VOGONS


First post, by ziggy587

User metadata
Rank Newbie
Rank
Newbie

Hey, everyone!

So I'm having an issue with the soft power switch on a Socket 7 motherboard. This is actually the same board from another thread I posted, I had an issue with voltage at the PS/2 ports. It's a Biostar MB-8500TUD-A.

An image of the motherboard.

s-l1600.png

So this is one of those baby AT boards that was on the verge of ATX. This board has the P8/P9 AT power connector as well as a 20-pin ATX power connector, although my exact revision did not populate the ATX connector. I tested continuity between the unpopulated ATX connector socket and the P8/P9 connectors, and everything checked out. So I decided to install the 20-pin connector so I can use an ATX power supply without any adapters, as well as use an ATX case. (and built in power button). I've been testing the board out for a few days now using a P8/P9 to ATX adapter, and aside from the aforementioned PS/2 voltage issue, everything seems OK. I don't think adding the missing 20-pin connector is an issue, I don't see any other missing components on the board that seem like they would have something to do with the power-on signal, but I thought I should mention it just in case.

Now here's the issue: When I plug an ATX power supply into the newly installed 20-pin connector, as soon as I flip the switch on the back of the PSU the board powers itself on.

I don't think I fully understand how the soft power works on ATX boards. I know that when the power-on signal (the green wire) is pulled low, the PSU will turn on. But I thought it was connected directly to the momentary switch, and only needed to briefly be pulled low to turn the PSU on or off. That seems to be how the reset switch works, so I assumed the power switch worked the same way. But it appears that there must be a circuit between the power-on signal from the PSU and the soft power switch. Now I'm guessing that the power-on signal needs to be HELD low, and this is accomplished on the motherboard instead of in the PSU. So that would mean the momentary switch activates a circuit that latches the signal low. And then hitting the power switch again will unlatch the signal. Is this correct?

I tried searching for the ATX soft power circuit.... Is there a standard circuit that all boards use? I'm assuming there is an issue with this circuit on my motherboard. I tried to trace it out to find the components, but that proved to be extremely difficult.

I have thought of a work around. I know I could just ground the power-on wire via a SPST switch, but I'd rather use the case's built in power button. So I thought of this: I could splice out the power-on wire from my 20-pin connector, then build my own soft power circuit, and hook the case's power switch up to that. But I would like to investigate the problem if I can. For one, so I can be sure it isn't going to cause any further problems. But also for the learning experience of it.

Reply 1 of 19, by PC Hoarder Patrol

User metadata
Rank l33t
Rank
l33t
ziggy587 wrote on 2022-11-27, 00:56:
Hey, everyone! […]
Show full quote

Hey, everyone!

So I'm having an issue with the soft power switch on a Socket 7 motherboard. This is actually the same board from another thread I posted, I had an issue with voltage at the PS/2 ports. It's a Biostar MB-8500TUD-A.

An image of the motherboard.

s-l1600.png

So this is one of those baby AT boards that was on the verge of ATX. This board has the P8/P9 AT power connector as well as a 20-pin ATX power connector, although my exact revision did not populate the ATX connector. I tested continuity between the unpopulated ATX connector socket and the P8/P9 connectors, and everything checked out. So I decided to install the 20-pin connector so I can use an ATX power supply without any adapters, as well as use an ATX case. (and built in power button). I've been testing the board out for a few days now using a P8/P9 to ATX adapter, and aside from the aforementioned PS/2 voltage issue, everything seems OK. I don't think adding the missing 20-pin connector is an issue, I don't see any other missing components on the board that seem like they would have something to do with the power-on signal, but I thought I should mention it just in case.

Now here's the issue: When I plug an ATX power supply into the newly installed 20-pin connector, as soon as I flip the switch on the back of the PSU the board powers itself on.

I don't think I fully understand how the soft power works on ATX boards. I know that when the power-on signal (the green wire) is pulled low, the PSU will turn on. But I thought it was connected directly to the momentary switch, and only needed to briefly be pulled low to turn the PSU on or off. That seems to be how the reset switch works, so I assumed the power switch worked the same way. But it appears that there must be a circuit between the power-on signal from the PSU and the soft power switch. Now I'm guessing that the power-on signal needs to be HELD low, and this is accomplished on the motherboard instead of in the PSU. So that would mean the momentary switch activates a circuit that latches the signal low. And then hitting the power switch again will unlatch the signal. Is this correct?

