Resistance measurements on a power rail are useless. Use diode mode. It would be in the realm of possibility for 74 series (non HC/T) to use that much current. The original bipolar versions could be quite hungry depending on which chips and the circuit.

Did you only test the +5V rail? You should check +12V with your lab supply, that's more likely to have the shorts on it for a few reasons.

Which diode failed in the PSU? One of the main flyback diodes?

Hello Kingcake

Thanks, I didn't know that.
In diode test mode the meter shows 0.02V drop, both polarities.
The 12V rail doesn't have a pad on the isa connector of this tape controller card, so we can rule that out.
Well most of the chips appear to be 74LS or just 74 chips. They're all from 1983-1986...

As for the IBM card
The diode test on the 5V rail says 0.049V both polarities and on the 12V rail it is 0.424V in the correct polarity and 0.394V in reverse
The -12V shows 0.792V in forward and open circuit in reverse
-5V rail shows 0.22V forward and reverse.
Unfortunately I don't know enough about electronics to come to any conclusions with these values.
My wild guess would be that 5V is the culprit because diodes and transistors shouldn't start conducting at such low voltages?
Would be happy to understand what's going on here.

As for the PSU diode, I'm not sure what it does I just looked for a shorted component after the bridge rectifier and that's what I found. In the picture there is 2 big power resistors and right next to it are four diodes. The top one was a BA159. I replaced it with a F107... should be alright the almost have the same spec.

Thank you for taking the time kingcake

audiocrush wrote on 2025-08-16, 11:20:

Hello Kingcake

As for the IBM card
The diode test on the 5V rail says 0.049V both polarities

Yeah that sounds like a short. Basically, you need resistance to cause a voltage drop. (i.e. you need resistance to flow to create pressure. same math). So very low voltage drop (your meter is showing voltage drop) means nearly a dead short somewhere on that rail.

You said you tried injecting +5V and checking for hot chips, right? What settings did you use on your lab power supply? Use 1V and high current. You need resistance to make heat. If a short is extremely low resistance, a dead short, it will make very little heat unless you pump quite a bit of amperage.

Also, judging by the location of that diode, it seems to be a part of one of the transformer snubbers. It's definitely not one of the main output diodes. Those are on the left and are much beefier. If that was bad, and a driving transistor/mosfet lost it's snubber network, it could have experienced high voltage spikes which caused breakdown. If the PSU still doesn't output I would suspect whicher transistor/mosfet is associated with that diode. Others more knowledgeable of PSUs might know more.

Hi,

The user modem7 on this forum appears to have a working card and his ohms measurements appear to be the same as on my card.
https://forum.vcfed.org/index.php?threads/ibm … -trouble.53100/

I find that a bit weird but when I do the maths...
5V/85Ohms should be 58mA right?
So thats 0.29Watts for the whole 5V portion of the card. Seems very reasonable to me.

Yes when I injected the 5 Volts on the other tape controller card I just set it to 5V and turned the current to 0mA, connected the 5V and ground pin and then turned the current up until it hit the 5V limit.

Hmm maybe Ill just give everything a try once I got back home from work tonight. Maybe these cards aren't as broken as I thought and it's been just the psu all a long.

*edit*
I spent some more time on the psu and I continued digging on the high voltage side of it.
There is a triac and I think it is blown... 0ohms between MT1 and MT2
I ordered a spare hopefully that'll do it...
Its really a shame that IBM used whichever PSU they could get for that thing so it is basically impossible to find a schematic for it
Desoldering the bridge rectifier really helped tracing the fault.
I had 220 ohms on the ac input which appears to be way too high.
I really wonder what this triac does... also since it does not appear in the "reference"-schematic of a similar psu on minuszerodegrees.net