Is the input voltage for that regulator 12V? Then you should expect it to get hot. Suppose the load on the 5V output is 100mA, then it would already have to dissipate 0.7W. That doesn't sound like much, but concentrated on such a small surface area, it would already be too hot to touch.

Yeah, the input voltage is 12V. I know it's possible for it to get that hot but I have not encountered in on a sound card yet. At least not to such a degree. That's why I would like some confirmation from another owner that other cards behave the same.

I don't have the same card, but I tested my Aztech card that also has a 12V to 5V regulator: https://theretroweb.com/expansioncards/s/azte … ro-16-ii-3d-pnp

It went to 43C after a couple of minutes. But your card has more parts, so it would be reasonable to assume it gets a bit hotter. One thing you could do is touch all the components on your card, especially the capacitors, to see if anything is hot, indicating a short or high leakage current which would also increase the current through the regulator. But as this card has lots of capacitors, even a smaller leakage current in all of them could result in a substantial current draw.

Already tried the "touch" test and nothing seemed like overhating. There is a weird small resistance between the GND and 5V out pin on the regulator (~0.8kOhm), other cards I have measured ~8kOhm. Took the regulator out and it is ok so some other parts is causing this small resistance. Checked some tantalums and some through hole ICs and it seemed that it originates from one of the Aztech chips which are powered by 5V. I don't know, maybe it is ok but the regulator is definitely in 60+°C as I cannot hold my finger on it longer than a second or two.

I have certainly had 5v regulators get really, really, really hot. It all depends on the voltage/current. I kind of doubt there is a problem. More recently I repaired some cheap PA speakers for a buddy - bad solder in both of them in the preamp section, right on the legs of the 7805's. These things were working with like 30volts in and NO HEATSINK for years....no wonder the connections cracked....BUT a resolder and heatsink and all was well, they can bake for another 15 years.

I don't think you have to worry since you added a heatsink. If it won't let you sleep you could try replacing some of the tantal capacitors on the rail that thing feeds, however iirc those usually fail with a bang or a short, so if the card functions fine I think that's unlikely to be the case.

Indeed aztech cards can be a pain driver wise. I have a newer one (end of ISA pnp era) and it's not much better in that regard.

As for the wave table connector missing- that was pretty standard across manufacturers to ship "cost reduced" cards like that, so I'd think that's factory, unless of course there are some obvious signs of it being de-soldered, unlikely but still possible I guess.

If everything is working as it should, but power delivery parts get hotter than usual and you get low resistance readings from power lines to ground, it is likely just leakage current from aging capacitors. You could desolder a couple of electrolytic capacitors that are connected to the regulator's output pin and test their leakage current.

bertrammatrix wrote on Yesterday, 15:52:

As for the wave table connector missing- that was pretty standard across manufacturers to ship "cost reduced" cards like that, so I'd think that's factory, unless of course there are some obvious signs of it being de-soldered, unlikely but still possible I guess.

The connector was definitely desoldered, and not exactly by a professional. Front side looks good but the bottom side had lot of flux residue and some damaged hole platings. I just filled the all the holes with solder as the connector is not MPU-401 compatbile anyway IIRC.

asdf53 wrote on Yesterday, 16:09:

If everything is working as it should, but power delivery parts get hotter than usual and you get low resistance readings from power lines to ground, it is likely just leakage current from aging capacitors. You could desolder a couple of electrolytic capacitors that are connected to the regulator's output pin and test their leakage current.

That's worth a shot. They look okay but the brand is not exactly confidence-inducing and they're several decades old. Still, I'm a bit sceptical this will help in a significant way. But it's worth a shot.

Does the regulator only get to 60C though? Because if so, I really wouldn't bother. The vast majority of the regulator's current will still come from the chips on the card. So even if you solder in new caps, don't expect it to drop by 10 or 20 degrees. The potential gain would not be worth the trouble.

I wrote 60+C as this is the temperature which burns skin so it's at least this. But I feel it's higher, maybe around 70-80. Unfortunately I don't have the means to measure it.