IBM5160 PSU output voltage issue

[hozone]

Hello,

I'm restoring an IBM5160 (find pictures here: https://forum.vcfed.org/index.php?threads/ibm-5160-restoration.1242913/)
It works fine, but after 1 month that I don't power it up I've a problem with the 25Mb HDD - find the problem here: https://forum.vcfed.org/index.php?threads/type-13-20mb-wd25-hdd-clicking-on-ibm-5160.1243668
Yesterday I was checking the power supply, and i find something strage. Maybe it could be the cause for the HDD?
With PSU disconnected I can read
+9.7V (on the +12V line)
+4.7V (on the +5V line)
With all connected (and PC powered on)
+10.5V (on the +12V line)
+4.4V (on the +5V line)

The 9.7V when disconnected and 10.5V when connected makes me think the caps on each device connected stabilize the output and take it up.
I'm wondering if the voltages I'm reading are "acceptable" or I have to investigate furter and eventually try to repair the PSU. Also I'm wondering it this may be the cause for the malfunctioning HDD.

Thanks!

 

[modem7]

With PSU disconnected I can read

For the power supplies used in the IBM PC family, it is usually invalid to measure the DC voltage when the power supply is unloaded. That is due to the design of the power supply.

With all connected (and PC powered on)
+10.5V (on the +12V line)
+4.4V (on the +5V line)

I'm wondering if the voltages I'm reading are "acceptable" ...

So, power supply loaded.

Per [here], both of those voltages are unacceptable according to the applicable IBM manual.

( Odd that the IBM power supply still produced the POWER GOOD signal. )

... or I have to investigate furter and eventually try to repair the PSU.

You are convinced that the power supply is faulty. Is that because you get good results when using a spare power supply ?

With all connected (and PC powered on)

So, motherboard, hard drive, floppy drives, keyboard, and all cards.
Does the +5V increase from +4.4V if you remove a significant portion of the load, say, remove the hard drive and its controller ?

Also I'm wondering it this may be the cause for the malfunctioning HDD.

Could be.

 

[hozone]

You are convinced that the power supply is faulty. Is that because you get good results when using a spare power supply ?

No, I did not try any other supply. I just think it's faulty cause the HDD does not work at boot. And cause the voltage readings are low.

So, motherboard, hard drive, floppy drives, keyboard, and all cards.
Does the +5V increase from +4.4V if you remove a significant portion of the load, say, remove the hard drive and its controller ?

Yes with motherboard, hard drive, floppy drives, keyboard, and all cards connected the 4.7V became 4.4V and the 9.7V became 10.5V.
In other words I read 4.7V when unloaded, 4.4V when loaded on what it's suppose to be the 5V line; 9.7V when unloaded, 10.5V when loaded on what it's suppose to be the 12V line.

One thing I forgot to mention is that I had a short on the +12V or -12V line, i forgot which one. The tantalum exploded and another was in short. Maybe that can have dameged the supply too.
I have to try with another supply i suppose.

 

[modem7]

With all connected (and PC powered on)
+10.5V (on the +12V line)
+4.4V (on the +5V line)

So, motherboard, hard drive, floppy drives, keyboard, and all cards.
Does the +5V increase from +4.4V if you remove a significant portion of the load, say, remove the hard drive and its controller ?

Yes with motherboard, hard drive, floppy drives, keyboard, and all cards connected the 4.7V became 4.4V and the 9.7V became 10.5V.

Power supply unloaded = {Very likely invalid - no point in measuring}
Power supply loaded with everything: motherboard, hard drive, floppy drives, keyboard, and all cards = +4.4V, +10.5V
Power supply loaded with only motherboard, floppy drives, keyboard, and video card = ?????????

The power supplies used in the IBM 51xx computers are switch mode power supplies (not linear power power supplies), and at the time, usually designed to require a minimum amount of loading for proper operation.

We have seen cases of deteriorated power supplies that work well when lightly loaded (and that load is still above the minimum specified by the maker), but start have a problem when the load is increased.

One thing I forgot to mention is that I had a short on the +12V or -12V line, i forgot which one. The tantalum exploded and another was in short. Maybe that can have dameged the supply too.

Very unlikely.

I have to try with another supply i suppose.

If you have a spare, that quickly proves that the power supply is the problem.

 

[hozone]

Power supply loaded with only motherboard, floppy drives, keyboard, and video card

It's +4.4V and +10.5V.
I've never notice that if floppy does not contains OS it boots with IBM basic!

If you have a spare, that quickly proves that the power supply is the problem.

I don't have one, I've to search for one.

 

[Timo W.]

Easiest way to check if the PSU is good or not: connect some other spare HDD to it (and ONLY that). This way, the PSU has enough load and you have nothing else connected from the system that acts odd and might be the culprit. Measure 5v and 12v again. Still too low? Replace the PSU.

 

[hozone]

Thanks, with just one "modern" HDD connected I can read +10.5V and +4.5V. I think my PSU has some problem.

 

[Timo W.]

Yes, that is too low. Might be just out of adjustment, but since working inside a PSU is very dangerous, I'm not going to give any hints on how this could be fixed.

Next thing to try is of course with a working PSU (any AT PSU will do) to make sure the system is working properly.

 

[hozone]

I have a maybe one AT, but I've to check. Does the connector are the same?

 

[Malc]

Thanks, with just one "modern" HDD connected I can read +10.5V and +4.5V. I think my PSU has some problem.

