You could take a multi-meter to the cables, check the connection voltage, and wire your own connection to it, remember to save the connectors so you can still use them,

Couldn't that be just the capacitors getting old ?
My 8088 became unstable after a few minutes running, and I had to wait for a while before it stops crashing. Changed the capacitors : rock stable. I ran it for hours without any problems. Same thing with my pcchips 486 (no it was not due to the poor quality motherboard 🤣). After recapping, no more problems.

I'm just giving my point of view, of course you can change it if you want. In that case, ampera's piece of advice is what you should follow

Deksor wrote:

Couldn't that be just the capacitors getting old ?
My 8088 became unstable after a few minutes running, and I had to wait for a while before it stops crashing. Changed the capacitors : rock stable. I ran it for hours without any problems. Same thing with my pcchips 486 (no it was not due to the poor quality motherboard 🤣). After recapping, no more problems.

I'm just giving my point of view, of course you can change it if you want. In that case, ampera's piece of advice is what you should follow

With an AT PSU it's probably easier to replace it than replace the caps.

If they're using standard color codes, I'd say the red wire is 5+, black is ground, and grey is probably PWR OK (Like an ATX power supply).

[EDIT]

That screw and tinning look rusty.

NEVER trust colours. It's a lot easier and cheaper to just test it than to be wrong and have a boned machine.

It's even cheaper to go out and BUY a dinky little multimeter and measure it than it is to loose an IBM PS/1. 120USD is more than 20 bucks.

You don't need a Fluke or anything, but if you're working with retro hardware, it's a great idea to invest in a decent multimeter, ideally not from Harbour Freight.
https://www.youtube.com/watch?v=ZoeUgMFLyAw
https://www.youtube.com/watch?v=n3WGaiYF2sk

Thanks all for the thoughtful responses!

Looks like I'll be investing in a multimeter and keeping you posted.

Also, I think I have a lead on a possible replacement PSU with the proprietary connector. Will see!

James

jmusic_man wrote:

Thanks all for the thoughtful responses! […]
Show full quote

Thanks all for the thoughtful responses!

Looks like I'll be investing in a multimeter and keeping you posted.

Also, I think I have a lead on a possible replacement PSU with the proprietary connector. Will see!

James

I'd personally trust an off the shelf modern PSU. AT Power supplies are still sold brand new and working with old tech on Newegg.

Ordered my multimeter.

I think I found a clue on PSU label: do you think 'Aux +5v 20mA' is the 3-pin connector in question?

jmusic_man wrote:

Ordered my multimeter.

I think I found a clue on PSU label: do you think 'Aux +5v 20mA' is the 3-pin connector in question?

As said, test, don't guess.

If it does supply 5v, just make a split off a molex or other 5v connector and into that connector.

DO NOT test current without a load. If you seriously want to test that, you are going to need to turn the machine on, strip a bit of insulation somewhere up the wire, and measure it that way. Otherwise you can blow the meter. I don't think 20mA will blow anything, but don't bother with current anyways. It's probably not that much.

jmusic_man wrote:

Looks like I'll be investing in a multimeter and keeping you posted.

Hi James, did you ever find out which wire was the standby +5v? I have the same motherboard in my PS/1 and the power supply has gone dead. I want to wire up an ATX supply but this particular connector has me stumped. I'm afraid of getting these wires mixed up and frying the board.

I know this is an old post, but my PSU just died, and I was having the same problem. Black (left) is ground, Violet (center) is on/off, Yellow (right) is 5V standby (always on) power. I am currently making an adapter cable for a standard ATX > PS/1 (AT based) connector.

Sorry, further to the above. Mine has different colors. So in the original post the J30 connector's black is ground, grey is on/off, red is +5VSB. J29 is the on/off switch at the front which would pass the signal to grey.

PixelOCDGuy wrote on 2022-12-30, 02:24:

Sorry, further to the above. Mine has different colors. So in the original post the J30 connector's black is ground, grey is on/off, red is +5VSB. J29 is the on/off switch at the front which would pass the signal to grey.

You are a super star man 😀 I have been scouting the Internet for this info for days!!

So basically on mine (same motherboard) the switch circuit is trashed. If you try to switch it off with the stock power supply or a new one it just resets over and over very fast making a bad clicking noise. Even with a new ATX power supply wired (just tried it ) same thing .. basically verifying the problem is on the motherboard and not the supply.

This could be bypass caps (at some point I'll get around to recapping) but I actually doubt it because all the caps that I've tested in this board have normal esr and capacitance readings. This is actually a problem because if you apply power in an "off" state the system firmware for some reason thinks you have a laptop of some sort ( oddly ) reveling a bunch of options inside the configuration utilities on page 3 and page one that look like corruption. If you place a jumper between these pins and bypass the switch (either by using a mechanical switch or my soft power switcher project https://github.com/lindoran/ps1pwfake) the computer boots up realizes the nv ram corruption and everything just works. I did go In and turn off apm completely on page three as well :

It does appear that the nvram stays correct even after you remove the jumper if your not comfortable leaving it in there forever like I did.

Wish we had schematics for the switch circuit and the apm stuff of the main board but heh - ibm