AlexZ wrote on 2022-04-09, 14:09:

Given this information the best option would be to downclock PIII 1Ghz to 750Mhz using 100Mhz FSB as at that frequency current draw will be only slightly higher than PII 450 and below 16 Amp. Celeron 1000 will be about as fast as PIII 750 anyway.

According to my limited benchmark stats in 3DMark99 they are slighlty faster than PIII 850/100. But other benchmarks may of course differ.

DrLucienSanchez wrote on 2022-04-09, 14:02:

Unfortunately I started to get instability again, cold booted, and BIOS hung, booted again, said primary HDD failed (it hadn't). So back to AMI BIOS, 32GB HDD and 550Mhz, alas, it's just not meant to be!

Pity...

DrLucienSanchez wrote on 2022-04-09, 14:02:

I didn't know it would run hot in all honesty, also expected all the SlotKets on ebay (about three of them) to just work, so best it didn't work out, as if it did, my wallet would've suffered a little.

There are currently some 10 slotkets on ebay but none of them come with the required voltage jumpers -plus- Coppermine support. Getting the 'right' type of slotket + celeron 1000 + cooling solution may set you back more than getting a 750 or 800 Slot 1 cpu (which seem to be within the safety margin). But this info is water under the bridge. Good luck with your next step.

PARKE wrote on 2022-04-09, 14:47:

According to my limited benchmark stats in 3DMark99 they are slighlty faster than PIII 850/100. But other benchmarks may of course differ.

Unfortunately Celeron 950 didn't shine that much when I benchmarked it - Re: Celeron, the good bad and ugly.. I used 3D Mark 99, 3D Mark 2000, 3D Mark 2001. That is why I use PIII 900E.

As far as slotkets go, out of my 11 slotkets only two have voltage jumpers.

AlexZ wrote on 2022-04-09, 15:30:

PARKE wrote on 2022-04-09, 14:47:

According to my limited benchmark stats in 3DMark99 they are slighlty faster than PIII 850/100. But other benchmarks may of course differ.

Unfortunately Celeron 950 didn't shine that much when I benchmarked it - Re: Celeron, the good bad and ugly.. I used 3D Mark 99, 3D Mark 2000, 3D Mark 2001. That is why I use PIII 900E.

You list the gpu benchmark results in your link.
My observation in 3dMark 99 is based on the cpu benchmarks. In some 70 or so results with varying mobo's / videocards / drivers / resolutions etc. the PIII 850 (SL4Z2 stepping cD0) scores between 12.131 and 12.758 and the Celeron 1000 scores between 12.778 and 13.347. From this particular perspective the Celeron 1000 seems to pick up where the PIII 850 leaves off.
But I am sure that there are plenty other ways to get to other results.

snufkin wrote on 2022-04-08, 16:54:

From the photos, looks like Vcore is controlled by an SC1152, datasheet here: https://pdf1.alldatasheet.com/datasheet-pdf/v … H/SC1152CS.html

That says the lowest it supports is 1.8V, but it also says that any Vid not specifically listed will cause a shutdown. So if the CPU asks for 1.75V then there's a chance it just won't start up.

Good spot. I noticed my version of the board (otherwise identical with same revision number MS6156 1.0 BX7) has a SC1152CS(W) voltage controller which supports 1.3-3.5 volts, Tualatin works with a slotket and pin mod. My chip SC1152CS(W) is dated 99 week 9, his SC1152CS 98 week 44.
Still I don't think the SC1152CS, that can supply 1.8-3.5 volts, would have any trouble with a PIII-900 @1.8v with heatsink and fan, he could mod the vid1-4 pins on the cpu to only supply 1.8v.

DrLucienSanchez wrote on 2022-04-09, 10:47:

Unfortunately I started to get instability again, cold booted, and BIOS hung, booted again, said primary HDD failed (it hadn't). So back to AMI BIOS, 32GB HDD and 550Mhz, alas, it's just not meant to be!

One think worth mentioning, I had odd problems with this board back in 2018, random freezes at Bios screen, in Dos and win98. Some times it would work for weeks on end without any problem at all, then suddenly not. It turned out to be the Power Supply.
It took me a whole year to find the problem because the PSU appeared to be working fine, I tested the 5v /12v feeds with a multi-meter and all readings were within range. After using a new 400w atx power supply all random issues went away, it has been rock solid stable ever since.
I don't know how old or what rating the power supply your using now is, worth trying if you have a newer spare PSU around.

I have a MSI MS-6147 Rev. 1.1 i440BX mATX board with no C letter on the Winbond I/O controller and have used a Pentium III 500@560 (112 FSB) Katmai for the last years. It has the SC1152CSW voltage controller, a newer version of the SC1152CS. Never thought, that this controller also supports <1.8V Vcore for Coppermines, the SC1153CSW supports VRM 8.4. MSI boards with the C letter on the I/O chip have the right SC1153CSW controller for official Coppermine support. So I have never tested a Coppermine on this board. For security reasons,...

But today I have tried it out on my MSI-6147: A Pentium III 700 1.65V Slot 1 Coppermine cB0 stepping (tested stable on a Abit BH6 board with 133MHz FSB=933MHz with default 1.65V VCore) installed, 500 Katmai out, and after the first switch on the Bios says: Pentium III MMX CPU at 784MHz, because 112MHz FSB was set in the bios. Tried Unreal, 3DMark2000, Incoming (Voodoo3 3000 PCI because no AGP slot) for the first time and everything works fine, also much cooler and less power consumption than the Katmai PIII clocked to 560MHz.

So the SC1152CSW is inofficialy able to handle below 1.8V. With the wrong not supporting <1.8V controller, you would get a black screen, because the needed voltage for the CPU cannot be provided and thus the controller interrupts the power supply of the CPU. Only a pin mod or a Slotket S370 with the voltage setting to 1.8V would help.