• Please review our updated Terms and Rules here

Replacement IBM 5150 Power Supply Making Continuious Tone

tempest

Veteran Member
Joined
Sep 16, 2014
Messages
930
Location
SE Michigan
I have an IBM 5150 that had a failing power supply, so I got a new one for it. It's not an original 5150 power supply, but an XT style one by Power Tronic. It has the P8 and P9 connectors and they're the correct style for the 5150/XT so I figured I was good to go. When I plugged it in the speaker emitted a continuous tone so I shut it off and tested the voltages. All the voltages are correct and I know I have it plugged in correctly (black ground wires towards the center). So what could the problem be? The system is ok because when I plug the old failing one in it still boots up, so it must be something about the new power supply it doesn't like. Any ideas?

powertronic.jpg
 
Check the level on the PG (power good) line--or just disconnect it temporarily.

I assume that you tested the voltages under load--i.e., while the PSU was connected to the motherboard.
 
Check the level on the PG (power good) line--or just disconnect it temporarily.

I assume that you tested the voltages under load--i.e., while the PSU was connected to the motherboard.

The PG line tests 5V. I'm not sure how to disconnect it short of trying to pull out the wire.

Yes I tested them connected to the board.
 
Okay, so let's think about this a second.

A power supply doesn't do much, so here's a plan:

Measure every pin on the P8 and P9; don't assume anything. Do this with the original and then with the replacement, changing nothing else. Compare. If there really are no differences, let's look at the PG pin. It could be rising too fast. With the replacement supply plugged onto the motherboard and the speaker screaming away at you, take a test probe or wire and ground that line for a second or two. Does that make a difference?
 
That worked! But after a while of the system being off I had to do it again. Is there a way to permanently fix this?
 
That worked! But after a while of the system being off I had to do it again.
Which supports Chuck's hypothesis of 'PG pin. It could be rising too fast.'

Per the diagram at [here], after turn on of the PSU, there is a delay before the POWER GOOD line goes high. The motherboard requires that delay. In the IBM 5150, that delay needs to be 840 ns (four CPU clock cycles) or more.

If you had a digital storage oscilloscope (and the knowledge of how to use it), you could measure the delay of POWER GOOD out of your PSU (an example shown at [here]).

Is there a way to permanently fix this?
Assuming that 'inadequate POWER GOOD delay' is the problem cause:

* Taking the necessary safety precautions, examine the circuitry in the PSU that generates the POWER GOOD signal to see if the circuitry is faulty. Maybe your PSU is of the 'cheap' type where the POWER GOOD signal is faked. The nature of the fake may be POWER GOOD circuitry consisting of simply a resistor and capacitor. If so, maybe the capacitor is bad.

* Add 'fake' POWER GOOD circuitry between the PSU and the motherboard. An example is shown at [here].
 
If you ground that line, your CPU will be held in reset mode. A simple comparator and an RC circuit should work as a substitute. Or you could go with a ready-made solution, such as MAX803.
 
The MAX803 is a three-terminal device, so there's not a lot of "there" there.

You can try to cheap out a bit, if you're game by trying something like this. Juat a cap and resistor.

I'd add a high-value resistor across the cap as a bleeder; maybe 100K. You could also drive the input of a comparator if you're fussy about voltage levels.
 
The issue seems to have gone away. The only thing I did was hook up all my peripherals to the power supply. Maybe the extra draw was enough to fix the issue? I have no other explanation. I'm going to keep testing it periodically before I declare the issue fixed though.
 
Back
Top