If you are desperate then why not mod the motherboard itself?

is the request about exact brand/model for collecting, or you need a solution to get stuff working ? Basically a practical reason.

I am trying to do sthg. very specific. Otherwise, I would indeed mod "a" board. My current project is to get a Tualatin to run on a Socket 8 (Pentium Pro) board.

This involves a Powerleap PL-Pro/II Socket 8 to 370 (PPGA) adapter. Only runs PPGA Celerons out of the box. (Reversibly) modified the Adapter with the well-known PPGA to FC-PGA1 mod. Now runs everything up to the highest Coppermine.

However, I would also like to get a Tualatin to run on it.

So you see, it's a pretty specific project. Already tried one of the Korean modded Tualatins, supposedly able to run on FC-PGA as well. But: no post.

Any alternative approach would be welcome though (except getting a Tualatin Board... which is not the point 😀.

If you can get 1.5 volts to that then it's just a socket mod like you've already done.

At some point I hope to make an adapter for this, in the meantime I am collecting info and cross linking relevant threads.

Re: Socket 370 Tweaker

1. I think a Lin-Lin 370 CPU Converter does the same thing as you are looking for, but they are equally rare
2. The most practical solution for you is to buy a pin-modded Tualatin from the Korean seller on eBay

He already tried #2

Frunzl wrote on 2022-10-01, 01:21:

I am trying to do sthg. very specific. Otherwise, I would indeed mod "a" board. My current project is to get a Tualatin to run o […]
Show full quote

I am trying to do sthg. very specific. Otherwise, I would indeed mod "a" board. My current project is to get a Tualatin to run on a Socket 8 (Pentium Pro) board.

This involves a Powerleap PL-Pro/II Socket 8 to 370 (PPGA) adapter. Only runs PPGA Celerons out of the box. (Reversibly) modified the Adapter with the well-known PPGA to FC-PGA1 mod. Now runs everything up to the highest Coppermine.

However, I would also like to get a Tualatin to run on it.

So you see, it's a pretty specific project. Already tried one of the Korean modded Tualatins, supposedly able to run on FC-PGA as well. But: no post.

Any alternative approach would be welcome though (except getting a Tualatin Board... which is not the point 😀.

IMHO, there are 2 main potential constraints/issues

a) Tualatin voltage requirements (may or may not be addressed by your Powerleap PL-Pro/II Socket 8 to 370 adapter, I don't know). It may be worth probing the adapter with a multimeter (while disconnected) to find a point where you can test Vcc while it is running to see what is actually supplied to the CPU

b) BIOS support or lack thereof (it may be necessary to add proper microcodes and/or otherwise mod the BIOS)

Either of these could be preventing a successful POST, AFAIK .

https://pdf1.alldatasheet.com/datasheet-pdf/v … SEM/US3004.html

Meets Latest VRM 8.4 Specification for PIII
On board DAC programs the output voltage from 1.3V to 3.5V.

http://brassicgamer.blogspot.com/2020/10/the- … ualatin.html#ct

The basic difference is that the Tualatin bus runs at 1.25V instead of Coppermine's 1.5V. This means Tualatin CPUs have an additional pin (for a total of 5) allocated for voltage identification. Depending on the logical state (hi or lo) of these pins, the CPU's core voltage can be set automatically. The additional pin makes twice as many voltage steps available (1.05V to 1.825V in 32x 0.025V steps) compared to Coppermine (1.3V to 2.05V in 16 steps). As mentioned, the VRM design guidelines changed from 8.4 to 8.5 to account for this more detailed level of voltage control.

You need one more mod to force your powerleap to always output 1.3V

Thanks for all your thoughts!

The Korean Tualatin does include the VSS-VID0-VID1-VID3 mod. It also only gets moderately warm when installed. I measured Vcore at 1.5 V.

Lack of bios support is unlikely, since I used Biospatcher (Award Bios) to (hopefully) enable Tualatin support. This was also necessary and sufficient to get the coppermines to work, so yeah...

