myne wrote on 2025-09-21, 10:52:

Ah. My bad. Refer back to previous post with datasheet. 1. Identify target voltage. 2. Identify relevant VID pins on slot. 3. […]
Show full quote

TheMysteriousGray wrote on 2025-09-21, 05:43:

the board is a first revision AB-BH6, not a CUBX. I've heard that the 1.0x BH6 boards don't have a voltage regulator equipped to […]
Show full quote

myne wrote on 2025-09-21, 04:21:

Have you considered the possibility that the cpu is faulty? […]
Show full quote

Have you considered the possibility that the cpu is faulty?

The cubx was built for coppermines according to other posts.
This strongly implies that the voltage support is inbuilt.

The only change from memory was the voltage controller, usually a 20-40 leg rectangular ic. Look for the datasheets of the 2 obvious ones. One will be the clockgen, the other the vrm.

You'll see that the vrm should be capable of ~1.3-2.8v.

Until it works in another board, I'm assuming the cpu is dead.

You could try washing it. I've never had electronics get worse after a good wash. Many have come good.

the board is a first revision AB-BH6, not a CUBX. I've heard that the 1.0x BH6 boards don't have a voltage regulator equipped to deal with the lower voltages of Coppermine CPUs. I was considering getting the CUBX to replace this motherboard, but unfortunately somebody scooped the one I was looking at out from under me.

I have considered the possibility that the CPU is busted, but I don't have a second system on hand to know for sure. I feel like if the CPU was well and truly dead, then the computer wouldn't respond at all, but it turns on for a second or two each time I plug it back in. It just shuts itself off immediately afterward and refuses to turn on unless I change CPUs.

I could start by just buying a different CPU online, and if that works, I'd know. Hopefully, finding one will be cheaper than finding a good slotket or replacing the motherboard.

Ah. My bad. Refer back to previous post with datasheet.
1. Identify target voltage.
2. Identify relevant VID pins on slot.
3. Sticky tape to block.
Test.
Any idiot can cut a piece of sticky tape. It is fiddly, but achievable.

So if I understand you right, in order to force the CPU to report a required voltage of 1.8V, I have to find a specific pin on the slot and cover it with tape? And when you say “slot”, are you referring to the slot on the motherboard or the insertable end of the CPU?

I’m only asking these clarifying questions so I don’t make a mistake, if I try this.

shevalier wrote on 2025-09-21, 11:43:

Once again, this is a jumper-free motherboard. It sets the voltage specified in the BIOS ONLY after it boots. If it fundamentall […]
Show full quote

H3nrik V! wrote on 2025-09-21, 10:08:

You could try going into BIOS with the 450 cpu and check which voltages you are able to select manually

Once again, this is a jumper-free motherboard.
It sets the voltage specified in the BIOS ONLY after it boots.
If it fundamentally doesn't recognize processors with voltages below 1.8 V, then it won't supply power to the processor AT ALL.
There's nothing to change the voltage with—the processor won't be powered at all.

And how Abit control CPU VRM —digitally or through the GPIO of the MIO chip—these are implementation methods.

I'm aware it's a jumper free motherboard - the pictures I posted are from my own board. 😀 my point is, if one could manually set the voltage in bios to less than 1.8V, I would suspect that voltage was supported by the VRM.

On further thoughts, though, if it's a Katmai cpu, it may not show all the lower voltages.

H3nrik V! wrote on 2025-09-21, 14:49:

shevalier wrote on 2025-09-21, 11:43:

Once again, this is a jumper-free motherboard. It sets the voltage specified in the BIOS ONLY after it boots. If it fundamentall […]
Show full quote

H3nrik V! wrote on 2025-09-21, 10:08:

You could try going into BIOS with the 450 cpu and check which voltages you are able to select manually

Once again, this is a jumper-free motherboard.
It sets the voltage specified in the BIOS ONLY after it boots.
If it fundamentally doesn't recognize processors with voltages below 1.8 V, then it won't supply power to the processor AT ALL.
There's nothing to change the voltage with—the processor won't be powered at all.

And how Abit control CPU VRM —digitally or through the GPIO of the MIO chip—these are implementation methods.

