relief to be flexible, I imagine you mean you want kink preventing cable gland? Try to use the existing one or tape the sgit or whatever means available.
>>3004836>crimp toolBuy N new controllers instead for the moneyIs that think terminated using a connector?>pinsThey should be fine. You mean terminals for a connector instead?
>>3004846All of that is pointless. Just replace the cable.>>3004847With a crimp tool you can make cables for all kinds of things and pay for the tool several times over. With a new controller you just have a new controller.>terminalsYeah.
>>3004846Yeah it might be possible to remove that ferrite bead and stick it back on.>Does the controller housing provide none?The ribbed/slotted protrusion seen wrapping the cable here >>3004825 is likely permanently mounded around the cable. To clarify, I'm talking about relieving radial strain, not axial strain, though both are necessary in this instance.>flexibleIf it's rigid, then the cable just bends back and forth and work fatigues right at the end of the strain relief, you've just moved the problem elsewhere instead of fixing it. You want a strain relief be flexible enough to prevent that stress concentrating at its end, but not so flexible that it bends too much within the strain relief. In other words, something that limits the bend radius of the cable. A conical section of a sufficiently flexible elastomer seems to do a good job of this.Barely any cables off the shelf come with this kind of proper strain relief. They almost always have ribbed/slotted strain reliefs that are far too rigid for the use they get, barely flexing more than a mm or so because they're made of a really sturdy TPU. The ribs/slots are meant to be there to provide room for the thing to bend only a certain angle, but by the time you've put enough force on the cable to reach that inherent end-stop, the cable is over-stressed and bending too tight. That or they're a thin sheath that extends like 4mm from the square back of the connector that's too flexible and too short to do anything, like what you see on apple cables. The third type is even worse, a cable connector housing that ends in a nice thin taper that would be a brilliant strain relief, if the connector housing weren't entirely solid. I've seen metal versions of this. Absolute cargo cult strain relief.Looks like XLR and other commercial connectors are a bit better for this. Oscilloscope probes are usually decent too.
>>3004855>If it's rigid, then the cable just bends back and forth and work fatigues rightWell it's splitting hairs at this point but just to clarify what I was about:A strain relief is not a cable gland. Strain relief is always axial and takes load off the termination shifting it onto the housing. You always want this to be rigid.A cable gland spreads out the location of bending as to avoid violating the minimum bending radius where this is most likely to occurr. In which case the cable either simply fatigues or work hardens (copper and aluminum do), leading to fatigue, but it does not 'work fatigue', no such concept in material science. Often the cavle gland is molded on and given a geometry that serves both distinct purposes, by simply locking to the enclosure (strain relief) and providing a flexible extension of gradually decreasing stiffness (ensuring large enough bend radii).
>>3004868>A strain relief is not a cable gland.A cable gland is just to fasten a cable through a panel, usually with some IP resistance. I don't think cable gland implies any form of "cable over-bending prevention", though they usually do perform some sort of "cable tugging prevention". I'd consider both to be forms of strain relief, and it looks like McMaster do too:https://www.mcmaster.com/products/cable-strain-relief/>spreads out the location of bending as to avoid violating the minimum bending radius where this is most likely to occurrThe minimum bending radius is ultimately an emergent property of the materials of the cable, they can only elastically deform so far before undergoing permanent deformation. One might call this deformation "strain".Seems like the term "cord grip" is used semi-widely for what I'm talking about, but this term also seems to cover entirely rigid sleeves that do nothing to limit bend radius. So I'll continue to use the term "radial strain relief", though "flexible cord grip" might work its way into my vocabulary. I don't know, the term "cord grip" also kinda implies some degree of axial location.
>>3004873>material propertynot really, geometry and construction as well as design cycle life play a big role. Saw this with a cable manufacturer once even. They offered suspiciously similar cables for energy chains with differing lifetime expectancy. Their engineer on the phone confirmed what I found: The minimum bending radius and cycle life they out into the datasheet was the only difference betqeen the not so distinct products.>strainIMO it's dishonest and you know it if you now bring up the material science definition of strain which 'strain relief' obviously really does not pertain to. It literally means 'keeping stress of the termination'.Anyways IMO strain relief is defined as exactly that function and that function only: Transferring stress from the cable to the enclosure in order to keep it off individual conductors and terminations.
