Maker Pro
Maker Pro

Old computer power supply has low voltage.

Landon

Sep 18, 2013
7
Joined
Sep 18, 2013
Messages
7
A friend of mine has an old 8088 computer that he wants to get running again, the computer turns on like it should but it gives keyboard and hard drive errors and there it sits, what I ended up finding out is that the +5 volt output of the power supply is fine but the +12 volt output is only putting out +9.6 volts.

Normally I would just replace the power supply but this thing is OLD very proprietary and only one came up on eBay and they wanted $50 for it, so I am trying to fix it.

The +12 volt output has a 1000uF 16v capacitor on it which I pulled out of the circuit which then caused the voltage to fall to +7.xx volts, the capacitor looked fine (no brown goo coming out of the top of it) but I replaced it anyway.

The output is still at +9.6 volts though, and what is odd is that the input voltage to the output diodes is reading as 20 volts AC which then goes through two PR3002 3 amp diodes that have their outputs soldered together, with a standard 0.7 volt loss per diode it should only be loosing 1.4 volts on the DC side.

But... I pulled them out of the circuit they each test infinity in one direction but they are both testing at about 0.478 on the "Diode Tester" function of my DMM instead of 0.7xx like silicon diodes should be reading as.

So...

My question is should these two PR3002 diodes be reading as 0.478 or should they be up higher in the 0.7xx range? Or am I barking up the wrong tree?

And before you refer me to the data sheet I already looked in several different places they all list of lots of useful info just not what a "Diode Test" on a DMM should read.

Data sheet: http://www.datasheets360.com/pdf/-6167918509033483484?comp=3559

Also, the circuit board in this is listed as a: "Leadman PC-130A".
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
Schottky diodes have a lower voltage drop than standard silicon diodes.
 

Landon

Sep 18, 2013
7
Joined
Sep 18, 2013
Messages
7
Schottky diodes have a lower voltage drop than standard silicon diodes.

Ok, so I was barking up the wrong tree, I figured that that those two PR3002 diodes were fine the way that they were testing I just wasn't sure.

Considering that I am getting 20 volts AC going into the diodes and there is no load on it to cause a brown-out/voltage sag on the AC side it has to be something on the DC side causing the voltage drop.
 

goldfist

Sep 18, 2013
31
Joined
Sep 18, 2013
Messages
31
Too bad I didnt see this post before destroying a good 8088 power supply to a IBM 5150 that I am modding to run on modern guts.

But anyways to fix your supply, if there are any tantalum capacitors in there, I would replace them. As for I have had old computer power supplies act like this and have a small tantalum fire when they flash over burn. When tantalum caps fail due to crystals, they short, and you might have one that is shorting ever so slighly to make a load on your 12v side and cause it to be weak. I would also remove all cards from the 8 bit slots as well as any drives and measure the power supply with minimal load and then if the voltage is better like 11v, shut it down and plug in a device and turn it on and measure etc, as for you may find that the problem is elsewhere than the PSU itself.

The last old computer to catch fire from aged tantalum capacitors was an old Northstar http://oldcomputers.net/northstar-horizon.html
 
Last edited:

Landon

Sep 18, 2013
7
Joined
Sep 18, 2013
Messages
7
Too bad I didnt see this post before destroying a good 8088 power supply to a IBM 5150 that I am modding to run on modern guts.

But anyways to fix your supply, if there are any tantalum capacitors in there, I would replace them. As for I have had old computer power supplies act like this and have a small tantalum fire when they flash over burn. When tantalum caps fail due to crystals, they short, and you might have one that is shorting ever so slighly to make a load on your 12v side and cause it to be weak. I would also remove all cards from the 8 bit slots as well as any drives and measure the power supply with minimal load and then if the voltage is better like 11v, shut it down and plug in a device and turn it on and measure etc, as for you may find that the problem is elsewhere than the PSU itself.

The last old computer to catch fire from aged tantalum capacitors was an old Northstar http://oldcomputers.net/northstar-horizon.html

The power supply is sitting on my bench with nothing hooked up to it (no load at all).

The original fan had a cracked fan blade which was hitting other things and making a lot of noise which is what prompted the PS to be taken apart in the first place, is was replaced with a fan that had a slightly higher amp rating but just to rule it out I unhooked it the voltage went from 9.6 to 10.2 so there apparently is so kind of load on the circuit board holding the voltage down.

The +12 volt DC path starts at the two 3 amp output diodes which I have tested and are fine, then it goes through some windings in a donut shaped ferrite core, then it goes past the + side of a 1000uF 16v capacitor that was just replaced, from there it goes to the wires that come out of the power supply, so there really isn't a lot to go wrong on the DC side of the circuit.

The +5 volt path also goes through the same ferrite core but the +5 volts is still +5 volts on the other side of it so I don't think there is any kind of weird dampening field going on inside the core that would cause the +12 volts to drop to +9.6.

Oddly though when I disconnected the output of the core for the +12 volt path it shot up to +17.x volts, apparently it does do something for voltage stablization.

Also I did not see any tantalum capacitors on the circuit board at all, they are either polarized electrolitics or disc ceramic.

So I am back to wondering if there is something going on on the AC side of the circuit and if the 20 volts AC that is getting to the output diodes is not high enough to end up +12 volts DC when it gets to the end of the circuit?
 

Nobody

Aug 31, 2013
12
Joined
Aug 31, 2013
Messages
12
Do not run old PC switcher without load! most of them do have minimum load requirements. At leas load them with dummy load, like 0.5A each. Otherwise it might go into over-voltage. Sometimes there are zenners parallel to buses, they might blow short to protect the rest of the electronics.

In switchers typically the highest power (in Watts) output is the one used for feedback/voltage stabilization.
8088 used to power CPU from 5V and 5V bus usually has highest power rating.
If it sag - it will be compensated by boosting duty of the driver (there is a limit when it blows). So, probably 5 V is the one with bad caps (counter intuitive, but it is the way it works).

You may determine primary feedback bus by tracing optocoupler (or small signal transformer), fix it first to be safe. Once primary is OK - you can do the rest.

It is possible to have two stage power with CPU getting 12V and then step down to 5V. So, 12V bus will be the biggest. It is possible to have two separate 12V buses, one for CPU and another for ISA and the rest.

Oh, and most power buses have two caps, first after the rectifier, then another one after the choke. First one will affect voltage, the second one will affect ripple/noise.

Just change ALL electrolytics in it and most probably everything will return to normal. If PC has second stage - 2 or 3 phase - replace these caps next to CPU also. Yeap, this is a lot of caps. 8088 got to be 20 - 30 years old, most caps by now can be dry without leaving brown goo.
 
Last edited:
Top