Soyo might have early VRM chip which does not support voltages below 1.8v. If the VID requests 1,65, 1,7 v etc then it does not output any vCore at all.
Set jumpers on the slocket to 1.8volts and try again.

Paadam wrote on 2024-09-19, 04:52:

Soyo might have early VRM chip which does not support voltages below 1.8v. If the VID requests 1,65, 1,7 v etc then it does not output any vCore at all.
Set jumpers on the slocket to 1.8volts and try again.

Ah, yeah that's probably the issue with the Soyo board then.

If you check this .pdf list of slotkets, mine is the Star Tech one on Page 32, which has no voltage jumpers. So, apparently it is useful only for later boards.

That really stinks. I don't think I have any 440BX boards that are newer than this that would likely have the later VRM.

In fact, the only other 440BX I can think of right now that isn't in the "to repair" box is an FIC VB-601 v1.4... and upon closer inspection, THAT board has one swollen cap so it is going to likely need a bunch of caps replaced.

It's kind of crazy, I thought I had a pile of 440BX boards in my collection, but it turns out they are pretty much all 440LX, VIA Apollo Pro 133A or 133T. Not counting the OEM boards in a couple of Gateway G6 systems. I did find one decent looking Intel 810... but, I mean... it's an Intel 810. 😮

Hmm... is it possible to upgrade the VRMs on these to support lower voltages? Seems a tad sketchy, but if the Soyo just supported newer CPUs it would be totally fine for my needs. I'm happy if I can run a PII 266 or a PIII 850 in this thing since that pretty much spans the CPU requirements of the hardware I test on this machine.

I guess I could just use one of the Via boards... I'm just used to the compatibility and lack of issues with this 440BX. If I end up needing a second dedicated machine to rule out motherboard compatibility problems, then the tester is a waste of space. As it is, I very rarely have issues with compatibility.

EDIT: Confirmed... my Soyo SY-6BE+ has an LM2635M regulator, which only supports 1.8v to 3.5v.

The FIC VB-601 v1.4, however, has an RC5053M, which is probably similar others like the RC5050, RC5052M and RC5054A which all support 1.3v to 3.5v.

Interestingly, the Wintech W6BXA has an SC1152CS which is supposed to only be VRM 8.1 compliant according to the datasheet. Yet, my experience has been that it works totally fine with 1.7v Coppermines. This poster seems to have had similar results.

So yeah... looks like if I want to get a 440BX tester system with a Coppermine CPU, I have some repairs to do one way or the other. Fantastic. >_<

Anyway, I guess on the plus side I have learned a ton about 440BX VRM versions and what to look for. Highly relevant information you know. 🙄 😜

... just to clarify though, I would still greatly prefer to get Board #1 working, so if anyone has any thoughts as to what may have failed that would prevent the board from receiving power at all (and how to test it) it would be greatly appreciated.

Ozzuneoj wrote on 2024-09-19, 19:04:

... just to clarify though, I would still greatly prefer to get Board #1 working, so if anyone has any thoughts as to what may have failed that would prevent the board from receiving power at all (and how to test it) it would be greatly appreciated.

I am now reading about problems with the two motherboards.
I can suggest some attempts, you don't have to do any repairs, maybe you may have already made some attempts, but you haven't solved the problem, since trying doesn't cost you anything, I would say to start with card 1.
I would try to understand if the problem is only the AGP slot, then I would use an ISA video card, if there is a boot image, try a PCI VGA, if you see the boot screen, an investigation should be carried out on the capacitors and mosfets of the AGP slot, and possibly need to be replaced, if any are faulty, it will not work until replaced.
When testing the card I would remove everything that is not essential, LAN sound card etc. only VGA ISA and only one memory bank, if you have others they must be tested individually, or if you have a working module, which you had tested in this tab, use that, if you get the screen, then you can test the other modules individually.
If you get a BIOS Recovery screen, it means that the BIOS needs to be reprogrammed, or replaced, if the card after reprogramming does not show any screen, the BIOS chip is gone, replace it with an identical one or one of the same capacity (compatible).
Card 2, I would use a Slot1 370 adapter, with voltage regulators on board, in a MB for Celeron Mendocino, thanks to the right adapter, I even started a Celeron Tualatin, maybe your adapter uses MB regulators, in that case not I would make any changes if you only need it for short tests.

