The 600E is already a Coppermine CPU, so the 850 should work too, all the VRM related caveats would have applied to the CPU you're currently using. Make sure your slotket actually supports Coppermine chips though, and not just older Celerons.

Official Abit line is no / not advised

https://web.archive.org/web/20020611094329/ht … eng/faq/bh6.htm

but plenty have done so anyway - I'd update to the latest BIOS (SS) and give it a go. A decent slocket may give you even more scope than Slot 1.

I found some old discussions:

https://arstechnica.com/civis/viewtopic.php?p … e178d#p16908755

https://groups.google.com/g/alt.comp.periphs. … t/c/Tc_kEGEjCwQ

In the Google groups one, the ABIT tech said "some later BH6 1.0x boards apply newer MOS-FET power amplifier (chip labeled 55N03, while earlier boards are 45N03), which could provide
sufficient current for Coppermine." And further down someone said the " L 3103 S" is an equivalent for 55N03, though I am not sure how to verify that.

We cannot officially claim that all BH6 1.0x support Coppermine, for they don't all fully VRM 8.4 spec. compliant. Some later BH […]
Show full quote

We cannot officially claim that all BH6 1.0x support Coppermine, for they
don't all fully VRM 8.4 spec. compliant.
Some later BH6 1.0x boards applies newer MOS-FET power amplifier (chip
labeled 55N03, while earlier boards' are 45N03), which could provide
sufficient current for Coppermine's application. We believe these boards are
the last several stepping of BH6 1.02.

To ensure if your BH6 1.0x is capable for Coppermine, please check your
MOS-FET chips' IC number.
Otherwise, you may also check your motherboard's backside. There would be a
PCB revision number. "0.5" is okay.

So it seems it would work fine on some later V1.0x boards with BIOS SS. Soggi's site has BIOS information, it does say that Coppermine over 700 isn't officially supported on V1.0X.

BTW was it not possible to change your FSB to 100 instead of 133 with your P3 600E? 133 MHz FSB CPUs cannot run at lower FSB?

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't have a 1/4 divider for PCI.

This means pci runs at 44.3MHz, including the Southbridge which is also connected to the PCI bus. 44MHz is too high for many PCI devices.

As for the tdp, in the manual I've found the max CPU listed is pentium 3 500.
Now to me it doesn't mean it's not compatible with newer CPUs but rather that the documentation was produced when newer CPUs didn't exist

The tdp of the maximum listed CPU (p3 500) is 28W.
The tdp of the pentium 3 850eb is 22W (probably due to the lower voltage). Correct me if I'm wrong, but this looks rather good. I think you'd be fine even with a 1GHz pentium 3 (which I did try on my bh6 although not for long because I wanted it to be a pentium 2 system) and there was no problem.

Deksor wrote on 2021-07-27, 10:54:

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't […]
Show full quote

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't have a 1/4 divider for PCI.

This means pci runs at 44.3MHz, including the Southbridge which is also connected to the PCI bus. 44MHz is too high for many PCI devices.

As for the tdp, in the manual I've found the max CPU listed is pentium 3 500.
Now to me it doesn't mean it's not compatible with newer CPUs but rather that the documentation was produced when newer CPUs didn't exist

The tdp of the maximum listed CPU (p3 500) is 28W.
The tdp of the pentium 3 850eb is 22W (probably due to the lower voltage). Correct me if I'm wrong, but this looks rather good. I think you'd be fine even with a 1GHz pentium 3 (which I did try on my bh6 although not for long because I wanted it to be a pentium 2 system) and there was no problem.

IIRC, the 1G P3 has another stepping, that will require another BIOS update, that might not be available.

CC0 Pentium III vs CD0 P3 ... CPUID 6-8-10... are there real differences???

Has the stepping discussion

Deksor wrote on 2021-07-27, 10:54:

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't […]
Show full quote

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't have a 1/4 divider for PCI.

This means pci runs at 44.3MHz, including the Southbridge which is also connected to the PCI bus. 44MHz is too high for many PCI devices.

As for the tdp, in the manual I've found the max CPU listed is pentium 3 500.
Now to me it doesn't mean it's not compatible with newer CPUs but rather that the documentation was produced when newer CPUs didn't exist

The tdp of the maximum listed CPU (p3 500) is 28W.
The tdp of the pentium 3 850eb is 22W (probably due to the lower voltage). Correct me if I'm wrong, but this looks rather good. I think you'd be fine even with a 1GHz pentium 3 (which I did try on my bh6 although not for long because I wanted it to be a pentium 2 system) and there was no problem.

Besides the TDP there is the amount of current drawn (which is discussed in one of the forum discussions posted above)
https://groups.google.com/g/alt.comp.periphs. … _kEGEjCwQ?pli=1
that is important. Following the info from Intel VRM spec sheets it looks as if the PIII 850 is not very far outside of the 'safe zone' anyway

.

svfn wrote on 2021-07-27, 10:44:

BTW was it not possible to change your FSB to 100 instead of 133 with your P3 600E? 133 MHz FSB CPUs cannot run at lower FSB?

It runs fine at default 100fsb, I just wanted to see if I could get it overclocked, hence the push to 133fsb (I could do the 112 or 124fsb too, but this is mainly for fun after all).

I had a look at my board, it looks like I have the older 45N03, yet the 600E runs perfectly happily at stock.

Deksor wrote on 2021-07-27, 10:54:

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't […]
Show full quote

I think a plausible cause of the instability has nothing to do with the vrms but rather because the 1.01 rev of the BH6 doesn't have a 1/4 divider for PCI.

This means pci runs at 44.3MHz, including the Southbridge which is also connected to the PCI bus. 44MHz is too high for many PCI devices.

