Yay 1k dumbasses!
I took a few weeks off work and I’m getting back into the building game. Made a resonant eq out of an aliexpress module. Good thing I paid attention to the module orientation and didn’t get all +/- signs backwards. Right???
Yay 1k dumbasses!
I took a few weeks off work and I’m getting back into the building game. Made a resonant eq out of an aliexpress module. Good thing I paid attention to the module orientation and didn’t get all +/- signs backwards. Right???
i love it even more that they’re backward
Maybe you can transform the + in - and the - in + with a black, and a white Poska
Not synth related and I’m not sure who’s dumbassery is to blame, but this might be amusing to some.
My Son wants to add a Leopard Gecko to his small menagerie and the latest in reptile comforts is a deep infrared heating lamp that requires our subject, a dimming thermostat. Looking to save a few dollars I opted to order a Thermostat from Amazon US. This device has no listed input voltage on ANY listing I could find, however the part number is identical at every store in every country I checked.
The device arrived and its labelled 120v, not 120-240v like I was hoping, local mains being 240v So I opened it up and started to figure out the circuits, triac is good for 600v, high wattage resistor, Zener diode, 78L05 voltage regulator leading to the microcontroller. So far it’s looking good.
Then there’s the sense line to the microcontroller, a few high value resistors, a blue capacitor labelled R1, weird, and a few more 100k resistors, again seems reasonable.
I connect the output to a spare halogen lamp and by slightly questionable means connect it to 240v mains with the power switch close at hand. Everything runs fine for a few minutes then magic smoke starts to waft up off the board, lamps still running, but I cut the power.
Curse a few times and inspect every SMD component for damage and find nothing. Draw the relevant power circuits in CircuitJS and simulate, everything is exactly as I would expect. Scratch me head for a day and then decide the best option it to hook it all up again.
Second run is short and I just verify voltage to the MCU is 5v and sense line is <5v.
Third time it works fine for a few minutes then, BANG! fireworks! WTF!
The blue “capacitor” labeled R1 has let out the magic smoke in the most spectacular fashion! Luckily, dispite it being wedged behind a filter capacitor I had managed to note the value of “271k”. I went to the random capacitor tray in my junk box and start rummaging for a replacement and the closest blue “capacitor” is labelled 10D391k with no clear voltage spec., so I google for a data sheet.
Dumb-arse! R1 is a Varistor! Turns out the part I have on hand is a suitable replacement, plus a new 3.15A250v fuse from a scrap power supply unit and I seem to have a working Thermostat (still needs a soak test!)
But that 10D271k Varistor? It should have been rated for 270v … local supply is 240v …
Am I missing something? Am I a dumb-arse for blindly assuming a product with the same part number will be the same spec regardless of country of sale? Was this a faulty unit from the get go?
Over the top solution.
Just say it’s “+/-” attenuation. Job Done.
It’s rated for 270 VDC typical, minimum 243 V. https://www.bourns.com/docs/Product-Datasheets/MOV10D.pdf
That minimum is pretty close to your nominal 240 V.
But… that’s VDC. Wall voltage is 240 VAC. Peak wall voltage is √2 times higher, or 339 V! It only goes above 240 a few ms per cycle so maybe that’s not so terrible, but maybe it is.
Oh wait, I see it says on the datasheet maximum RMS voltage is 175 V. So no, 240 VAC is not ok!
Your replacement is rated for 250 V RMS, which is closer than I’d like to 240 V but is over so in theory should be all right.
Thank you! That explains it.
The replacement is likely pulled from a 240v rated power supply, so hoping that it was used in a similar manner. NZ “240v” mains is actually 230v RMS, so a little more breathing room. Going to give it a good soak test on the bench this weekend.
Realized that I left in an accidental 1k/100k voltage divider on the V/Oct input of this board a day after JLC accepted my order, welp.
(Should be fixable in software, or by removing a few SMD resistors and normalling the jacks to the ground, though.)
It would look a good deal less awful, and be much easier to check the routing, if it weren’t for the ratsnest lines running from one board to the other. I use different net names on different boards including different power symbol names, e.g. GND and GND1. Then when KiCad says there are unconnected items you know it’s something you have to fix, not a fictitious connection between boards.
This is something I’m still conflicted on. I’ve tried it both ways. The downside of separating the boards like that is that the schematics become a great deal harder to read. What I would really like is a kicad plugin that allows me to draw then hide wires between the connectors, so the ERC sees them as connected. Seems like a pretty straightforward thing to code, but I haven’t found it yet. And yes, I should probably just write it myself at this point…
Wiring up the last transformer panel made me realise I might have forgotten something on the first two…
Oh who doesn’t love exposed 240V connections in the back of the synth? Some excitement…
Mine go in their own case, which reminds me I have one that’s finished and needs a case. I think i have some power con connectors somewhre spare…
OK I’ve got one. I was building a single channel slew limiter based on a stripboard design from unauthorised service on youtube. None of my builds were working until I realized I was shifting the power headers one row down. I was feeding -12v up the back end of a capacitor lol. I’m lucky nothing exploded. It made some weird and wonderful noises I recorded on my 4 track. If there’s any interest, I’ll upload it.
Here’s the stripboard if any of you are interested.
Oh man I fucked up so bad on one of these boards
It’s not the mistake that’s getting to me, it’s the not noticing mistake until after all of my last pots have been soldered to the board, everything is soldered in place, it’s time to put the chips in aaaaand… I’ve used a 16 pin footprint instead of a 14 pin one. Everything is routed wrong.
I’m going to try to make a frankenchip with wires for legs and hope that fixes it.
Closest leg of the encoder is shorted to the switch contact by a solder bridge