I tried searching for the ATX soft power circuit.... Is there a standard circuit that all boards use? I'm assuming there is an issue with this circuit on my motherboard. I tried to trace it out to find the components, but that proved to be extremely difficult.

I have thought of a work around. I know I could just ground the power-on wire via a SPST switch, but I'd rather use the case's built in power button. So I thought of this: I could splice out the power-on wire from my 20-pin connector, then build my own soft power circuit, and hook the case's power switch up to that. But I would like to investigate the problem if I can. For one, so I can be sure it isn't going to cause any further problems. But also for the learning experience of it.

Try jumpering pins 10 & 11 on the FP connector (J19...next to the VRM heatsinks)

The attachment J19 Front Panel Connectors.jpg is no longer available

Reply 2 of 19, by Horun

User metadata
Rank l33t++
Rank
l33t++

Another thing: make sure in the BIOS under Power management you do not have anything odd like "AC Power Restart on Loss" or some such... set to enabled ! Or it will power up anytime board/bios detects possible power...
Other thought: I have a older PcChips board with both ATX and AT PSU connectors and the ATX always acted flakey (not powering on sometimes, etc) but the AT always worked as should...

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 3 of 19, by ziggy587

User metadata
Rank Newbie
Rank
Newbie

One other thing I forgot to mention in the first post... If I hook a switch up to the front panel connector pins 10 and 11, it does nothing. The computer will power itself on, and pressing or holding the power button doesn't do anything. Sorry for leaving that part out, seems kind of important. 😀 But this is why I'm thinking there might be something wrong with the soft power circuitry on the motherboard. Perhaps it's damaged in a way that's permanently holding it low?

PC Hoarder Patrol wrote on 2022-11-27, 01:57:

Try jumpering pins 10 & 11 on the FP connector (J19...next to the VRM heatsinks)

If I jumper pins 10 and 11, then what would I connect the power switch to?

For reference, here's a snip from a manual for this board that I found online...

FP Connector Pinout

33v6Zhs.jpg

Horun wrote on 2022-11-27, 03:41:

Another thing: make sure in the BIOS under Power management you do not have anything odd like "AC Power Restart on Loss" or some such... set to enabled ! Or it will power up anytime board/bios detects possible power...
Other thought: I have a older PcChips board with both ATX and AT PSU connectors and the ATX always acted flakey (not powering on sometimes, etc) but the AT always worked as should...

Thanks, I'll double check that. That's one of the things I normally look for in the BIOS, I definitely like to keep it as "stay off after power loss." I'm pretty sure it's already set that way, but I'll double check to make sure.

Reply 4 of 19, by ziggy587

User metadata
Rank Newbie
Rank
Newbie

OK, so I was poking around a little and found some things that confuse me further...

The front panel connector pins 10 and 11 are for the power switch.

Pin 10 goes to pin 1 of a 74HCT14. I have not (yet) been able to trace out any more. The 74'14 is located between two ISA slots and is very hard to trace out from there.

Pin 11 of the front panel connector goes to the anode side of diode D5. The cathode of D5 goes to JP6 pin 1. JP6 is the jumper to hold or clear the CMOS (close pin 1 and 2 to hold the CMOS). The positive side of CR2032 goes through a diode and a resistor to pin 1 of JP6, and pin 2 of JP6 goes to a pin on the CMOS chip. So that part makes sense. But I can't figure out why one of the pins for the power switch is connected to JP6 via a diode, that's not what I was expecting.

I tested D5 in circuit and got some odd results. When using the diode setting on my DMM, it reads (negative) -0.584v. It reads OL when I reverse the probes on D5, so I assume that means the diode is still functioning correctly, aside from a negative voltage reading. Reading the resistance across D5 is -3.3 MOhm, and OL when I reverse the probes. Does the negative voltage reading indicate that it's reverse bias?

I've tried searching for an example schematic of an ATX soft power circuit but haven't been able to come up with anything. Does anyone know of any you could link me to?

