Currently as of now these days from long searches, there is currently no available PSU for compaq deskpro M series. I need that pinout so others can build adapter to use ATX PSU on it. If I have a PSU for deskpro M series, this will be a golden time even loaned one in semi-working, broken state or working state is very essential as what I see, there is pretty large number of Deskpro M owners with their power supply potentially bound to fail one day and to make adapter just for this will be much nice. I'm willing to fix and document Deskpro M power supply.

Notes:
The Deskpro M power supply is very special design and I need to document this very carefully:
I suspected that may be a signal loop-back so PSU knows is complete circuit is made from the both slot board and I/O board is connected together to power up. Conjunctive theory. I verified and this need to be documented fully since that because -5V path goes through I/O board from PSU connector first, then back out to slot board to the EISA slots pin for -5V, I had to plug both EISA slot board and i/o board together to get -5V circuitous path made complete!). How I know, I recently bought a set of boards to make a working Deskpro M computer on bench but without PSU, I'm stumped on these unknown 3 pins left to document. Oh, there is no LED connection since PSU is not butting against front panel like my Deskpro 386/25e motherboard is (PSU pinout already documented except two (easy!), just matter of guessing which one is power good pin therefore know other pin LED output through PSU's board to illuminate hard drive activity LED), 386/25e PSU is full length PSU front to back with two LEDs in the PSU for power and hard drive activity.

Second, one of unknown pin has to have a "power good" signal that is held low (forced low held the reset line in the computer's board, then pulled up to 5V allows computer to boot. This is still present to the current time in all computers to the day even the Intel new 2020 ATX 10pin 12V design. That was IBM idea first started with PC 5150, the circuit to signal power good comes from internal PSU monitoring 4 voltages 5 and 12 both minus and positive are all correct then few milliseconds later then power good is enabled, thus releasing reset circuits allowing motherboard to run.

Cheers,

Verified on rear left pin is N/C next to pin 1, with light shining through the slot board PCB to see if the via is connected carefully looked for buried trace at a via, not there either compared to other locations of buried trace at any via. The photos of PSU I found so far on VCFed for Deskpro M shows previously said 3 pins are connected on the PSU board itself cast more doubts.

PS: via means drilled hole with conductive metal plated through so tracks can make through connection on both sides or to make a track connection at a soldered via.

Here's the semi-complete pinout for Deskpro M line series (386DX, 486 and early Pentium socket 60/66)

This is 24 position connectors with last two positions has metal pins not present, facing front. Pin 1 is towards rear of chassis as "o".
There are 3 unknown signal or two signals and one power on rear most pins, marked "*".
N/C means no connect.

This is for Deskpro M series PSU pinout measured on the motherboard's.

*N/C (?) X 0 pin 1 * (voltage sense or something?)
-5V X X (?) *
GND X X -12V
GND X X GND
GND X X GND
GND X X GND
+5V X X +5V
+5V X X +5V
+5V X X +5V
+5V X X +12V
N/C X X +12V
N/C X X N/C

Quick add on:
Deskpro 486/33L has CPU board while 486/50L has CPU on planar.
Some L and Systempro CPU board are single use only and cannot be used for Systempro dual CPU, while 386/486 combo is supported.

I probably can provide info on the Systempro and Deskpro N due to complete system reference manuals.
Got some more Compaqs that can be added to the PS-pinout database, but mostly Prosignia and Proliant servers.
Deskpro 286, 386/n, 386/s and 386/e are hiding somewhere in my basement too.
I will try to add these to the vogonswiki at some time.

I have two turn-of-the-millennium Compaq boards with non-standard PSU connectors. One is an i810 board with 24p ATX-like connector , but definitely not ATX pinout, the other is an i840 board (SW400, from an AP550 workstation) with the same 24p connector (seemingly same pinout, but haven't verified that yet) and an additional 8p aux connector. Tried to find pinout online when I got the SW400, failed and just bought the Compaq PSU. It's a bit later than your stated period, but I could work out the pinout if desired to document that too.

I am recapping my M power supply; I’ve ordered replacement capacitors and am slowly making my way through removing the existing ones. It’s extremely slow going, I’ve only been able to pull 5 of them despite generous amounts of flux and solder wicking. These things were well built, that’s for sure.

UPDATE! Confirmed working great using adapter rewired to Deskpro M pinout PSU using a ATX power supply.

This is for Deskpro M series PSU pinout measured on the motherboard's.

*N/C X X pin 1 (one), still unknown purpose and has connection, no voltage.
-5V X X <---------Power Good signal.
GND X X -12V
GND X X GND
GND X X GND
GND X X GND
+5V X X +5V
+5V X X +5V
+5V X X +5V
+5V X X +12V
N/C X X +12V
N/C X X N/C

Greetings,

I got excited when I stumbled upon a *recent* (or any, for that matter) post regarding one of these old proprietary Compaq power supplies.

