Anonymous Coward wrote on 2022-06-02, 15:08:

The overdrive DX/2-40 was relatively common, but the plain version was indeed pretty rare. That was meant for portable devices, was it not? If that's the case, you'd only find them if you were salvaging laptops. At the time, a DX/2-40 on the desktop made close to no sense. But, then we have chips like the SX-16...

Dx2-40 odpr made zero sense even if common
Might as well have a dx2-33 or dx4-50

Also I’m not aware of any motherboards with the sx20 that had a locked bus speed but I guess anything is possible and yeah, socketed oscillators in 486 boards of that vintage were a thing, I had a lot of old Micronics ISA 486s that were apparently upgraded from sx25’s to dx2-50s during their life, board had a socketed oscillator and was rated to 50mpg FSB, tech could have upgraded these to dx2-66’s if they wanted .

In so far as sx16’s at least they all overclocked to 50mhz+

Cause jumpers are hard

DX/2-40 ODPR makes zero sense to people like us, but I'm sure it must have made at least a little sense as a drop in upgrade for corporate buyers. Or maybe the US government...who knows. The whole concept of 487SX and 5V overdrives was pretty much retarded from the get go.

It's not exactly rare, but I once met a guy who recreated an 8086 or 8088 using bread board a bunch of wires, transistors (or relays, I don't know), the thing was the size of the entire table.

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

Tetrium wrote on 2022-06-03, 07:21:

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

According to cpuworld the fastest p3 with 100 mhz fsb is the 1000 mhz (there is also a 1100 probably just an ES?)

Nemo1985 wrote on 2022-06-03, 08:39:

Tetrium wrote on 2022-06-03, 07:21:

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

According to cpuworld the fastest p3 with 100 mhz fsb is the 1000 mhz (there is also a 1100 probably just an ES?)

https://www.cpu-world.com/sspec/SL/SL5QW.html

I can confirm the existence of the PIII-1100 @100MHz FSB (SL5QW).
It was the last PIII I used and I had it run @1200 MHz for a long time without problems. Bought it in 2002 and still have it 😀
If it's some kind of rare I cannot say.

I also recently came across a PIII 1133 (SL5B2), a standard Coppermine in FCPGA-2 package (the one with heatspreader, just like the later Tualatins).
Don't know either if that's a rare one. Always thought they were recalled by Intel because of stability issues. But not sure if that was only a thing with the Slot 1 parts.

Edit: just in case, the Coppermine-T (as opposed to the standard Coppermine) @1,13GHz has sSpec SL5QK.

I don't think SL5B2 was the recalled one. I think it was SL4HH. I think it was also an SECC2 Slot 1 CPU, not S370.
Those should indeed be pretty rare. I think all or most of them went to IBM and Dell. Dell never shipped, so probably all of the bad 1.13s went back to Intel.
IBM was the other big customer, and they apparently did ship machines with the CPU. Probably a lot of them were recalled, but if you're going to find one it'll probably come from an IBM.

Anonymous Coward wrote on 2022-06-03, 15:06:

I don't think SL5B2 was the recalled one. I think it was SL4HH. I think it was also an SECC2 Slot 1 CPU, not S370.
Those should indeed be pretty rare. I think all or most of them went to IBM and Dell. Dell never shipped, so probably all of the bad 1.13s went back to Intel.
IBM was the other big customer, and they apparently did ship machines with the CPU. Probably a lot of them were recalled, but if you're going to find one it'll probably come from an IBM.

It was indeed the SL4HH slot 1 cpu with stepping cC0 that was recalled.
https://www.anandtech.com/show/594/3
As far as the Intel records show there were no s370 cpu's @ 1.13Ghz issued in 2000 - these came a year later and were all from the cD0 stepping.

Tetrium wrote on 2022-06-03, 07:21:

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

You're right I didn't think of that!!! I'll whip out my Abit BE6-II and see if any of my cheap passive slotkets will work with these chips. I don't know if I'll leave a SL5QW in the BE6 rig tough... I've built 4 "GHz" class PC's already.... two 1.4 P3-s tualatins, a 1GHz coppermine mATX compact build built on an Asus CUVX-M and the dumpster find NEC PowerMate 1100Mhz tualeron + Asus TUSI-M PC I built a couple of weeks ago.