>>3004832>>I have isolated>are you sure about that?quite sure. Except no because I wrote red circle when I meant blue circle, and vise versa. But I've held the thing totally rigid and flexed just that one inch segment and it causes disconnects, everywhere else doesn't.>>repairing would be alot easier>are you sure about that? I see no differencde.>>3004834>Reterminating that wire inside the controller is probably the way to go>I’d want some way of adding a proper flexible strain relief.My main concern is getting strain relief right. I've repaired other cables before but none with tiny wires like this and usually the way the repair fails is from bad strain relief.>>3004847>Buy N new controllers instead for the moneyI actually tried this. Maybe I'm stupid but none of them feel right. That made sense for the two third party ones but I even got an unopened OEM official microsoft one and the stick sensitivity was all wrong. Like half the travel is deadzone, I was unironically getting nauseous when I tried to play Marathon with it. Maybe this is the universe telling me it's time to move on to an xbox 1 controller...I do have 2 third party controllers who's cables I would willingly sacrifice cause that does sound easier than cutting out the damaged section and splicing it back together. My main concern is that to solder a whole new cable on I would need to open up the controller. I've done this exactly once with a different knock off controller after onions sauce got into the buttons and not only did the buttons still stick after cleaning, but the left stick then had a slight drift that wasn't there before.On the other hand, my concern with splicing the existing cable and not having to open it up is that I don't know if I have heat shrink small enough to securely hold around then and I don't know how else to insulate the repairs without using electrical tape which can only end badly.
First time doing audio repair. I have a stereo amplifier from the early 90s. I've managed to find out that at least one of the transistors on the board is dead.Should I replace only the dead transistor(s) or all of them pre-emptively?Also, eBay and AliExpress were the only places I was able to find the correct type of transistor (UN4211). Are those an acceptable source?
>>3004958> one of the transistors on the board is deadSilicon transistors are probably good for hundreds if not thousands of years.As long as your replacement is a close enough match for the original one, you should be fine.As a general rule, I always order extra parts because the shipping eclipses the parts price anyway. If it’s a fail-y part, I include extras inside the device when I fix it (i.e. drill a couple of holes in the case and bolt in the extra TO220s without connecting them).You’re probably thinking about electrolytic capacitors which have a fluid inside them and have a more definite limited lifetime. It makes more sense to replace those even if you’re replacing the transistor.I don’t replace caps personally unless they’re bad, and when I do, I go up in voltage and get >= 105 deg C where I can.
>>3004958>ebay....aliexpress...(UN4211). Are those an acceptable source?not reallyi'd solder a couple of 10K resistors to a general purpose 2N3904 acquired from a reputable supplieror salvaged from ancient gear, such as an Apollo lunar capsule
>>3004958EBay sellers from outside of China might be reputable, aliexpress not so much. Transistors age, so you will want to adjust the Vbe-mult down before installing the replacement transistor.I’d avoid using a different transistor unless it’s an extremely close match in specs (e.g. same part made from a different manufacturer, or slightly updated version of the same part), because it’s hard to know what influence the different specs will have. But if they’re acting as voltage followers, (e.g. class B or AB) there’s not too much to go wrong.>>3004986Looks like you could use almost any of those part numbers, so long as you can add another resistor in parallel to get closer to that 10k value for R2. Adding resistance in series for R1 would be messier but still doable.
Digikey shipping went from $8 CAD to $15 CAD. They've killed their competitve edge. Might as well be ordering from LCSC now unless I need something overnight.
>>3005005Amazon also sucks now. Even with Prime, if you want one relay, you can't: comes in packs of 5 or 10 at least so your are forced to spend at least 10 bucks per order.
>>3005005I just tell a lab assistant what I need and it magically arrives at my desk over the course of the week. Feels good.
>>3005047It really hurts the economics on repairability. A few cents or dollars worth of parts + $8 was a fantastic deal. Now it's $20 baseline after taxes. Not worth it to repair shit worth much more than that anymore.