PC@LIVE wrote on 2024-09-22, 07:34:

I am now reading about problems with the two motherboards. I can suggest some attempts, you don't have to do any repairs, maybe […]
Show full quote

Ozzuneoj wrote on 2024-09-19, 19:04:

... just to clarify though, I would still greatly prefer to get Board #1 working, so if anyone has any thoughts as to what may have failed that would prevent the board from receiving power at all (and how to test it) it would be greatly appreciated.

I am now reading about problems with the two motherboards.
I can suggest some attempts, you don't have to do any repairs, maybe you may have already made some attempts, but you haven't solved the problem, since trying doesn't cost you anything, I would say to start with card 1.
I would try to understand if the problem is only the AGP slot, then I would use an ISA video card, if there is a boot image, try a PCI VGA, if you see the boot screen, an investigation should be carried out on the capacitors and mosfets of the AGP slot, and possibly need to be replaced, if any are faulty, it will not work until replaced.
When testing the card I would remove everything that is not essential, LAN sound card etc. only VGA ISA and only one memory bank, if you have others they must be tested individually, or if you have a working module, which you had tested in this tab, use that, if you get the screen, then you can test the other modules individually.
If you get a BIOS Recovery screen, it means that the BIOS needs to be reprogrammed, or replaced, if the card after reprogramming does not show any screen, the BIOS chip is gone, replace it with an identical one or one of the same capacity (compatible).
Card 2, I would use a Slot1 370 adapter, with voltage regulators on board, in a MB for Celeron Mendocino, thanks to the right adapter, I even started a Celeron Tualatin, maybe your adapter uses MB regulators, in that case not I would make any changes if you only need it for short tests.

Hi there, thanks for the reply.

Just to explain the situation better: When I made the thread Board #1 had already been tested outside of the case with nothing but a power supply attached (no cards) and it is still dead (no power at all). I have tried with different power supplies, CPUs, no CPU, etc. and there is no power at all to the board. It does not receive any power so it is not getting to the point of initializing the BIOS or cards at all. It is an electrical component on the motherboard itself that has failed in some way. So, I am trying to diagnose which component on the motherboard may have failed so I can repair it.

Regarding board #2, yes it appears it cannot run processors with voltages under 1.8v. It is unlikely that I will find a slotket with a dedicated voltage regulator these days because they are very rare and expensive, so finding one that will at least allow the voltage to be clamped at 1.8v (which should make little difference in heat on a 1.75v CPU) would probably be the easiest option. I would just need to find one (or modify one that I already have). Because of this added complication, I'd rather just fix Board #1.

Regarding the SOYO SY-6BE+, is there any reason that the LM2635M couldn't be replaced with an LM2636M? I have attached the datasheets. It looks to me like the pinout and other values are all basically the same between them, but I'm no engineer.

Would this be sufficient to enable support for less than 1.8v processors, since the latest BIOS seems to support them?

Also, it's kind of lame that the LM2635M datasheet says it can support the lower voltages but they are "disabled" and can be enabled upon request... whatever that means. Maybe they mean "if you want to support newer processors, pay us for the LM2636 instead."

HOLY COW, I FIXED IT! 🤯

(Solution in bold for those that don't have time to read the whole story.)

