Have you checked mosfets voltages? One should generate the +2.0V required by the cpu (or the V needed).
With no cpu does it power on?
No +12V on fans can be a dead chip driving the fans: check traces from +12V to one of the chips onboard, +12 doesn't come straight from the PSU. Also "sense" line.
Once found look up for the corresponding datasheet and check.

Check I/O chip https://www.alldatasheet.com/datasheet-pdf/pd … W83977TF-A.html

Could also be a dead cap on the PSU rails, checked resistance?
It's gonna be boring as hell and up to months but hang in there.

Checked JP9?

What does "No fan, no POST card lights, nothing" mean, exactly? POST cards have LEDs for power rails, these also do not light up? Is the ATX PSU starting at all?
I would assume, since the 3V3 lines seem to be N/C and the mobo does have AT power connector too which lacks 3V3, that this voltage is generated by the mobo and these pins are not used. So my next step would be to use AT PSU instead of ATX to see if the problem is perhaps with the PSU itself. If the chipset does not start the PSU, other than simple wiring failure between power switch and mobo, would suggest some chipset issue.

Nexxen wrote on 2023-03-21, 15:17:

Have you checked mosfets voltages? One should generate the +2.0V required by the cpu (or the V needed). With no cpu does it powe […]
Show full quote

Have you checked mosfets voltages? One should generate the +2.0V required by the cpu (or the V needed).
With no cpu does it power on?
No +12V on fans can be a dead chip driving the fans: check traces from +12V to one of the chips onboard, +12 doesn't come straight from the PSU. Also "sense" line.
Once found look up for the corresponding datasheet and check.

Check I/O chip https://www.alldatasheet.com/datasheet-pdf/pd … W83977TF-A.html

Could also be a dead cap on the PSU rails, checked resistance?
It's gonna be boring as hell and up to months but hang in there.

Checked JP9?

Thanks for your reply, Nexxen! I have not yet checked the mosfet voltages. I assume there won't be any current as I don't think the power supply is delivering anything but +5V standby. It seems like the board does not ask the PSU to turn on. I will work on checking out the Winbond IO chip as well, thanks for the datasheet link.

JP9 toggles startup from keyboard or mouse, I tried it in both positions but no change unfortunately.

Would the PSU rail caps be the 6 big caps behind the slot 1 connector on the motherboard? I believe the PSU itself is okay, it tested okay externally, but I have another PSU I can try as well. I traced one +12V rail going down through the AGP, PCI, and ISA slots along the back of the board, so at least that one isn't going through any caps I can see.

Deunan wrote on 2023-03-21, 16:18:

What does "No fan, no POST card lights, nothing" mean, exactly? POST cards have LEDs for power rails, these also do not light up? Is the ATX PSU starting at all?
I would assume, since the 3V3 lines seem to be N/C and the mobo does have AT power connector too which lacks 3V3, that this voltage is generated by the mobo and these pins are not used. So my next step would be to use AT PSU instead of ATX to see if the problem is perhaps with the PSU itself. If the chipset does not start the PSU, other than simple wiring failure between power switch and mobo, would suggest some chipset issue.

Thanks Deunan. Indeed the POST card does not light up on any rail, it's as if pressing the power switch has no effect at all. The PSU has a fan which should briefly spin and it does not.

Thanks for confirming what I thought about the +3.3V, it seems not all boards use it directly from the PSU even though it's available. I have another working ATX PSU that's wired up with an AT adapter, so I will give that a try to see if there's any difference.

Update: ATX pin 14 "PS_ON#" is supposed to turn on the PSU when it's shorted to ground. Without the PSU plugged in I confirmed it does not short to ground when I press the power button. Instead I can see this pin routing over to the Winbond IO chip before it disappears into a via, so I assume the Winbond chip handles the actual PSU power on function.

Update: The Winbond chips accepts +5V on pins 15 and 56 and ground on 25, 40, 65, and 90. I'll check these.