>>3005005I noticed that. Disappointing. My last order needed 14 cents to make it to $100 so I added some precision resistor.
>>3004825it doesn't mattert. guy who fixed his own in the worst kind of way by cutting it off, burning off a ground pad when soldering it back and having to glue it back together with hot glue
>>3004846the reason these cables fail is because the cable itself is too flexible. with how flexible the cable is, the strain relief doesn't do anything
>>3005090The maximum possible bend radius should transition continuously from near infinity at the chassis, to the maximum possible bend radius of the cable itself at the end of the strain-relief / cord-grip, as to not concentrate stress. And ideally this continuous curve remains continuous in various test scenarios, ranging from very light stress to near-breaking stress. Having a discontinuous bend radius curve in either direction defeats the purpose.Reminds me of impedance matching.
>>3005107i dont really know what you just said but if you've ever had an xbox 360 controller in hard you know its cable bends really really easily and eventually permanently at thatthe strain relief on the other hand is so stiff it just doesn't move with the cable at all
>>3005108Exactly, there’s a discontinuity between the strain relief and the cable. The cable has a small bend radius while the strain relief has a large one, if you graphed the bend radius as a function of distance from the controller, it would drop suddenly at the end of the strain relief. This is a discontinuous graph.
Install a mini/micro/type C USB port in the controller and use a regular USB cable to connect it, or stuff an ESP32 inside with a battery and charging module to make it wireless.
I am coming to understand why people just emulate synths in software, rather than do everything analogI am almost there lads
>>3005124>why people just emulate synthsbeen saying this for years on /ohm/takes 15 minutes to pirate and install Reason 5, and you're all finishedand it cost you nothing in effort or shekels
>>3005131An EE student came to the master and asked “why do we need build electronic circuits when we can just simulate them in SPICE instead?”The master responded “where?”The student was enlightened.
>>3005131even better: Can just ask claude to make the music for you.
>>3005124Is that conformal coating? Or just a fuckton of no-clean flux?
>>3005144I was wondering the same thing, but the chips look dry. I think it's just the angle and lighting.
>>3005145Dried flux residue then probably.
>>3005146It's just the angle, here is a close up of one of the boards >>3004351
pee hole anon here. I finally got some boards assembled. I messed up the orientation on all the opamps because the loupe I used to check orientation when pick-and-placing rotates the image 180°. so I had to resolder them all by hand. the yield so far is an abysmal 10%, mostly because the opamps got deballed. I can fix that in the next run. and hey! the design works!
would it be bad to route PCIE gen 3 tracks under a dc-dc converter like picrel? it will just be under a small section on it, not directly through, should i just rework the pcb layout a bit instead?
>>3005150You can see the meniscus around each resistor in that photo too.>>3005155Wow, congrats, hope you show us the rest of the project.>>3005184If that’s a DC-DC converter in a metal can, and the PCIE signals are differential pairs, I wouldn’t expect any problems whatsoever. You can put a PCB ground plane between them I’d you’re worried.Are you designing a PCIE card?
>>3005184Are you building a riser to convert 19V from a laptop brick to 12V for the PCIe bus?
>>3005187>>3005186No its for a robotics project with an M.2 NPU, I ended up redoing some of the power line wiring and i moved the dc-dc converter so that i was able to place the PCIE lines next to it instead
>>3005186>You can see the meniscus around each resistor in that photo too.It's just a little bit of flux OK!?!?!Actually this brings to mind a question. These were soldered with a hot plate, and I noticed that the solder paste spread out once it got warm, but before melting, it appeared to dry out and turn dull. Then it took a little while longer for the solder to melt. Would the dull/dry appearance suggest that the flux burned off/evaporated too fast and the heating rate was too slow?
>>3005197Oh so you’re actually designing the motherboard? Impressive if so.