I'm aware it's a jumper free motherboard - the pictures I posted are from my own board. 😀 my point is, if one could manually set the voltage in bios to less than 1.8V, I would suspect that voltage was supported by the VRM.

On further thoughts, though, if it's a Katmai cpu, it may not show all the lower voltages.

There’s no risk of the BIOS displaying a voltage the VRM can’t actually target, is there? Like I said before, I only ask to make doubly sure I’m not making a mistake by doing this or that.

No.
Look on the board for the rectangular ICs. It will have 20-50 legs and be near the row of big capacitors.
Google "xxx datasheet" and it will tell you exactly what it can do.

Example: https://www.analog.com/media/en/technical-doc … eets/1553lf.pdf

If you want to be absolutely sure, post the model of the VRM IC on your board (little black chip near the power connector). Its specs will tell you what is possible and what happens if selecting something impossible.

Tbh though we already know that: if you insert a CuMine CPU which asks for <1.8V, it doesn't get any power. If it were possible to do the same via BIOS, the same would happen, as this is VRM behaviour, not BIOS behaviour.

Is this the chip?

TheMysteriousGray wrote on 2025-09-21, 17:03:

Is this the chip?

The attachment D42DF513-4915-43AD-96EC-558B524518CD.jpeg is no longer available

Looks like this:

https://www.onsemi.com/products/power-managem … trollers/cs5165

Datasheet: https://www.onsemi.com/download/data-sheet/pdf/cs5165-d.pdf seems as it goes down to 1.25'ish volts. Same that's on my BH6

Off course, it will only work if all 5 bits for voltage selection are implemented, and not just pulled high or low.

So unless the VRM is faulty, failing or not physically configured right, then the most likely culprit is that the CPU itself is busted, it would seem.

Honestly, if that’s true, bit of a relief. Slot 1 Coppermines are surprisingly cheap compared to slotkets or motherboards, at least on eBay. Could just buy a new one, stick it in and see what happens

Try the sticky tape. You only have to block vid1, pin a120 to get 1.8v.
It's the second pin from the end on the front.
A121 is vid4 which is 0 anyway, so you can block both the last 2 pins.

It can't get much easier than that.

myne wrote on 2025-09-21, 17:59:

Try the sticky tape. You only have to block vid1, pin a120 to get 1.8v. It's the second pin from the end on the front. A121 is v […]
Show full quote

Try the sticky tape. You only have to block vid1, pin a120 to get 1.8v.
It's the second pin from the end on the front.
A121 is vid4 which is 0 anyway, so you can block both the last 2 pins.

It can't get much easier than that.

Blocked the pin on the CPU like I think you’re saying, even counted to make sure it was A120. Exact same thing happened. I even tried putting tape over the VID1 pin on the voltage chip for a lark and nothing changed. Either I somehow did it wrong or this chip is borked.

Heh it should be labelled.
Oh well, worth a shot

https://upload.wikimedia.org/wikipedia/common … slot1_naked.jpg

It was the second to last pin on the “A” side like you said. I only counted because, as you can probably gather by now, I am an EXTREMELY cautious person who does not like handling equipment like this without knowing exactly what he’s doing 🤣. I do appreciate the tip since I’m willing to try just about anything short of ripping the board apart and putting it back together to fix an issue that’s not really THAT important in the grand scheme.

I figure the next step is just to buy another Slot 1 Coppermine and put it in to see if anything changes. Probably a cheap one. If I get the same issue as before, then this board is just not cut out for bigger badder CPUs, though I can’t imagine why.

TheMysteriousGray wrote on 2025-09-21, 09:35:

Is the PCI divider not determined by software in any way? because the BIOS showed that a 1/4 PCI divider kicks in at 133 MHz even before I flashed it to the SS version. If so, that’s an incredibly stupid oversight for the BIOS to offer a feature the board can’t do.

It is couple of years now when I had that combo running last time and can’ remember what BIOS shows, but it isn’t just something that is defined by software. You need ICs to make that happen.

There are several discussions about this here in Vogons, as well as vintage forum posts. Like this example:

P3-850 on Abit BH6 v1.01?

But as I said, I had no trouble running 800/100 Coppermine on my 1.01.