I'm at the point of trying to determine 'power good' signal and any other gotchas. I thought perhaps what I have so far could be beneficial to someone else, or perhaps someone else might chime in with missing piece of info. PC is Compaq Deskpro 50M. Power supply uses 24 pin molex atx-style connector as mentioned by previous poster. Skinscreen on PSU PCB "BOARD NO. 002021-001 A/W REV C, DIAGRAM NO. 002020, SPARE NO. 129069-001" and "ASSY. NO 002019-002"

The other avenue I explored is repairing the original PSU, I was able to identify values for all the electrolytic. Unfortunately, C22 and R29 were badly damaged and I was unable to verify their values with certainty. If someone could verify those values, that would be very really helpful. Hoping what I documented can be of use to someone going down similar path.

DISCLAIMER: proposed pin-out is derived from dead PSU, so connections are based off of continuity checks, reference to ground and tracing PCB traces, etc. Please don't let the smoke out of your nostalgic precious... and if you do, it's not my fault, you've been warned. 😉

Couple of observations: On PSU side, pin 3 goes to 10k resistor that connects to gate pin of 2N7000 mosfet. Pin 2's trace dead ends on another to-92 via that is not populated on this board.

Actually there is connection on pin 1, please edit your article, as I have documented this and offered a theory on what pin 1 is for.

Once you get the PSU fixed, please measure the pin 1 when on with fan running, and you can block the fan from turning then turn on again and measure pin 1 again, using meter set to voltage.

Please show us the photo of the unidentified capacitor?

Cheers,

Thanks for the prompt response! I re-read you previous post and I think I follow. I did find on the PSU side, there is a 'sneaky' trace on connector side, partially obscured by the 24 pin connector, that ties pin 1 directly to the 5v rail.

Unfortunately components were pretty badly damaged. I realize there is damage to PCB and traces themselves, but believe it or not, I've repaired worse.

Here's notes on electrolytics proposed to replace..

Capacitors uF voltage height diam
C3 68 35 15.6 6.3
C8 820 250 51 30
C9 820 250 51 30
C12 470 25 20 10
C24 68 35 15.6 6.3
C29 1000 25 26 14
C35 100 35 10 15
C36 4700 6.3 33 13
C37 4700 6.3 33 13
C41 470 25 22 10

C22, based on size and other values used on board, best guess, might be 47nF?

R29 looks carbon comp, maybe 2W.. first band is red (2), last band is gold (5%), hard to say beyond that.

So far, that's good of your work, keep doing that with few pointers:

That burnt resistor also the burnt capacitor area, are all in simple circuit area, I recommend draw a partial circuit around that to make sense, which is easier since you are in small area near hot side AC to DC rectifier and filtering circuit which converts AC to DC at about 300VDC through large capacitor before feeding the chopper transistor that in turn drives the power supply main transformer.

Sure you are spot on on 5V is directly tied to this pin 1? Checked using ohms with multi-meter? Also please take good photos on both sides of heatsink of this pin 1 trace so I can retrace this one.
Remember, I do not have original PSU. They do not exist anymore from long search. Ones that still works are in working Deskpro M computers at too much price and don't want to buy another computer.

Cheers,

pentiumspeed wrote on 2020-08-12, 00:53:

So far, that's good of your work, keep doing that with few pointers: […]
Show full quote

So far, that's good of your work, keep doing that with few pointers:

That burnt resistor also the burnt capacitor area, are all in simple circuit area, I recommend draw a partial circuit around that to make sense, which is easier since you are in small area near hot side AC to DC rectifier and filtering circuit which converts AC to DC at about 300VDC through large capacitor before feeding the chopper transistor that in turn drives the power supply main transformer.

Sure you are spot on on 5V is directly tied to this pin 1? Checked using ohms with multi-meter? Also please take good photos on both sides of heatsink of this pin 1 trace so I can retrace this one.
Remember, I do not have original PSU. They do not exist anymore from long search. Ones that still works are in working Deskpro M computers at too much price and don't want to buy another computer.

Cheers,

Sorry for delayed response. To follow up, please see attached photos. It proved difficult to get a good photo of pcb being candled, highlighted candled traces. Trace on connector side follows solder side trace once out from under connector. Good suggestion to draw up partial schematic, though, having carefully pinout out atx power supply to retrofit, I fear psu might have taken motherboard with it when it failed. Board does not post, no LED light.

A couple more photos... Also, to your question, yes. Confirmed several times, measuring continuity between pin 1 and 5v rail on PSU connector measures 0.3 ohms (which is the same measurement I get when shorting meter leads together.)

Use my diagram to make a adapter using a ATX extension cable and a connector extractor tool. -5v can wait, as I tested my Deskpro M with Seasonic Prime 650W.

Still make a partial schematic and show us around this burnt components including the AC in circuit as I can tell it is part of this.

Cheers,

Looking for PSU pinout for Compaq board 010546-101, is a Slot 1 motherboard with Intel 810(E?) chipset.
Found more info about it in the Compaq Deskpro EP Maintenance & Service Guide an seems to be used in a Compaq Deskpo EP 6500.
The service guide mentioned it should be used with a 200W power supply that has part number: 103748-001
I've tried a normal ATX psu and even a 20pin psu from a Deskpro EP 6400 but can't get it starting, power led onboard is lit when a psu is connected.
Hope someone at Vogons has more knowledge about this motherboard/psu.