I know driving an LCD is babby's first project, but I am a little bit proud that I have rawdogged it by sending bits into some shift registers rather than just importing some Arduino library that handles it for me
>>3005229>I am a little bit proudi'll be impressed when you create your own special characters and incorporate them into the chip's character tableand none of them are shaped like dicks or swastikas
>>3005184as long as you got shielding in betwix and balanced signals I see no issue. Put a power plane in between, or leave a GND island under the converter and route a diff pair.If your signal is single ended you're down to shielding, or choosing a converter that operates at a frequency that won't be an issue or filtering this very noise.
>>3005258>in betwixI appreciate the old fashioned parlance, but it's "betwixt". Just trying to help, Anon.
I was looking through an old DAW I had lying around and found this internal Pre amp. Thought it would be neat to transfer it to my usual PC since I use an external, but for the life of me I can't find what connector this should use. Its from a Carillon AC1, which I could not find the manual for either.Any help or advice would be greatly appreciated!
>>3005352Pick one.https://www.molex.com/en-us/products/connectors/wire-to-board-connectors
>>3005354Sweet, thanks. Do you think that would give a workable signal that can be plugged into a motherboard?
>>3005355I don't know how they laid out the boards, but it should be usable if you can power it. If they split the preamp between different PCBs then you'll have to reconstruct that part of the circuit.
>>3005356Awesome, thanks a bunch!
>>3005357Np problem, Anon. Good luck.
Can someone recommend a good short but practical book or guide to KiCAD that does not assume much in terms of electronics experience or knowledge? I'm just a brainlet learning some basic electronics and most resources I looked at were either way too in-depth or at least seemed to assume I know how parts work, what a copper pour is etc.For now I just want to design some very basic boards for now for stuff I am doing on a breadboard, like a simple nfc chip + onboard antenna, or a mount for a keyboard switch and a small display, things like thatAlso what the fuck is up with amazon being flooded by what is obviously AI generated garbage spam books when looking for KiCAD or FreeCAD?
>>3005404>Can someone recommend a good short but practical book or guide to KiCAD that does not assume much in terms of electronics experience or knowledge?Why and what for do you think you need KiCAD or any ECAD when you don't have 'much in terms of electronics knowledge'?I hope you understand the D in CAD is misleading as in it hardly assists you in designing, indtead it helps you to cleanly present a design you have already come up with.
>>3005404>very basic boards>nfc chip + onboard antennaRF isn't exactly babby's first PCB material. You should at probably have a qualitative understanding of impedance and transmission lines before you get your teeth into that kind of circuitry, plus a basic understanding of passive and active devices.
>buy xtar vx2 pro, start testing capacity of a bunch of cells i have laying around (300 mAh discharge test)>most sit at ~95% of rated capacity, but one Wurkkos cell measures at 108%is that a thing, or a measurement error? Are the tolerances on li-ion batteries that high?
>>3005404>Also what the fuck is up with amazon being flooded by what is obviously AI generated garbage spam books when looking for KiCAD or FreeCAD?It's less Amazon being flooded with AI KiCAD and FreeCAD books and more the internet as a whole being flooded with AI in absolutely every category.
Anons, you have seen me posting about this synthesizer project on and off for about a year, and today I would say I reached the stage of minimal proof of concept.I know not everyone is set up to play soundposts so I uploaded a video to catbox: https://files.catbox.moe/zo9mut.mp4Next step is to convert all of the clutter on the breadboards into a kind of "master" PCB that drives the 4 voice cards you can see in the pic.I think it's miraculous that it works at all given that it's programmed in micropython, now I need to decide if it's worth the time to rewrite the entire codebase in C, which will require me to learn C.Thanks to all the anons that looked at my circuit board layouts, answered my amateur questions, etc
>>2855435and it does work by toggling GPIO pins in python, actually, so it looks like I had the last laugh, anon
>>3005544What's the latency like? Is this a MIDI setup?
>>3005548yes, the Pi gets MIDI data over UART and uses it to control the voice cards. I can update all the cards and the LCD at about 100 updates per second. The actual audio frequency is generated by the PIO systems on the Pi, so the low update rate is sufficient to control the envelope, LFOs, etc.