Reply 5 of 19, by Horun

User metadata
Rank l33t++
Rank
l33t++

Hmm something does not sound proper. maybe your board rev was not meant to have the ATX PSU part ?
From one Intel TX schematic one pin from PWRBT# is pulled to ground thru resister then goes to 74c14 Schmidt trigger then to 7404 buffer, then to the PWRBT# pin on the PIIX4 chips which is pulled high to SB3v thru resistor.
Other PWRBT# pin to SB3v. When Button is pushed it goes +3v which goes to Schmidt and drops pin U20 low causing the power up (if I read the schem correctly) so if PWRBT is pulled low (grounded) it Powers up.
another example: From one Intel 440MX chipset schematic (I know not same but yours is Intel 430TX chipset so "could be" similar) one pin of PWRBTN pin at front panel header goes to the chip set pwrbtn#, it is also pulled high thru resistor.
Other pin is ground. When Button is pushed it forces pwrbtn# line to ground initiating the power on/power off cycle by taking pin U2 of 443MX low.
Think I read the schems right 😁

Last edited by Horun on 2022-11-28, 02:12. Edited 2 times in total.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 6 of 19, by Horun

User metadata
Rank l33t++
Rank
l33t++

440MX schem parts pictures.
The TX schematic is Intels "TX Desktop Rev 2.1, Dec. 19 1996" pdf
The 440MX schematic is from "Intel 440MX Scalable Low Power Development Kit sch_27363201" pdf
Note: Those are not the original names of the PDF's, I got them digging thru Intels old Developer website stuff some time ago....

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 7 of 19, by Horun

User metadata
Rank l33t++
Rank
l33t++

Fixed my boo-boo in first post. Here is one more cut from the "Intel Low-Power Pentium Processor Evaluation Platform Manual"

added: anyone wanting some of the Intel schems you can get them here: http://web.archive.org/web/*/http://developer … intarch/schems/*
there are some others too but that is where I found most of them last year....just fixed my effe-up here too. I blame old FF for it's bad caching ;p

Last edited by Horun on 2022-11-28, 03:02. Edited 1 time in total.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 8 of 19, by ziggy587

User metadata
Rank Newbie
Rank
Newbie

Thanks for the schematics! When I get the time I will review the motherboard again and see what I can come up with. In the meantime I will study the schematics.

Reply 9 of 19, by rasz_pl

User metadata
Rank l33t
Rank
l33t
ziggy587 wrote on 2022-11-27, 23:31:

Pin 10 goes to pin 1 of a 74HCT14. I have not (yet) been able to trace out any more. The 74'14 is located between two ISA slots and is very hard to trace out from there.

pin 2 goes directly to BGA chip so you wont find it, but you can measure and compare pin 1 and 2 (should be inverted output) to see if 74HCT14 is functioning

ziggy587 wrote on 2022-11-27, 23:31:

Pin 11 of the front panel connector goes to the anode side of diode D5. The cathode of D5 goes to JP6 pin 1. JP6 is the jumper to hold or clear the CMOS (close pin 1 and 2 to hold the CMOS). The positive side of CR2032 goes through a diode and a resistor to pin 1 of JP6, and pin 2 of JP6 goes to a pin on the CMOS chip. So that part makes sense. But I can't figure out why one of the pins for the power switch is connected to JP6 via a diode, that's not what I was expecting.

why? thats a good spot guaranteed to always be powered by 3V no matter what. Pressing power button sends that 3V to 74HCT14 just like in Horuns P_eval1.jpg and P_eval2.jpg

AT&T Globalyst/FIC 486-GAC-2 Cache Module reproduction
Zenith Data Systems (ZDS) ZBIOS 'MFM-300 Monitor' reverse engineering

Reply 10 of 19, by Horun

User metadata
Rank l33t++
Rank
l33t++
rasz_pl wrote on 2022-11-28, 03:32:
pin 2 goes directly to BGA chip so you wont find it, but you can measure and compare pin 1 and 2 (should be inverted output) to […]
Show full quote
ziggy587 wrote on 2022-11-27, 23:31:

Pin 10 goes to pin 1 of a 74HCT14. I have not (yet) been able to trace out any more. The 74'14 is located between two ISA slots and is very hard to trace out from there.