Depending on the PSU you may need to load it using more than just one 'modern' HDD, Try it with 2 or 3 modern HDD's connected, Better still have you got an old power hungry MFM HDD that spins up ?.

 

[Chuck(G)]

I'd probably start by opening the power supply and measuring all of the electrolytics. But then, that's just me.

 

[Timo W.]

Depending on the PSU you may need to load it using more than just one 'modern' HDD, Try it with 2 or 3 modern HDD's connected, Better still have you got an old power hungry MFM HDD that spins up ?.

The PSU in the 5160 does not need a huge load. That was only the case with the PSU of the 5170.

Also, what he measured is the same as it was with the whole PC connected, so that's all the PSU delivers.

I have a maybe one AT, but I've to check. Does the connector are the same?

Yes, they are the same. When connecting, make sure the black wires meet in the center.

 

[Malc]

The PSU in the 5160 does not need a huge load. That was only the case with the PSU of the 5170.

Also, what he measured is the same as it was with the whole PC connected, so that's all the PSU delivers.

Note i said, 'Depending on the PSU', IBM used various PSU manufactures, Not to mention the Clones around, The OP's 5160 may or may not have the original 130W PSU fitted. I wouldn't condemn a PSU until i have tested it on the bench suitably loaded.

 

[modem7]

Easiest way to check if the PSU is good or not: connect some other spare HDD to it (and ONLY that). This way, the PSU has enough load ...

Depending on the PSU you may need to load it using more than just one 'modern' HDD, Try it with 2 or 3 modern HDD's connected, ...

The PSU in the 5160 does not need a huge load. That was only the case with the PSU of the 5170.

No.

For example, the power supply that IBM provided in the IBM 5155, an XT-class computer) requires approximately four to five IDE drives.

As Malc wrote, the load requirements comes down to the particular make-model of power supply, i.e. the particular circuitry in use. We know that in the lifetime of the IBM 5160, that IBM used different power supplies.

 

[modem7]

Do I suspect that it is the power supply? Yes.
Do I know that it is the power supply? No.

Before expending any of my time on a 'suspect' power supply, I will first prove that the power supply is faulty.

( Yes, I have the luxury of having spare power supplies, motherboards, drives, etc. )

 

[modem7]

Also, what he measured is the same as it was with the whole PC connected, so that's all the PSU delivers.

But there are variables that have not changed. An example is the motherboard.
Yes, it would be very rare, but imagine the combination of:
- A power supply that does not de-assert its POWER GOOD signal when that is expected. (Which appears to be the case here.)
- A capacitor on the motherboard that has failed in the failure mode of: extremely leaky.
- The current draw from the leaky capacitor is such that it slightly overloads the power supply, but not enough for the power supply to shut down.

( People do win significant sums in the lotteries. )

 

[Hugo Holden]

If the power good circuit is anything like the one in the IBM5155 psu, you cannot place much stock in it as it is not specific as the previous graph presented might suggest.

Have a look at the last page of this article (circuit analysis of the 5155 psu), last sentence, and the remarks I made about the power good line:

https://www.worldphaco.com/uploads/The_IBM_5155_POWER_SUPPLY.pdf

The 5155 psu has some fairly elaborate under & over voltage and over-current detection circuitry, but this is not linked with the power good signal at all.

I'm not sure how the 5155 supply differs from the 5160 supply, because I have not investigated and performed a circuit analysis on the 5160 supply, but the 5155 supply is a very good starting point to understand computer SMPS and help with their repairs.

 

[Chuck(G)]

Most SMPS use an electrolytic capacitor in the primary circuit before the oscillator. When this goes empty-can, the output drops substantially. I recall doing some troubleshooting on a couple of Bob Boschert's swtichers. In both cases, the cap had cooked out.

 

[hozone]

Thanks all.
A few points that may help:
1) my PSU is like this one inside: https://www.vogons.org/viewtopic.php?t=73543
2) I have an electronic load, I can load one line at time, like +12V or +5V if needed
3) I'm searching for a spare, but it's not so easy to find at a reasonable price here in Italy
4) Below the reason why I write here before starting the repair of a switching, that's no so easy for my

- A power supply that does not de-assert its POWER GOOD signal when that is expected. (Which appears to be the case here.)
- A capacitor on the motherboard that has failed in the failure mode of: extremely leaky.
- The current draw from the leaky capacitor is such that it slightly overloads the power supply, but not enough for the power supply to shut down.

5)

https://www.worldphaco.com/uploads/The_IBM_5155_POWER_SUPPLY.pdf

Any change to get a document like this for my Schrack PSU?

 

[modem7]

I have a maybe one AT, but I've to check. Does the connector are the same?

Yes, they are the same. When connecting, make sure the black wires meet in the center.

You could not find that AT-class power supply?

( And for ATX power supplies, an adapter can be purchased - see [here]. )

4) Below the reason why I write here before starting the repair of a switching, that's no so easy for my

Assuming that the power supply is faulty, then based on the symptom, it is is as Chuck stated: The main filter capacitors at the front end of the circuit are certainly under suspicion.
In the circuit diagram at [here], which is for a particular make-model of power supply used by IBM in the 5160, they are the four large 440 µF capacitors (C1/C2/C3/C4) (of type: aluminium electrolytic).

Be sure to read the safety information at [here], especially the bit about discharge of the subject filter capacitors.