>>3005560PIO is so neat
>>3005599The project couldn't exist without it, I am using PIO for frequency generation, frequency measurement (calibrating the filters), controlling the address lines, and a custom SPI implementation that also handles chip select and saves me having to manipulate the pin using slower methods from inside micropython.I admit that some of those wouldn't be necessary at all if it were programmed in C but it's still neat.
>>3005613>I admit that some of those wouldn't be necessary at all if it were programmed in C but it's still neat.what do you use besides the PIO stuff? Python?
Does anyone know where or how or is it even possible for an individual to obtain this part? how would you even go about ordering this? PE12864WRM-026-L-Qit's a tft screen for a sewing machine, mine has the ribbon cable bitten into.unless there's a way of fixing flexible ribbon cables maybe? i don't know
>>3005633https://youtube.com/shorts/1NagraXOMsc?is=MEdt0ceHwdUhUwbk
>>3005653the ribbon cable pretty much bends 180 degrees back to the pcb the screen is glued to
>>3005655You can do it. It's been done.
>>3005658Been there. Done that. Would do again.
>>3005653I feel like this could be done with some anisotropic conductive film and a lot less effort.
>>3005669Very mid solution
im working on a 4 layer PCB with two motor drivers communicating through 10MHz SPI with a microcontroller as shown, if im separating the grounds between the driver and microcontroller, with the star point at the connector, will it cause issues with EMC for the SPI line? should i connect the ground plane on layer 2 across the whole board instead?
>>3005752I don't think at 10 MHz it will turn out to be a problem but generally for good signal integrity you should have continuous ground along the length of the signal line.
>>3005752I’d want to run the SPI lines to the star ground area and then to the drivers. This is probably why you don’t see star grounding on digital designs.
>>3005653so i'll need magnet wire, solder paste(?), uv glue, uv source, probably a magnifying glass and a soldering iron. some kind of scalpel toomaybe i'll make do without the paste?
>>3005781>so i'll need......the patience of a godmost flex circuits will melt instantly as soon as you touch a soldering iron to themthis completely ruins themin that case, you might get some compatible cable from aliexpressor, bypass the broken traces using thin flexible wire if there are solder points at the connectors
So probably not technically the right thread for this, but the best thread I could find on DIY. I need to get some new routers. I'm currently running some Linksys WRT54GL routers. They're like tanks and have lasted quite a long time. Should I just keep on keeping on with the same, or upgrade to something better now? My internet speed is a blistering 11.89mbps download and 4.48mbps upload... Really it's been doing everything I need it to just fine. Would rather not pay for faster internet unless I actually needed it. So the old school routers really aren't bottlenecking my speed any at all. Any reason to upgrade to something newer?
>>3005820Security updates for firmware, more modern settings for ddns, maybe firewall stuff or DNS content blocking. The ability to run self-flashed openWRT and fully customise your service. But also some newer routers are outright anti-ownership, if I recall a few years ago Cisco or Linksys bricked some routers on purpose, or witheld security updates deliberately, so know the risks of using firmware that connects to someone else's server.But I'm hardly an expert, just shooting in the dark. We mainly do soldering and low-level design / repair here. If anyone on 4chins would know, it's the guys on /g/'s home server general. One of the less vapid parts of /g/, it's not too fast and the guys aren't obsessive about new tech, they run shit off old optiplexes all the time. But they do talk about TB how normal people talk about GB. The friendly gnu/linux general might be helpful too, to a lesser extent, since openWRT is based on linux.
>>3005820>keeping on with the same, or upgrade to something betteressentially, only three things have improved- wifi speeds have risen to 150m, 600m, 1500m, etc- MIMO technology lets you connect to more devices simultaneously- WPA2-Professional, and WPA3 have better wifi password securitythe main deficiency now is wifi speed; your WRT54 can do 54m, so it's faster than your internet but if you wanna transfer big files, like backups, it's kinda slow
>>3005824NTA but the WRT54G stuff still gets FreshTomato builds. Ancient kernels of course, but it still works. There's loads of cheap OpenWrt hardware out there if they wanted newer software and faster WiFi/LAN though.But I agree.>>>/g/hsg
>>3005807>most flex circuits will melt instantly as soon as you touch a soldering iron to themThe flex CABLES you posted do, indeed, tend to melt the instant you touch them with an iron.Flexible CIRCUIT BOARDS are made with polyimide (Kapton) and have no trouble with soldering temperatures (how TF would you make a flexible circuit if the substrate melted before the solder?).If I'm looking at the right thing (pic related), its cable has the distinct color of polyimide. You can burn it if you fuck up, but it won't melt at any normal soldering temperature.