pin 2 goes directly to BGA chip so you wont find it, but you can measure and compare pin 1 and 2 (should be inverted output) to see if 74HCT14 is functioning

ziggy587 wrote on 2022-11-27, 23:31:

Pin 11 of the front panel connector goes to the anode side of diode D5. The cathode of D5 goes to JP6 pin 1. JP6 is the jumper to hold or clear the CMOS (close pin 1 and 2 to hold the CMOS). The positive side of CR2032 goes through a diode and a resistor to pin 1 of JP6, and pin 2 of JP6 goes to a pin on the CMOS chip. So that part makes sense. But I can't figure out why one of the pins for the power switch is connected to JP6 via a diode, that's not what I was expecting.

why? thats a good spot guaranteed to always be powered by 3V no matter what. Pressing power button sends that 3V to 74HCT14 just like in Horuns P_eval1.jpg and P_eval2.jpg

Might need a scope to see if the Schmidt 74c14 is actually working, maybe not.
Excellent thought raz_pl ! Using the battery for the VSB makes sense at a time when ATX was first coming out and being a AT/ATX psu board. Good thinking.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 11 of 19, by Repo Man11

User metadata
Rank l33t
Rank
l33t

A few years ago I had a thread about my Biostar MB-8500TTD and Feipoa told me that he added an ATX power connector to his MB-8500TTD but it did not work.

"We do these things not because they are easy, but because we thought they would be easy."

Reply 12 of 19, by rasz_pl

User metadata
Rank l33t
Rank
l33t
Horun wrote on 2022-11-28, 04:06:

Excellent thought raz_pl ! Using the battery for the VSB makes sense at a time when ATX was first coming out and being a AT/ATX psu board. Good thinking.

its not even battery, this is what OP traced:

The attachment rtc_rst.jpg is no longer available

except this is part MSI MS-7024 Pentium 4 motherboard 😀 In AT there is no standby so with computer Off CMOS is sustained using battery. In ATX CMOS is sustained with Standby voltage until unplugged from wall socket, only then battery takes over. Many tales of people with fully working computers ... until the day power went out or they had to unplug from the wall for a second and ended up with a problem (in case of certain modern Asrock motherboards that can even mean bricked board).

>Pin 11 of the front panel connector goes to the anode side of diode D5.
means pin 11 is wired to 3(or 5) Volt Standby supply. OPs D5 is D9 from my picture

ziggy587 wrote:

I tested D5 in circuit and got some odd results. When using the diode setting on my DMM, it reads (negative) -0.584v. It reads OL when I reverse the probes on D5, so I assume that means the diode is still functioning correctly, aside from a negative voltage reading. Reading the resistance across D5 is -3.3 MOhm, and OL when I reverse the probes. Does the negative voltage reading indicate that it's reverse bias?

negative reading in diode test mode? was this on system plugged in? Never do that! "Reading the resistance" NEVER measure anything other than voltage on plugged in system. Speaking of which - on plugged in system measure voltage on D5, Pin 1 and 2 of 74HCT14, press and hold power button and measure Pin 1 and 2 again.

Horun can you post or link to your TX MX diagrams? I only have HX one

AT&T Globalyst/FIC 486-GAC-2 Cache Module reproduction
Zenith Data Systems (ZDS) ZBIOS 'MFM-300 Monitor' reverse engineering

Reply 13 of 19, by Horun

User metadata
Rank l33t++
Rank
l33t++

Here is one TX: http://web.archive.org/web/20000616111232/htt … gnex/tx_sch.htm
http://web.archive.org/web/20000816233911/htt … gnex/tx_sch.pdf
and another : http://web.archive.org/web/20040723101316/htt … ls/27323502.pdf
one of the 440MX: http://web.archive.org/web/20030628045856/htt … 635_sch_pdf.pdf
will post the other if I can find it again...

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 15 of 19, by ziggy587

User metadata
Rank Newbie
Rank
Newbie
Repo Man11 wrote on 2022-11-28, 04:53:

A few years ago I had a thread about my Biostar MB-8500TTD and Feipoa told me that he added an ATX power connector to his MB-8500TTD but it did not work.

In what way did it not work?

PS - Hello, motherboard brother! 😀

rasz_pl wrote on 2022-11-28, 05:38:

negative reading in diode test mode? was this on system plugged in? Never do that! "Reading the resistance" NEVER measure anything other than voltage on plugged in system.

