I've got a D12246 Williams power supply board in a Taxi machine with no 5V. (No load it gives 4V and even a small load, pulls of the display drivers, it pulls it down to 3V). The transistor is a ST 2NG057. I desoldered it and I'm planning on replacing it with a 12V to 5V regulator that is rated for 5A. It should be more efficient, and won't use the old comparator circuitry.

I verified the rest of the voltages are rock steady with no ripple issues. (+12V, -12V, +100V and -100V). I found and fixed the shorted coil which toasted the transistor and shorted the 5V supply. Since the 5V is the only voltage that is bad, it seems best to just fix that portion.

Has anybody tried anything similar? It seems like it should work and be significantly more efficient than the previous design. Anything I should be worried about?

Thanks for the info. Your transistor number helped me track down another part that is still around. Since I can't find the spec's on the transistor, I'm really wondering how much 5V current is needed on a Taxi machine? I have a 5V LDO coming in which can support up to 5A. Some of these transistors can handle up to 12A, so I'm not sure if I under specified the part.

Short on the 5V line was caused at Q73 and Q75. Those transistors caught on fire and caused a hard short between the base and the emitter. I removed those two transistors (or what remained of them) and verified the short was cleared. (I also replaced the shorted coil that may have been the root cause of all this.) (Funny part is that I was promised this machine when it was working just a few months ago. I had a friend play some games on it before agreeing to the deal. What did they do to it?)

At that point I unplugged all the connectors from the power supply board and verified voltages. I also verified there is only a couple of millivolts of ripple. I looked at the ripple just to make sure that ripple wasn't causing my 5V issue.

5V with only the AC connector plugged in was reading 4V. I tried to load it a little bit by adding the display connector. It went down to 3V.

I've had problems with aged caps before and large amounts of ripple, but I'm pretty sure that is not the case with this machine. Currently I'm just hoping that they didn't cause destroy the MPU board. I have yet to see this thing boot successfully and that is the point when I know I can fix anything else that is wrong.

Thanks for all the info guys.

So C7 looks like the decoupling cap with C8 being the power supply smoothing capacitor going to the voltage regulator supply voltage. If C7 is not working, I should see a lot of ripple across C8. I already desoldered the transistor at this point, but I'd love to know what the root issue is. If there is a large amount of ripple on that pin, it would certainly reduce the voltage on the regulated output. I'm going to plug the board back in and see. Should be a quick 5 or 10 minute test tonight.

Regardless, I'm probably end up by just using a voltage regulator to get around that whole circuit and make the power supply 30 or 40% more efficient. (Damn, now I can't keep my coffee cup warm by sitting it on top of the pinball head). We'll see how it goes. Bummer is I won't get the parts until Monday. Arrggggh.

I got a switching regulator that I was planning on putting in, but it is definitely a point well taken that it will look like a kludge when I put it in. I just measured the voltage and across c7 is 5.2VDC with around .2 VAC ripple. That seems a little high for ripple to supply to that chip. Pin 4 is measuring about 1.1 VDC, and Pin 5 is measuring about 2.4 VDC. I inspected both C7 and C8 and I don't see any leakage or bulging at the end.

There is a little toastiness noticeable on the AC connector to the power supply board, which may indicate it had a high current event.

Thinking about that circuit more, I'm guessing the voltage across c7 should be more than 5.2VDC. Anybody have any good guesses on what a good value is for that voltage? If that is the case is C7 (the cap being used for the boost) or C8 (the cap being used to reduce ripple) the most likely candidate?