>>3005807Just get the right flex cables.Pic related was my N64 mod.You can solder right onto it. It even has SMD parts on it.
>>3005633https://octopart.com/part/powertip/PE12864WRF-055-H-Q pick a supplieror i guess you could get a datasheet for it and then try to find something similar on aliexpress, search for 12864 with the same driver chip and then compare the cable pin out.>>3005781you don't *need* uv glue. and definitely a microscope, magnifying glass won't do.and its not called fucking magnet wire there is nothing magnet about it, its enamelled wire or enamel coated wire and there are two kinds, one where the coating kind of disintegrates when you solder it and another kind where you have to sandpaper or scrape it off first. choose wisely.>>3005838yeah normally if it goes between two fold down connectors its a cheaper white PET plastic backed thing, to save the cost of the fold down connectors the alternative is a PI (essentially kapton like you say) that is soldered so needs to withstand the heat. personally i think flexible circuit board is kind of a misnomer because either its a circuit board that you put things on, or its flexible. if its flexible then its only really flexible until you solder stuff on to it. then if you flex it, the bits can ping off.
Flat Flex Cable vs. Flexible Printed Circuit
What is the smallest device that I can plug into an electric outlet? I'm moving into a new home and know some of the outlets are linked to wall switches. I want to find out which ones before I get everything setup. I realize I could plug in a phone charger or something and test them that way. Is there a faster way? Ideally it'd be something small I can throw in my tool bag. Imagine a tiny LED with no buttons, just on/off dependant only on whether or not it's plugged in.>plug in device >flip wall switch >mystery solved I have an outlet tester but to my knowledge it's only good for testing correct wiring, not whether or not the outlet is linked to a wall switch.
>>3005853Neon lamp voltage probe. The switched outlets should be the top ones.
>>3005853>I'm moving into a new homethis is about as light and small as you getbut you really should use your socket testercoz electricians make mistakes90% are high at work, so it'll protect you from unsafe wiring
>>3005859pic
Babbies first scope time.What is the current state of cheap/handheld oscilloscopes? Anything worth buying or is it still "at least Siglet or kill yourself"?
>>3005824>>3005830>>3005831Thank you Anons, I'm quite technology retarded so this helped more than you can imagine.
>>3005853An outlet tester. Small, cheap, simple, and they can diagnose a bunch of conditions. Even basic ones (three LEDs) will tell you if there are swapped wires or a missing ground. More advanced ones with displays will tell you voltage or if an outlet is GFCI protected.
>>3005854>>3005859>>3005861>>3005890Got the probe and the tester. So between these two, I should be able to know whether or not an outlet is hooked to a switch.
>>3005866The cheapest possible FNIRSI scope set me back about 40 dollarydoos and was fine for audio frequency, single channel stuffI don't think it would be any good for monitoring digital data transmissions or stuff like that
>>3005866>>3005894A 2C53T is 50mhz (at least on paper). So it should at least cover a lot of microcontroller related stuff pretty well like PWM and I2C.
>>3005898>is 50mhzthat's the 3db point im not sure if you understand that since you seem to imply this points to an upper limit or so, only the sampling rate will give this via nyquist>>3005894>I don't think it would be any good for monitoring digital data transmissions or stuff like thatDid you use one of those awfully long ground crocodile clips and expect it not to be tardo or what is this? This is clearly not the scope and the scope instead literally shows it's quite capable.
>>3005633can you show the damage?Alot of stufd can be repaired qith some basic crafting and creativity. Lot's of atuff to get you there, never leave home without conductive ink, adhesive copper tape, kapton tape, knife and literally paper.