As for the tdp, in the manual I've found the max CPU listed is pentium 3 500.
Now to me it doesn't mean it's not compatible with newer CPUs but rather that the documentation was produced when newer CPUs didn't exist

The tdp of the maximum listed CPU (p3 500) is 28W.
The tdp of the pentium 3 850eb is 22W (probably due to the lower voltage). Correct me if I'm wrong, but this looks rather good. I think you'd be fine even with a 1GHz pentium 3 (which I did try on my bh6 although not for long because I wanted it to be a pentium 2 system) and there was no problem.

I did run this board with the semi-famous Celeron 566 @ 133fsb as my main computer for a long time with no stability issues back in the day. I guess something could have aged enough to where it is presenting issues now when it comes to the southbridge tolerance for overclock. Now that I think about it I must have had a PCI sound card in it too that tolerated the PCI overclock (I think it was a Monster Sound MX300). The ISA sound card I have in the machine now is the first ISA card this board has ever had in it.

In any event, at the right price I may just pick the cpu up and give it a try either way.

Ah I didn't read that you overclocked it. According to soggi's site, Coppermines below 700 mhz is fine so I assume that up to 650, 667 should be ok, though the BIOS only officially states support added for PIII 600. It's the faster Coppermines that I am unsure about.

Since P3 1Ghz slot 1 CPUs are hard to find and costly, I wonder if slotket is a better way, though finding models that support Coppermines is also hard haha, but then S370 1Ghz CPUs would be plenty.

Is it the exact same board you had back in the days or did you buy another one ? The rev 1.1 of this board has a 1/4 divider.

As for the microcode, even though the latest bios rev don't mention the p3 1GHz, I can tell you that it's supported because I ran it on my board (same rev and latest bios).

This is the exact same board I had back in the day. It's been sitting in my old closet at my parent's house for 20 years along with the two CPUs I had on it back then (first a Slot 1 Celeron 300A and then a socket-370 Celeron566 on an Abit slotket).

There are a couple of Slot 1 P3-850s on ebay right now, I may buy one just to see if it works.

teiresias wrote on 2021-07-27, 13:05:

I did run this board with the semi-famous Celeron 566 @ 133fsb as my main computer for a long time with no stability issues back in the day. .

So, 1133 on a 566? Pretty impressive! Must have been a late stepping?

H3nrik V! wrote on 2021-07-27, 16:08:

So, 1133 on a 566? Pretty impressive! Must have been a late stepping?

Actually no, that was a 66fsb part wasn't it? So I must have upped that to 100fsb to get 850 on its 8.5 multiplier, so the bus overclocking wouldn't have come into play anyway.

teiresias wrote on 2021-07-27, 16:18:

H3nrik V! wrote on 2021-07-27, 16:08:

So, 1133 on a 566? Pretty impressive! Must have been a late stepping?

Actually no, that was a 66fsb part wasn't it? So I must have upped that to 100fsb to get 850 on its 8.5 multiplier, so the bus overclocking wouldn't have come into play anyway.

Yeah, seems more plausible. I had a 600 running 900, those were the days 😀

Just a quick update. I was able to get a P3-850 SL4BW off of ebay, and it works just fine on my BH6. It wasn't able to autodetect the settings on it, but I just had to manually set the FSB and it's been running for a few hours without a hitch.

I ran my standard batch run of 3dMark99 and it actually has a slightly lower 3dMark score than when I was running the P3-600E overclocked to 800Mhz (paired with an FX5600 Ultra). I'm guessing the difference is maybe down to the RAM running faster with the 133fsb on the overclocked 600E, but seeing as how the overclock was unstable it seems a small price to pay.

teiresias wrote on 2021-07-31, 18:10:

Just a quick update. I was able to get a P3-850 SL4BW off of ebay, and it works just fine on my BH6. It wasn't able to autodetect the settings on it, but I just had to manually set the FSB and it's been running for a few hours without a hitch.

I ran my standard batch run of 3dMark99 and it actually has a slightly lower 3dMark score than when I was running the P3-600E overclocked to 800Mhz (paired with an FX5600 Ultra). I'm guessing the difference is maybe down to the RAM running faster with the 133fsb on the overclocked 600E, but seeing as how the overclock was unstable it seems a small price to pay.

nice! thanks for the update. at least not stuck with P3 600/650 and can go for higher, i wonder if it can go 1ghz, but slot 1 1ghz is pretty rare.

There is negligible price difference between PIII 1Ghz and 800Mhz. I would advise to buy 1Ghz and underclock it if the board isn't fully stable. It has lower multiplier than PIII 850Mhz which is a good thing. Based on the TDP list posted above it's clear boards that support Katmai 500 and support 1.7V should be fine with at least PIII 900Mhz. Many old boards supported up to 550-600Mhz Katmai so running 1Ghz PIII should not be a problem on them.

My current CPU is also an underclocked PIII 1Ghz as I have no time to verify stability with higher FSB and 750Mhz is enough for now.

I just couldn't find a 1GHz slot1 that was in a price range I was willing to pay. Right now it looks like the top results on ebay for a Slot-1 1GHz are in the $200 buy-it-now range, whereas the 850 I bought was $50 with a bit extra cost to buy a new slot-1 cooler to replace the one it came with (which looked like it was built to work with some proprietary ducting system in a pre-built system).

Oddly, just yesterday for some reason the PC stopped booting (no post beep, no video) but everything powered on. On a hunch, I plugged the old 600E back into the slot and it booted right up. It seems for some reason the BH6 had lost the manual voltage setting and had gone back to 1.65 from the 1.7 I'd set it to (to support the 850 CPU). Once I was able to boot with the 600E I set the voltage back, swapped the CPUs again and the 850 booted right up. I guess I'll be keeping at least one lower voltage slot 1 CPU around (not that i tend to sell these things anyway, 🤣).