The basic difference is that the Tualatin bus runs at 1.25V instead of Coppermine's 1.5V. This means Tualatin CPUs have an additional pin (for a total of 5) allocated for voltage identification. Depending on the logical state (hi or lo) of these pins, the CPU's core voltage can be set automatically. The additional pin makes twice as many voltage steps available (1.05V to 1.825V in 32x 0.025V steps) compared to Coppermine (1.3V to 2.05V in 16 steps). As mentioned, the VRM design guidelines changed from 8.4 to 8.5 to account for this more detailed level of voltage control.

You need one more mod to force your powerleap to always output 1.3V
[/quote]

By the way, I do not really understand what 1.25 V Bus voltage means. As per spec, they require 1.45 - 1.5 Vcore, which is provided by the Interposer/Adapter. 1.3 V would be heavily undervolted, although it might still post.

I also tried several Tualatin/Tualeron models (Korean mod or modded myself), since I read somewhere that high multipliers can hamper CPU recognition... But 7.5x Tualatins don't work either, and on the other hand, the 11.5x Coppermine Celeron does run, so it doesn't seem to be that either. I also thought that maybe L2 cache size might be a problem, but a regular 256k coppermine also works fine...

I really don't know, what to do next, except getting the PowerLeap PL-370/T (or the lin lin version), since it's the only one, which is explicitely compatible with PPGA.

Cpus around this time had a io and a core voltage.

Ah, sorry for my ignorance, I only knew that from the MMX and K6 CPUs, where it often was an issue if the board could do split voltage or not...
After socket 7, that ceased to be a problem for me, but should have been clear that this actually persisted, especially if attempting to run a chip from 2001 on a board from 1996 😀

Could that truely be the issue here? I remember that the MMX 233 ran happily at 3.3V on both core and io (or at least it posted)... but I will look into it.

Im getting old 😀 and posting nonsense 🙁 sorry.
1.25V refers to AGTL+ bus voltage and not Vcore, so no, you dont need to mod anything to run at 1.3V. Tualatins bus runs at a lower voltage because they were made on newer process node, but you can safely run them on 1.5V like Coppermines and Mendocinos.
So your modded tualatin has 0.025V and VTT_PWRGD taken care of?
Do you have a POST card to see if the CPU runs at all, and where the bios is getting stuck?

some old resources
https://www.tomek.cedro.info/files/old-www/kszak/bx_mod/
https://www.oocities.org/_lunchbox/
https://www.overclockers.com/forums/threads/r … owerleap.67417/

rasz_pl wrote on 2022-10-01, 20:52:

Im getting old :) and posting nonsense :( sorry. 1.25V refers to AGTL+ bus voltage and not Vcore, so no, you dont need to mod an […]
Show full quote

Im getting old 😀 and posting nonsense 🙁 sorry.
1.25V refers to AGTL+ bus voltage and not Vcore, so no, you dont need to mod anything to run at 1.3V. Tualatins bus runs at a lower voltage because they were made on newer process node, but you can safely run them on 1.5V like Coppermines and Mendocinos.
So your modded tualatin has 0.025V and VTT_PWRGD taken care of?
Do you have a POST card to see if the CPU runs at all, and where the bios is getting stuck?

some old resources
https://www.tomek.cedro.info/files/old-www/kszak/bx_mod/
https://www.oocities.org/_lunchbox/
https://www.overclockers.com/forums/threads/r … owerleap.67417/

Please do not worry, posting credible nonsense is the highest skill there is 😀

Yes:
- VSS is connected to VID25mV
- VttPWRGD (AK4) connected to PWRGD (AK26)
- DYN_OE (AN3) as well as RESET2 (AJ3) were removed and insulated.

In order to get PPGA -> FC-PGA compatibility, I also removed AM2 on the CPU and connected AH4 (RESET) to X4 (RESET2 on FC-PGA) on the socket (= Interposer).... come to think of it, I just realize that X4 is VSS on the tualatin and would therefore bridge RESET to VSS.. that can't be good, right?

But I need to go to X4 for reset in the socket... so I need to remove the pin on the CPU, while kepping it bridged on the socket (AH4 to X4), right?