>>3005866I bought the cheapest dual channel portable scope I found on some eevblog forum thread. The UI is pretty shitty because it has like 3 buttons and a touchscreen, but it’s handy enough. I don’t think it decodes digital comms, but it can save waveforms to a flash drive to analyse later. Plus I’ve got a nanoDLA for logic analysing.>>3005904Never seen a scope with a bandwidth higher than it’s nyquist rate. Seems like a recipe for aliasing. They’re usually sampling 5-10 times faster than the stated bandwidth, but who knows what the filter rolloff is like.
Signal generator means I can use the device to control things like PWM motors and fans?
>>3005924>control ... PWM motors and fans?maybesome will have controls for adjusting the duty cycle, some wontand, of course, you'll need transistor buffers, coz sig-gens have weak and fragile outputs
>>3005914> bandwidth higher than nyquistMy tek scope has this mode where it does random sampling and builds up the wave form over time. It’s pretty cool, but it obviously works on periodic repetitive waveforms
>>3005927>it obviously works on periodic repetitive waveformsConsidering common pre-digital CRT scopes only worked on periodic waveforms, I'm surprised more scopes can't do this. Actually, maybe my 100MHz hybrid CRT scope does this?>Maximum sampling rate:>VC-6045: 40Msps one-channel sampling, 40Msps two-channel sampling>Maximum storage frequency:>A single-shot signal (Maximum amplitude error 30% or less): 5MHz>A repetitive signal:>VC-6045: 100 MHz (20 MHz at 2 mV/DIV)Would explain why the higher digital sampling frequencies are less useful, and jump all over the place when triggering is improper.It's also way out of calibration, gotta go through the proper steps for that.
>>3005927an ancient GHz scope I worked with had an extra module to do this. Just sampled at a frequency that is not a harmonic and overlayd it.
>>3005914>Never seen a scope with a bandwidth higher than it’s nyquist rateSuch thing has not been my claim. But yeah like I just said those used to exist with obvious limitations in the olden days.
>>3005925> some will have controls for adjusting the duty cycle, some wontCan you elaborate? Do you mean that some controllable motors need more sophisticated pwm signals, or are you just mean that there are PWM motors that are not meant to be controlled via a pwm signal?
>>3006023>Can you elaborate?the topic was signal generators, not motorsif you get a cheap one from China, for example, it may not have a knob to adjust duty cycle(PWM control is exactly that: varying the duty cycle to control speed)so, if your ONLY need is to control speed, you dont need a general purpose sig-genyou can get a PWM controller for 10-50 times less money
Everything looked so wonderful on the schematic and the breadboardbut now I gotta connect U4 to U7 and it's just awful
>>3006406Put SMDs on the other side of the board from THTs, so all the soldering is done on the same side. That or put all of those top traces (U7 pins 15-18 at least) on the back of the board so you can put U4 above U7, but then the trace order would be inverted for U4.>inb4 single sided boardTough luck, lots of jumpers.
>>3006406fuck i love laying out so much. wipe it out start again lol. the best kind of puzzle where you don't even know if there is an answer, have to just kind of intuit if something more neat is possible or not.you can clearly get those lines out by just widen the gap between U7 and U1.
>>3006408>>3006409I mean it works with vias but looks like real amateur hour.Putting just one SMT chip on the backside of an otherwise entirely THT board seems a bit rogue, but it might just solve the problem............
>>3006428If anything, I’d swap all the traces. Have all the traces going to the THT components on the back side, like you would with a single sided board.
>>3006430>I’d swap all the tracesYou know what, that is extremely sensible and didn't occur to me at all. Of course I should be running the traces on the back if everything is THT. Thanks anon, sometimes it takes another person to see things from a different perspective.It's a board for a Pi Pico to plug in to and interface with a bunch of other stuff. That SMT IC looks so out of place, maybe I should just get the PDIP version so that everything matches...
>>3006441What capacitors are those? Really small disc caps? I should probably get me some of those, they'd make routing perfboard a decent bit easier.