No, that negative reading (diode and resistance test) was done when the motherboard was powered off and unplugged. However, I now realize that when the motherboard is unplugged, then battery voltage is then applied to JP6 pin 1. So that's where the negative reading was coming from. I removed the jumper and the battery, and now D5 is reading positive in the diode test as you would expect it to.

rasz_pl wrote on 2022-11-28, 05:38:

Speaking of which - on plugged in system measure voltage on D5, Pin 1 and 2 of 74HCT14, press and hold power button and measure Pin 1 and 2 again.

OK, with the mobo powered on here's what I've found...

D5 - 3.566v on the anode side, 3.501v on the cathode side

74'14 (without holding power button)
Pin 1 measures 0v, but will occasionally read 0.001v or 0.012v
Pin 2 is measuring 0.03v, but will briefly measure 4.85v at times

74'14 (holding the power button)
Pin 1 = 3.656v
Pin 2 = 0.030v

For reference, here's what I was able to trace out.

The attachment Capture.JPG is no longer available

Reply 17 of 19, by ziggy587

User metadata
Rank Newbie
Rank
Newbie
rasz_pl wrote on 2022-11-29, 02:37:

>Pin 2 is measuring 0.03v, but will briefly measure 4.85v at times

Is that measurement error - not pressing probe down hard enough/missing the pin?
if not then you found your problem

I thought so too, but no. And I just checked it again last night to be sure. Pin 1 will be near 0v and jump to 3.65v when I press and hold the power button. Pin 2 is definitely acting weird. I did some more testing last night. It will sometimes hold at 4.85v for a while, then drop to 0v for no apparent reason (no buttons pressed). If it's at 4.85v and I press the power button, it will drop to 0v, but then it will stay at 0v after releasing the power button. What's strange is that pin 2 will stay at or near 0v at that point, but if I probe pin 1 (at 0v) then pin 2 will magically jump back to 4.85v for a few seconds.

So if I'm understanding correctly, pin 1 is the input and pin 2 is the output. If pin 1 is low then pin 2 should be high and vice versa. So the fact that pin 2 is low when pin 1 is low shows that there's a problem with the 74'14. Correct? Is it possible that there could be any pull-down components between the power button and pin 1 of the 74'14 that have failed, and cause this problem? Or would the best assumption be that the 74'14 is bad? And then my question would be, is it common for 74 logic to fail on its own? Or could something have caused it to fail?

Reply 18 of 19, by rasz_pl

User metadata
Rank l33t
Rank
l33t
ziggy587 wrote on 2022-11-30, 15:12:

What's strange is that pin 2 will stay at or near 0v at that point, but if I probe pin 1 (at 0v) then pin 2 will magically jump back to 4.85v for a few seconds.

is it common for 74 logic to fail on its own? Or could something have caused it to fail?

not common, I would suspect cracked trace first, but if you measure directly chip legs then its a good candidate for first resoldering and then replacement

AT&T Globalyst/FIC 486-GAC-2 Cache Module reproduction
Zenith Data Systems (ZDS) ZBIOS 'MFM-300 Monitor' reverse engineering

Reply 19 of 19, by ziggy587

User metadata
Rank Newbie
Rank
Newbie

OK, I think I found out what the issue is.

I removed the 74'14 and replaced it with a brand new part, but pin 2 still exhibits the same behavior as before. So I don't believe there is anything wrong with the original 74'14. I've also observed that pin 3 is always high, pin 4 is always low, and pin 4 seems to be connected directly to the ATX power-on signal at the 20-pin connector.

There is a foot print for a missing IC, looks like SO14 package, probably another logic chip. Pin 2 of the 74'14 goes to pin 3 of this missing IC. Around this missing IC, I can see traces going to a missing cap (probably for decoupling) and a missing resistor. I also see traces going to a diode footprint that is populated with a jumper, and another diode footprint that is unpopulated.

I suppose I could try and figure out what the missing circuitry is, but it might be a lot of trial and error and guess work. Another option would be to go back to using the P8/P9 adapter, and try and create my own latching circuit that I hook up to the case's power button. I'm open to opinions and suggestions.

Thank you, everyone, for you help with this!