Anushri regulators overheat, +5 supply fails

I see another thread about how hot the regulators get. My case seems to be an extreme, because the regulators get hot enough to inflict burns and the power supply itself fails. Specifically, the +5V starts out at full value, but then sags as the 7805 heats (7905 hot too, but not as much). Before long, the circuitry gives up entirely. Have tried this in both Eurorack and standalone (LT1054) mode. No difference. Have pinged and scoured visually for shorts with not joy. I do notice a low impedance between VCC and VDD, which follows the number of chips that are in the their sockets. Fully populated, it shows about 30 ohms and that enough for my voltmeter to call it a short circuit. If I take out some of the chips, the supply remains stable although the regulators still get hot. btw, this is all on the main board. Panel board is not attached.

Genuinely perplexed. All but ready to start dykeing out bypass caps…

It does sound suspiciously like a short somewhere… It’s certainly not intended behavior :wink:

Things to try:

1/ Temporarily solder the 9V power jack and try from a wallwart with 9V, or 7.5V. The regulator surely gets hot with 12V, but with 9V or a bit less, it shouldn’t be.

2/ Are you measuring the impedance between VCC and VDD with the power on? Don’t do that. I don’t have a built board at hand but this should be around 100 ohms.

1/ Done that. Result same.

2/ Measuring with power off.

Trying to isolate, I find that with the processor only inserted, the regulator overheats and fails. So far, I cannot find any shorts of pins on that chip to Gnd, VCC, VDD or anything else. I’m still looking.

So the power supply works without the ATMega(processor)?

It could be anywhere on the board… and could also be something really small… Try without the control board attached if it still gets hot…

V cent. That’s correct. I’ve also tried an unprogrammed part in there just in case. No difference.

shiftr, I separated the control bd some time ago.

Next plan is to lift all pins on the ATMega (no other chips on the board), except VCC, VDD, the 2 clock pins, and rst. But that’s tomorrow. It’s 4 am where I am and I’m fried… Thanks, everyone.

Revisiting this morning, I find that it does appear to be the ATMega processor itself that has failed.

1. With all chips removed, and no power connected, I’ve clipped my ohmmeter between VCC and Gnd. It measures 8100 ohms.

2. Placing the programmed Anushri processor in the socket, the measurement plunges immediately to 30 ohms

3. Replacing the Anushri part with the ATMega chip from my Arduino board, the measure resistance is the same as if the part is not there, 8.1 kilohm. Return to step 2, it drops once more to practically a short circuit.

At this point, I don’t really see an alternative to replacing the chip. I anyone has other suggestions, I’m open. Also as to how to proceed when the new part arrives. Needless to say, I’ll proceed with care at that point.

For shipping, please use a registered service. If need be, I’ll cover the cost. Ordinary post has an unfortunate history of getting lost on its way to me here.

Thanks much. I find this issue quite unexpected, and I might doubt it myself if it weren’t sitting in front of me.

Replacing the chip seems the way to go then - MI sends via DHL Express, that’s at least extremely fast, and trackable…

Thanks much!

An unprogrammed ATmega boots into a state in which all pins are high-Z. The programmed chip boots into a state in which some pins are configured as outputs - some of them high, low, or wiggling… So if the problem is that one of the output pins of the microcontroller is shorted to ground or to VCC, you won’t see the problem with an unprogrammed chip, but you’ll see it with the programmed chip.

Have you checked for that?

Yes. The test described above was done without any power, so chip initialization would not be a factor. I went back to check again and isolated the fault yet further, but still internal to the chip I’m afraid.

On the chip itself (outside the socket and not under power), pin 12 (I/O SCK) is shorted to pin 22 (GND). Again that’s on the unpowered chip itself, unconnected to anything. It does appear that the Anushri part is shorted internally.

Please do send the new part. I think we’ve done everything possible to qualify this. Thanks. If you like, I can send the failed part back to you for post mortem.

Send me a private message with your address.