I have been mulling over the various ways to get my tester system up and running again, which is not as trivial as it may sound. The system has 30+ Windows folders on it that are fully set up for whatever hardware I am testing and I do not want to have to reconfigure them for a new motherboard, even (especially?) temporarily. As a sort of last-resort I ordered some updated LM2636 regulators to replace the coppermine-incompatible LM2635 on the SOYO SY-B6E+ since that board should be nearly identical as far as resource allocation and chipset drivers are concerned. Those are coming from Aliexpress and will take weeks, but hey, at least I got that ball rolling.

Tonight though, I wanted to take another feeble stab at fixing the WinTech. After repeatedly poking at every diode and capacitor and beeping out the ATX connector checking for shorts, I found nothing. One thing I remembered was that if I connect the power supply to the board and manually jump the green+black pins (my ATX extension has an in-line connection for a power switch for AT-style boards) in the ATX connector it actually did make the power supply turn on while connected to the board... it just wouldn't POST. I thought, "Hey, why not connect one of my POST cards while I do that to see what it says?"

So, I did that, and as expected the code read out was simply "--". All of the voltage and diagnostic LEDs were on though, except for one (I think it was "Ready"), which actually did surprise me, since I was expecting a voltage issue. At this point, I realized that the "Reset" LED was also lit. Hmm... didn't seem right to me. So I disconnected everything and sure enough, the pins on the motherboard to connect the Reset button were shorted without any button attached to them! To sanity-check myself I tested some other similar boards and none had pre-shorted reset pins. So, I hooked the board back up, powered it on directly with the PSU, and when I shorted the reset pins the lights did respond a bit, but didn't do anything new. When I shorted the power switch pins (keep in mind, the power supply was turned on manually by shorting the green and black wires), the diagnostic LEDs all changed around and after holding it for a while the code readout started showing numbers!!

After doing this, I shut it down, put in a video card and RAM, forced it on again and it booted!! So, naturally, after this I shut it down, disconnected the "force-on" switch and just used the normal power switch header and it WORKS!

Does anyone have any idea what would cause something crazy like this?

The only possibilities I can come up with:
1. When first started diagnosing this after it died, I found that the CMOS battery was insanely low... like 0.2v. I have no idea how it was still holding CMOS settings. Replacing it did not fix the issue but perhaps the low voltage caused some issue at some point.
2. The buttons on the tower I use this in are a tad funky... maybe some one-in-a-million fluke thing happened with the buttons being stuck or toggling on and off due to a poor connection and some logic on the board got "stuck".
3. The caps are original and they aren't name-brand, but none are swollen. Perhaps one has wandered too far out of spec and caused the reset\power logic to get stuck.

(By the way, the G400 Max (which I was trying to test when the board died) is working perfectly fine, even in this board now, so that was probably unrelated.)

One of the reset button pins appears to be connected to a nearby 14pin SMD Ti chip marked F08, which is most likely one of these. Whatever happened must have tripped up a gate in this SN7408 IC... I guess? I don't know. This is beyond my level of understanding\education. 🤣

Any input would be greatly appreciated. I would love to "fix" whatever caused it so that it never happens again... though it may be as simple as holding the power and reset buttons for a few seconds (possibly with the board forced-on by the PSU) until it goes back to normal.

EDIT: Just found some posts where others reported the reset being stuck on, though their boards actually powered on:
Motherboard stuck at reset
Socket 7 board stuck reset

Excellent work my friend, what caused the problem, well it could be a causal block (in the sense that only a certain combination causes the problem), I would exclude the problem with some faulty component, it would instead be possible that there is, for example, a capacitor that certain temperatures can block the start, but perhaps it is more likely that it is a bad soldering, in short, if the failure to start no longer occurs, I wouldn't worry too much, but if it recurs, you should understand if it is due to something you have done (changing cards for example).
However, I don't know if your problem solving method works with other motherboards, I have some motherboards that no longer boot, and the Reset LED stays on, I'm curious to try and see what happens if I do what you did, in the past from some research, I saw that the circuit is very simple, and goes directly to the SB (if I remember correctly), so if there is no problem at the SB, it should be quite simple to follow the circuit, and check the SMDs.