It doesn't matter what kind of VRM controller it has.
There's a sort of "magic relay circuit" between it and the processor, controlled by software.
Considering that dynamic voltage regulation only appeared in Tualatin mobile CPU, for descktop Mendos/Katmai/Cooper/Tualatin it's handled by SOME vendor-specific solution.
And if Abit hadn't included the CooperMine detection feature in the version 1.0 motherboard, it "magic relay" wouldn't have enabled or switched.
If this "software control" hadn't existed, and the VID signals were transmitted directly, everything would have worked.
However, version 1.01 may differ from 1.0 by a single bootstrap resistor, which sets a different initial value for one of the VIDs.
For example, with "11111" - no CPU (or 1.2V for this PWM controller from On-semi) to something else.

Don't get hung up on "starting at 1.3V" and "which PWM controller is used."
This is a "black box" that behaves exactly as it does, and there is no other behavior.

The only other thing I can think of is that grounding bsel1 causes the cpu to reset.bsel1 was a ground on p2 boards

Bsel1 is pin a14

shevalier wrote on Yesterday, 05:21:

It doesn't matter what kind of VRM controller it has. There's a sort of "magic relay circuit" between it and the processor, cont […]
Show full quote

It doesn't matter what kind of VRM controller it has.
There's a sort of "magic relay circuit" between it and the processor, controlled by software.
Considering that dynamic voltage regulation only appeared in Tualatin mobile CPU, for descktop Mendos/Katmai/Cooper/Tualatin it's handled by SOME vendor-specific solution.
And if Abit hadn't included the CooperMine detection feature in the version 1.0 motherboard, it "magic relay" wouldn't have enabled or switched.
If this "software control" hadn't existed, and the VID signals were transmitted directly, everything would have worked.
However, version 1.01 may differ from 1.0 by a single bootstrap resistor, which sets a different initial value for one of the VIDs.
For example, with "11111" - no CPU (or 1.2V for this PWM controller from On-semi) to something else.

Don't get hung up on "starting at 1.3V" and "which PWM controller is used."
This is a "black box" that behaves exactly as it does, and there is no other behavior.

I get your point, but it's a hard fact that some motherboards do have VRMs, that doesn't support lower than e.g. 2.0Volts. My own Intel SE440BX-2 in early revision is an example. But agreed, it will only work if the "black box" is wired correctly. But then again, if BIOS allows selecting 1.3 Volts, I'd suspect that the board is wired to be able to do that.

H3nrik V! wrote on Yesterday, 09:02:

I get your point, but it's a hard fact that some motherboards do have VRMs, that doesn't support lower than e.g. 2.0Volts. My own Intel SE440BX-2 in early revision is an example. But agreed, it will only work if the "black box" is wired correctly. But then again, if BIOS allows selecting 1.3 Volts, I'd suspect that the board is wired to be able to do that.

Intel motherboards are not among those that allow for software override in terms of frequency and voltage.
This is a very different case- VRM8.1 and 1.8(or 2)V

I put in the old CPU and checked the voltage settings in the BIOS, and the lowest option is indeed 1.3V, but if I understand you guys correctly, that option is irrelevant if the VRM isn’t actually wired to go that low?

The only thing I can think to do to know for sure is undervolt the old CPU and try to boot into Windows to see what happens, but some part of me says that’s a bad idea.

TheMysteriousGray wrote on Yesterday, 12:16:

I put in the old CPU and checked the voltage settings in the BIOS, and the lowest option is indeed 1.3V, but if I understand you guys correctly, that option is irrelevant if the VRM isn’t actually wired to go that low?

The only thing I can think to do to know for sure is undervolt the old CPU and try to boot into Windows to see what happens, but some part of me says that’s a bad idea.

Pretty sure that the Katmai won't go anywhere near 1.3V - maybe 1.9 or 1.8 but probably not lower.
What if you set FSB to 66 with the Katmai, turn off and then try the swap?

FWIW, my BH6 is kind of an a*s when it comes to changing CPUs - often, when I have swapped, it powers on, but no life AT all. I can usually get it running by holding the reset button while I power on, and then 10-15 seconds more, then it usually defaults to 66 MHz FSB ... But YMMV, at least I don't feel it's being a consistant success.

The attached image is from the BH6 manual. Seems my "hold reset" method isn't the right one 🤣