Compaq PDP-124P - loud screeching when running

[kotel studios]

Hi,

Recently I got my compaq PDP-124P to "work" and it now screeches loudly when running (you can hear it from 60cm away!).
There's nothing shorted on it, caps were already checked and replaced when they were close to being above 20% of rated spec, apart from the primaries. Fan isn't the issue too.

I didn't have time to measure the voltage rails (beside 12V, that one had 11.7V) due to the unexpected sound.
Load was only an single sacrificial board w/o CPU power since these PSU's don't work without mainboards.

Any ideas what could be the issue?

[kotel studios]

Actually, after removing all of the transistors I noticed the dip8 component on the primary side had two legs shorted. Odd, don't think I've done that while desoldering the transistors and whatnot...
And now after putting everything back together I get no life at all from this PSU. Not even the bulb protection. I guess the 230vac wires got screwed.

[stj]

what are you testing caps with?
are you checking ESR & Leakage?

[PeteS in CA]

This may seem odd, but since the P/S is probably more than 20 years old, there is the possibility that:

* Output caps you replaced now have capacitors that are "too good", their impedance is too low, causing instability;

* Output caps you did not replace are high impedance, causing instability.

[kotel studios]

This may seem odd, but since the P/S is probably more than 20 years old, there is the possibility that:

* Output caps you replaced now have capacitors that are "too good", their impedance is too low, causing instability;

* Output caps you did not replace are high impedance, causing instability.

I don't think it's the first assumption since I used the same caps as with my LC clones, but who knows what hp did with their stupid engineering.

what are you testing caps with?
are you checking ESR & Leakage?

lcr t2 tester. Sure, it's nor that reliable but it gives me an rough example of what the real capacitance should be. ESR is always 0ohm (tester limitation) and V loss is either 0.2% or 0%.

[kotel studios]

Okay still same sound, even after replacing an 10uF 50v cap connecting to 5v rail (I didn't see that one of the legs was broken). Maybe even worse. This time I measured 5v rail and it was 5.14v
Another idea would be that the primary caps are cooked and need to be replaced, but the only 470uF 200v caps I have report as 320uF which is not normal (they are from an LC clone). I think for testing they should be enough.

Any ideas?

[kotel studios]

Another idea would be that the primary caps are cooked and need to be replaced, but the only 470uF 200v caps I have report as 320uF which is not normal (they are from an LC clone). I think for testing they should be enough.

Primaries were replaced but the sound is still there..... So this can't be the primary side. Gotta take a closer look at the secondary or better yet, put the original caps in.

[stj]

screaming usually means it's running at the wrong frequency - often caused by overloading or instability.
so i'm sure the problem is on the output side

[kotel studios]

screaming usually means it's running at the wrong frequency - often caused by overloading or instability.
so i'm sure the problem is on the output side

Overloading shouldn't be the issue. The board takes waaaay less than 180W the PSU is rated for.

Old caps in and the original fan and still the same issue. 12v is 11.84v.
I think the sound comes from near the transformator. If it had shorted windings then the PSU wouldn't start up right?
I will need to pull out the transistors near it and check them out of circuit (in circuit they are fine).

I am running out of ideas now. I should also replace all of the smaller caps on the secondary side, but who knows if that'll help.

[stj]

take something insulated like a bic pen without the refil and poke at the transformers and coils to see which is doing it

[kotel studios]

Actually, after removing all of the transistors I noticed the dip8 component on the primary side had two legs shorted. Odd, don't think I've done that while desoldering the transistors and whatnot...
And now after putting everything back together I get no life at all from this PSU. Not even the bulb protection. I guess the 230vac wires got screwed.

[kotel studios]

Yep, it was the short on the dip 8 component at fault. Or that and the out of spec 220uF 25v cap. Odd since there wasn't any short before on it.
Now the PSU works without any fuss with the original caps. I will do further testing to confirm whether this PSU is safe for normal use.

[stj]

what is the part number of the 8pin device?

[kotel studios]

UC3845B PWM chip. I think it was ground with output shorted. Gotta recheck once I get to finalization of this repair.

[kotel studios]

Okay, since everything seems to work fine I decided to rework the case a bit and guess what? The 4pin CPU power is proprietary. Instead of 12v or 12.4v high power it's 12.8v (no load is 13.20v while with an 12v 5w bulb it's 12.9v)...
I'm not gonna write about "downgrading" the rail here since this post was just to fix the screeching noise.

[stj]

it's probably "within spec"
atx is pretty lax with it's numbers

[kotel studios]

Okay under general load (as I would run it) the cpu 12.8v goes down to 12.66v and seems to be going slowly up. So ig it should do the trick.

Generally this whole repair could've been avoided if I cleaned this PSU and waited like 6 months before testing it without a sacrificial board.

[stj]

i always clean stuff before starting on it - makes everything a lot easier that way.