Can you measure with a multimeter in 'diode/continuity' mode on the mb without anything connected (not even the cmos battery) atx pin 'power ok'(power good) , atx pin 5v SB(standby) and atx 5v (main rail) related to ground in both ways (black lead to ground/red lead to ground) ?

3.3V is normal, AT/ATX combo boards generate it themselves
fan header might be speed controlled, wont have direct connection to 12V
ATX pin 14 "PS_ON#" is controlled by Winbond as you traced

https://pdf1.alldatasheet.com/datasheet-pdf/v … 83977EF-AW.html on first glance this chip just handles buffering/latching power button signal and has no voltage monitoring facilities, doesnt even check powergood.

Unplug everything form the motherboard except the supply, stick multimeter probe into ATX connector pin 14 "PS_ON#", set multimeter to voltage mode, press button and see if reading changes.

There are two possibilities
1 W83977EF doesnt get power or is somehow broken
2 there is a short on the board/excessive power draw and ATX supply detects that not letting itself power on

Have you tried a AT PSU ? It seems those dual AT- ATX boards failing come up often more lately.
Have had a few die recently with no visible damage or explanation other than one still worked OK from AT PSU but not from ATX. Sorry cannot help other than that..

It's alive! Using an ATX to AT adapter worked. Thanks everybody for the ideas and suggestions.

I still don't know why the ATX adapter worked then stopped working seemingly for no reason, but I planned to fit it into an AT case anyway, so it works out.

bogdanpaulb wrote on 2023-03-21, 22:03:

Can you measure with a multimeter in 'diode/continuity' mode on the mb without anything connected (not even the cmos battery) atx pin 'power ok'(power good) , atx pin 5v SB(standby) and atx 5v (main rail) related to ground in both ways (black lead to ground/red lead to ground) ?

This was an interesting troubleshooting step I hadn't tried before. Here are the results of the diode tests:

• ATX pin 8 (power good) - black to ground = open loop, red to ground = no beep 1.2V
• ATX pin 9 (standby) - black to ground = no beep 2.2V, red to ground = beep 0.4V
• ATX pin 5 (+5VDC) - black to ground = beep 0.5V, red to ground = beep 0.3V

rasz_pl wrote on 2023-03-22, 01:21:

3.3V is normal, AT/ATX combo boards generate it themselves fan header might be speed controlled, wont have direct connection to […]
Show full quote

3.3V is normal, AT/ATX combo boards generate it themselves
fan header might be speed controlled, wont have direct connection to 12V
ATX pin 14 "PS_ON#" is controlled by Winbond as you traced

https://pdf1.alldatasheet.com/datasheet-pdf/v … 83977EF-AW.html on first glance this chip just handles buffering/latching power button signal and has no voltage monitoring facilities, doesnt even check powergood.

Unplug everything form the motherboard except the supply, stick multimeter probe into ATX connector pin 14 "PS_ON#", set multimeter to voltage mode, press button and see if reading changes.

There are two possibilities
1 W83977EF doesnt get power or is somehow broken
2 there is a short on the board/excessive power draw and ATX supply detects that not letting itself power on

I measured open loop between PS_ON# and ground with the PSU connected and pressing the power switch. I accidentally bumped the leads together and shorted PS_ON# to ground (ask PSU to turn on) and heard the PSU click, but nothing else happened, granted there was also no CPU or RAM. I wonder if manually turning on an ATX PSU would still work, but I'm afraid to test now that it's working!

I also checked all of the VCC, VSB, and ground pins on the W83977EF IO chip. It was challenging with how small the pins on the PQFP package are, but continuity tests to these pins passed / no shorts.

Horun wrote on 2023-03-22, 01:43:

Have you tried a AT PSU ? It seems those dual AT- ATX boards failing come up often more lately.
Have had a few die recently with no visible damage or explanation other than one still worked OK from AT PSU but not from ATX. Sorry cannot help other than that..

Thanks for the suggestion. I wonder if it's something to do with modern PSUs on old 20-pin ATX connectors. I have another combo board I used the same PSU on that also doesn't start, maybe I need to try the AT connector there too.