Frunzl wrote on 2022-10-01, 20:38:

Ah, sorry for my ignorance, I only knew that from the MMX and K6 CPUs, where it often was an issue if the board could do split voltage or not...
After socket 7, that ceased to be a problem for me, but should have been clear that this actually persisted, especially if attempting to run a chip from 2001 on a board from 1996 😀

Could that truely be the issue here? I remember that the MMX 233 ran happily at 3.3V on both core and io (or at least it posted)... but I will look into it.

Sucess or not, don’t run a mmx at 3.3v core 🤦‍♂️

Sphere478 wrote on 2022-10-01, 22:12:

Frunzl wrote on 2022-10-01, 20:38:

Ah, sorry for my ignorance, I only knew that from the MMX and K6 CPUs, where it often was an issue if the board could do split voltage or not...
After socket 7, that ceased to be a problem for me, but should have been clear that this actually persisted, especially if attempting to run a chip from 2001 on a board from 1996 😀

Could that truely be the issue here? I remember that the MMX 233 ran happily at 3.3V on both core and io (or at least it posted)... but I will look into it.

Sucess or not, don’t run a mmx at 3.3v core 🤦‍♂️

not saying you should, but you could, especially if you wanna get 300 MHz out of it... but that is a totally unrelated issue 😀

Frunzl wrote on 2022-10-01, 21:26:

Please do not worry, posting credible nonsense is the highest skill there is :) […]
Show full quote

rasz_pl wrote on 2022-10-01, 20:52:

Im getting old :) and posting nonsense :( sorry. 1.25V refers to AGTL+ bus voltage and not Vcore, so no, you dont need to mod an […]
Show full quote

Im getting old 😀 and posting nonsense 🙁 sorry.
1.25V refers to AGTL+ bus voltage and not Vcore, so no, you dont need to mod anything to run at 1.3V. Tualatins bus runs at a lower voltage because they were made on newer process node, but you can safely run them on 1.5V like Coppermines and Mendocinos.
So your modded tualatin has 0.025V and VTT_PWRGD taken care of?
Do you have a POST card to see if the CPU runs at all, and where the bios is getting stuck?

some old resources
https://www.tomek.cedro.info/files/old-www/kszak/bx_mod/
https://www.oocities.org/_lunchbox/
https://www.overclockers.com/forums/threads/r … owerleap.67417/

Please do not worry, posting credible nonsense is the highest skill there is 😀

Yes:
- VSS is connected to VID25mV
- VttPWRGD (AK4) connected to PWRGD (AK26)
- DYN_OE (AN3) as well as RESET2 (AJ3) were removed and insulated.

In order to get PPGA -> FC-PGA compatibility, I also removed AM2 on the CPU and connected AH4 (RESET) to X4 (RESET2 on FC-PGA) on the socket (= Interposer).... come to think of it, I just realize that X4 is VSS on the tualatin and would therefore bridge RESET to VSS.. that can't be good, right?

But I need to go to X4 for reset in the socket... so I need to remove the pin on the CPU, while kepping it bridged on the socket (AH4 to X4), right?

Removed X4 on the CPU, I had modded myself. No post....

However, I noticed that on the Korean modded Chip, he bridged X4 (Vss) to W5 (Vcc)... So maybe X4 on the Tualatin needs voltage for him to live at all... When I removed X4 from my modded chip, I unfortunately ripped off the pad, so I cannot make the X4-W5 connection on mine anymore and would have to mutilate the Korean chip..

But first, I would very much like to have some idea if this makes any sense at all.

Frunzl wrote on 2022-10-01, 23:23:

When I removed X4 from my modded chip, I unfortunately ripped off the pad

🙁
When I was modding motherboards for coppermines back in ~2000 I would drill socket plastic a little bigger and sleeve insulation stripped from thin wire over the pin I wanted disconnected.

if you dont have POST card you can try oscilloscope or logic probe / logic analyzer plopped into BIOS data/address pins to see if there is any CPU activity at all.