HP Power Supply help

[pentium]

I've been trying to revive the power supply from an *ANCIENT* HP 9845 computer for almost a year now. I borrowed an identical power supply from a friend that was known to work. The PSU's are broken into five boards. I can confirm through swapping with the known good power supply and its good boards that my power supply will work with all but one of my boards.






HP's schematics refer to this as the "Power Supply Low Voltage Board".
The schematics for the power supply can be found *here* between pages 37 and 45.
(NOTE: The schematics show a different revision of board from mine, so the layout is different and some components are not present)


My loaned and working card and my own but defective card are identical in layout and component values. With the defective board installed voltage regulator U6 on the "Power Supply Mother Board" (page 39) overheats and won't do a voltage above 2v as if the rail is either heavily loaded or shorted.
The following components (on the RIGHT side of the component-side photograph above) have been swapped between boards with no change in the fault (IE: the fault remains on MY card and doesn't move to the loaned board):

5 Volt regulator
Optoisolator
All five IC's
All three transistors

This leaves me with four diodes, sixteen resistors, two axial capacitors, five decoupling capacitors and two larger polyester (?) capacitors unverified and no obvious board damage, component damage or clear shorts to ground according to a DMM and this is a two layer board. I have never gone this far down a rabbit hole. Please by god help me.

[budm]

Re: HP Power Supply help

So if you check the resistance on the output side of the regulator of the good board and the bad board, do they have the same resistance?

[pentium]

Re: HP Power Supply help

Both swapped and tested out of circuit. They're fine.

[budm]

Re: HP Power Supply help

The load side of both board have same resistance reading then? The bad one must have heavier load than the good board, or some of the loads may have to be switched on first so resistance reading just on the output of the regulator may not show the problem, you will have to follow all the loads (switched) to see if any one of them show bad compared to the good board. I will have to look at the SCH to see how many loads are connected to the output of the regulator, or I wonder if the output is Oscillating.

[budm]

Re: HP Power Supply help

Both swapped and tested out of circuit. They're fine.

Are you talking about the regulator? I am talking about the loads to the regulator.

[pentium]

Re: HP Power Supply help

Both, actually.
U6 and the regulator on the Low Voltage board were pulled and tested both with and without a load. They operated to spec fine.
I also looked at the resistance of the output side of both and could not see anything that would put an excessive load on it. There's literally *no reason* I can find that explains what is causing this.

[budm]

Re: HP Power Supply help

Both, actually.
U6 and the regulator on the Low Voltage board were pulled and tested both with and without a load. They operated to spec fine.

Well, it is the bad loads on the bad board that are causing over current draw, that is you need to trace out the loads (switched and unswitched).
Load dictates how much current will flow in the circuit.

[pentium]

Re: HP Power Supply help

I've powered the board from a bench supply from pin P5R-A. There is no excessive load. It makes no sense.

[budm]

Re: HP Power Supply help

Is this the page 39 you are referring to?
Did you look at the output with the scope?

[pentium]

Re: HP Power Supply help

Yes, also the output seemed fine for a regulated DC voltage.

[budm]

Re: HP Power Supply help

Since you are getting only 2V output it is obviously not regulating any more.
What is the input Voltage to the regulator when you are seeing such the low output?
The output is also connected to L2 for the stacking power supply generated by T1 which I have no idea how T1 is powered.
How bad is the ripple showing on the scope?
You can also remove fuse F2 and feed the input of the regulator with external power supply setting to the same DCV input as the good board and see how much current you are drawing and see how the output behaves.

[pentium]

Re: HP Power Supply help

If I remember the input of U6 was something like 24v or 12 or...something that was supposedly in spec.I don't remember off the top of my head. Again, no ripple.
We can rule out anything at U6 or before because I throw a known good card in and everything works. We are not dealing with ripple. We are dealing with a fault on the card.

Also the reason U6 drops to 2v is because the regulator is kicking its thermal protection in because it's acting as if it's grossly overloaded and overheating.
Yet if we check, we find there is no load issues.

[CapLeaker]

Re: HP Power Supply help

Lift a leg off all them little diodes and check those. Same for the axial caps. I have the same opinion on the PSU not regulating. Btw... did you replace LM311 IC? Or better mark on the pic what exactly you have replaced.

[budm]

Re: HP Power Supply help

If I remember the input of U6 was something like 24v or 12 or...something that was supposedly in spec.I don't remember off the top of my head. Again, no ripple.
We can rule out anything at U6 or before because I throw a known good card in and everything works. We are not dealing with ripple. We are dealing with a fault on the card.

Also the reason U6 drops to 2v is because the regulator is kicking its thermal protection in because it's acting as if it's grossly overloaded and overheating.
Yet if we check, we find there is no load issues.

Well I do not know what else to tell you, no load = no current flow, too much load = too much current flow. May be I know nothing.
Look at how L2 is connected to the output of the regulator.

[Longbow]

Re: HP Power Supply help

I would like to have more specific information about the item you are working on. What is the exact model number? What is the HP part number of the faulty pcb? What is the HP board designation - i.e. there is a U6 IC on at least 2 of the power supply boards. Which one? On one LV board U6 is an 74LS122 and on another U6 is a 3 terminal reguator.

These schematics are production only. Certainly not the complete manuals that HP is famous for. This power supply drawing should be available somewhere showing voltages...

Are you aware of the many sites that are dedicated to restoring 9845's? Such as http://www.hp9845.net/9845/ and this one:
http://www.hpmuseum.net/display_item.php?hw=149 This supply should not be a big deal to fix as long as you have a bit more information availiable.

[Longbow]

Re: HP Power Supply help

Please show a closeup of the board and circle U6. Take voltage readings all over U6 input and output as well as around T1 with a "working" board. Then take the same readings with your own board.

[Longbow]

Re: HP Power Supply help

If I remember the input of U6 was something like 24v or 12 or...something that was supposedly in spec.I don't remember off the top of my head.

We'll need a bit more detail than that!!

We can rule out anything at U6 or before because I throw a known good card in and everything works. We are not dealing with ripple. We are dealing with a fault on the card.

I think you are rushing to judgement. You pointed out that the schematics aren't exactly right and that your comparison board is different. It may seem that the board is the problem but we have some interesting stuff going on with T1. I think this is an early form of overvoltage protection but you can't see the whole circuit on this one diagram. I'll bet your friend's board is a newer revision (HP is famous for making multiple revised pcb's for the same unit.) Right now, I think the most accurate statement is that your friend's board does not make U6 get hot. Period. That is just one fact amongst others.

It was typical for HP to make a revised board available for retrofit, if a common problem was found. The new board would come along with instructions on what other parts in the unit needed to be changed as a result of installing the revised board. Slow and careful is the word.

[Drack]

Re: HP Power Supply help

I would check that silver part with 2 screws( Don't know the name on English) that look like it may be the culprit those are something like a thermal fuse or thermal something

[Stefan Payne]

Re: HP Power Supply help

You mean the Power Transistor? (or maybe a FET)

[budm]

Re: HP Power Supply help

I sure like to see the pictures of the board that he is referring to the SCH on page 39.