Here is the board finally restarted. I like to think the POST card is two segments away from saying YE-HA.

And here it is in case running 3DMark2000:

Cheers everyone for the help!

Cosmic wrote on 2023-03-29, 05:52:

It's alive! Using an ATX to AT adapter worked. Thanks everybody for the ideas and suggestions. […]
Show full quote

It's alive! Using an ATX to AT adapter worked. Thanks everybody for the ideas and suggestions.

I still don't know why the ATX adapter worked then stopped working seemingly for no reason, but I planned to fit it into an AT case anyway, so it works out.

bogdanpaulb wrote on 2023-03-21, 22:03:

Can you measure with a multimeter in 'diode/continuity' mode on the mb without anything connected (not even the cmos battery) atx pin 'power ok'(power good) , atx pin 5v SB(standby) and atx 5v (main rail) related to ground in both ways (black lead to ground/red lead to ground) ?

This was an interesting troubleshooting step I hadn't tried before. Here are the results of the diode tests:

• ATX pin 8 (power good) - black to ground = open loop, red to ground = no beep 1.2V
• ATX pin 9 (standby) - black to ground = no beep 2.2V, red to ground = beep 0.4V
• ATX pin 5 (+5VDC) - black to ground = beep 0.5V, red to ground = beep 0.3V

rasz_pl wrote on 2023-03-22, 01:21:

3.3V is normal, AT/ATX combo boards generate it themselves fan header might be speed controlled, wont have direct connection to […]
Show full quote

3.3V is normal, AT/ATX combo boards generate it themselves
fan header might be speed controlled, wont have direct connection to 12V
ATX pin 14 "PS_ON#" is controlled by Winbond as you traced

https://pdf1.alldatasheet.com/datasheet-pdf/v … 83977EF-AW.html on first glance this chip just handles buffering/latching power button signal and has no voltage monitoring facilities, doesnt even check powergood.

Unplug everything form the motherboard except the supply, stick multimeter probe into ATX connector pin 14 "PS_ON#", set multimeter to voltage mode, press button and see if reading changes.

There are two possibilities
1 W83977EF doesnt get power or is somehow broken
2 there is a short on the board/excessive power draw and ATX supply detects that not letting itself power on

I measured open loop between PS_ON# and ground with the PSU connected and pressing the power switch. I accidentally bumped the leads together and shorted PS_ON# to ground (ask PSU to turn on) and heard the PSU click, but nothing else happened, granted there was also no CPU or RAM. I wonder if manually turning on an ATX PSU would still work, but I'm afraid to test now that it's working!

I also checked all of the VCC, VSB, and ground pins on the W83977EF IO chip. It was challenging with how small the pins on the PQFP package are, but continuity tests to these pins passed / no shorts.

Horun wrote on 2023-03-22, 01:43:

Have you tried a AT PSU ? It seems those dual AT- ATX boards failing come up often more lately.
Have had a few die recently with no visible damage or explanation other than one still worked OK from AT PSU but not from ATX. Sorry cannot help other than that..

Thanks for the suggestion. I wonder if it's something to do with modern PSUs on old 20-pin ATX connectors. I have another combo board I used the same PSU on that also doesn't start, maybe I need to try the AT connector there too.

Here is the board finally restarted. I like to think the POST card is two segments away from saying YE-HA.

And here it is in case running 3DMark2000:

Cheers everyone for the help!

ATX pin 9 (standby) - black to ground = no beep 2.2, red to ground = beep 0.4 seems to be on the off side, but maybe is just the design of the board.
Can you enable power on via keyboard from the bios and test it with ATX psu plunged in directly (jp9 should be placed in pos 2-3 for this)? There should be no chances of breaking something in doing so.

I remember PC Partner board with AT/ATX and i440LX, where was jumper for setting, which PSU is used (seems it is not presented on your board) and even in BIOS setup was option for AT/ATX selection. But must say, board work with ATX PSU even with option AT in BIOS, as this was default settings.