>>3006454just 100 nF power decoupling capacitors with 2.54 mm pitch yes
almost ready to send to JLC :^)
>>3006534Is that a ground plane? Highlight the ground net (~) and see how continuous it is. I suspect you'll want at least some via stitching traces to prevent it being so broken up. Or maybe you'd be better off putting the ground plane on the top layer, since it's less broken up with traces. Or have ground on both, and just put 4 vias around where traces on either side cross, pic related.
>>3006537I did a ground fill on both layers, so I'm hoping I can get away with it, but you are right, it does still look a bit broken up in the top right. I think I can move some of the power traces to the bottom layer which should help. Thanks for the cross-stitching tip.
>>3002473I'm designing a prototype PCB on those one sided copper plated boards but I can't avoid crossing traces. How do I avoid cursed design in such case, given my limitations? Add 0ohm resistors to jump over traces? Make dozens jumper wire of terminals all over the board?
>>3006586Yeah, jumper wires are the standard way to do it. Based on the number of old arcade PCBs I've seen, it was a pretty common solution when trying to keep things simple and cheap.
>>3006586There’s a chance you can route traces between the legs of DIP packages, but your tolerances need to be pretty good, and I’d want a solder mask. To the extent that you’re able you may be able to change footprints (e.g. TO-92-wide, staggered TO-220, longer resistors, etc.) to make more room for traces, or even change parts (e.g. 7400 logic to 4000 logic equivalent, or a quad op-amp to two dusl op-amps), and to shuffle around the components of a chip (e.g. swap U1A and U1C in an LM324). Also keep an eye out for components with internal jumpers, like pins 2 and 4 of a SOT-223 voltage regulator, or the multiple ground pins of an arduino board, though KiCAD will tell you to connect the two anyway. You can also use unused inputs of some chips, especially microcontrollers, to route signals through. I wanted a trace coming from one of the ADC pins of my arduino to the other side, so I routed it straight through a high-Z digital input, so long as I never turn it into an output it won’t be a problem.Otherwise, yeah jumper wires. I see plenty of solid core uninsulated wire used for this in old circuits, but I prefer insulated wire, I use my collection of solid core breadboard jumper wire. It may even be better to use jumpers in specific places, a long snaking ground path could form a large loop area with a high frequency signal, or just be higher resistance and allow some power supply ripple into a sensitive component.
Might be the wrong thread but is this safe? It's behind a wall socket
>>3006534 (me)This will be obvious to the rest of you I'm sure, but I was surprised to see the effect of a series resistor on my serial clock signal on the breadboard prototype. Seems like I should use these termination resistors to avoid frying downstream devices with negative voltage spikes at the edges
>>3006616That looks European. I am in the US so I don't know if our building codes apply. Here, if you splice a wire, it should be in an electrical box. Picture related. That said, that looks like good work. Maybe someone clipped a wire driving a nail or something. It looks like they used a standard wire block connector. I don't see any exposed wire. As long as it is tight, there are no holes in the insulation, and it is properly rated for the current it will be carrying it should be fine. "Safe" but, again, I don't know if it meets your building code.
>>3006620Yeah it's UK. I'm not sure on the codes but AI says you shouldn't have one like this inaccessible legally, so if I had done this as a job for someone professionally. It's behind a socket now which is working. Basically I nicked a cable with a hack saw so clipped it and bridged it with what I had so there wouldn't be a hotspot but AI says the screws could come loose over time from vibrations or if someone moves the cable (it's in a wall) and it's supposed to be done with wagu or resin clips. Now I'm worried I could open it back up and wrap it all with electrical tape but I may as well replace it if I'm opening back up -_-
>>3006616Looks like some.special vampire splice. Just shove it in the cable and crimp. Kinda cool.
I know nobody asked and maybe I'm spamming a bit, but after sleeping on this, I came up with a much more sane layout with the logic level shifter in line with the pins on the Pico. The red ground plane is less broken up now. Even added the mounting holes I forgot about the first time.
>>3006621AI is not weong. A WAGO clip would be the cleanest and imo safest solution. There is nothing wrong with your block connector and it should be fine if installed correctly. But if the screws do come loose (and create a potential fire hazard), additional electrical tape wouldn’t help.