I would like to compare the performance of the SL5QW + Abit BE6-II against the 1.1Ghz Tualeron + Asus TUSI-M and the Asus CUVX-M + 1Ghz Coppermine with 133MHz FSB. It might be interesting.

Thanks for the great idea Tetrium! I was wondering what to do with my sunday afternoon 😀

rmay635703 wrote on 2022-06-02, 15:13:

Dx2-40 odpr made zero sense even if common Might as well have a dx2-33 or dx4-50 […]
Show full quote

Anonymous Coward wrote on 2022-06-02, 15:08:

The overdrive DX/2-40 was relatively common, but the plain version was indeed pretty rare. That was meant for portable devices, was it not? If that's the case, you'd only find them if you were salvaging laptops. At the time, a DX/2-40 on the desktop made close to no sense. But, then we have chips like the SX-16...

Dx2-40 odpr made zero sense even if common
Might as well have a dx2-33 or dx4-50

Also I’m not aware of any motherboards with the sx20 that had a locked bus speed but I guess anything is possible and yeah, socketed oscillators in 486 boards of that vintage were a thing, I had a lot of old Micronics ISA 486s that were apparently upgraded from sx25’s to dx2-50s during their life, board had a socketed oscillator and was rated to 50mpg FSB, tech could have upgraded these to dx2-66’s if they wanted .

In so far as sx16’s at least they all overclocked to 50mhz+

Cause jumpers are hard

Maybe all of these were remarks 😜
I mean this would point towards all SX CPUs being overclockable to 50+ MHz. Are they really all overclockable to 50+MHz? I mean they can't even be overvolted since they already run at 5v.

Socket3 wrote on 2022-06-03, 18:30:

You're right I didn't think of that!!! I'll whip out my Abit BE6-II and see if any of my cheap passive slotkets will work with t […]
Show full quote

Tetrium wrote on 2022-06-03, 07:21:

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

You're right I didn't think of that!!! I'll whip out my Abit BE6-II and see if any of my cheap passive slotkets will work with these chips. I don't know if I'll leave a SL5QW in the BE6 rig tough... I've built 4 "GHz" class PC's already.... two 1.4 P3-s tualatins, a 1GHz coppermine mATX compact build built on an Asus CUVX-M and the dumpster find NEC PowerMate 1100Mhz tualeron + Asus TUSI-M PC I built a couple of weeks ago.

I would like to compare the performance of the SL5QW + Abit BE6-II against the 1.1Ghz Tualeron + Asus TUSI-M and the Asus CUVX-M + 1Ghz Coppermine with 133MHz FSB. It might be interesting.

Thanks for the great idea Tetrium! I was wondering what to do with my sunday afternoon 😀

I'd definitely like to learn the results of a Coppermine 1100 vs Tualeron 1100 comparison 🙂

ChrisK wrote on 2022-06-03, 13:37:

I can confirm the existence of the PIII-1100 @100MHz FSB (SL5QW). It was the last PIII I used and I had it run @1200 MHz for a l […]
Show full quote

I can confirm the existence of the PIII-1100 @100MHz FSB (SL5QW).
It was the last PIII I used and I had it run @1200 MHz for a long time without problems. Bought it in 2002 and still have it 😀
If it's some kind of rare I cannot say.

I also recently came across a PIII 1133 (SL5B2), a standard Coppermine in FCPGA-2 package (the one with heatspreader, just like the later Tualatins).
Don't know either if that's a rare one. Always thought they were recalled by Intel because of stability issues. But not sure if that was only a thing with the Slot 1 parts.

Edit: just in case, the Coppermine-T (as opposed to the standard Coppermine) @1,13GHz has sSpec SL5QK.

Coppermine-T is an interesting chip. Iirc it was made for a feature that ended up never getting implemented.
FC-PGA2 Coppermines aren't rare, I'd say they are more uncommon. There's also Celeron (Coppermine variety) with the IHS, from top of my head from 850MHz and up most speedgrades have FC-PGA2 variants. Whether or not this is of any real advantage is hard to say. People who are too afraid to crack a die installing CPU HSFs may prefer these but the disadvantage is the increased height of the CPU which significantly increases risk of breakage of the socket tabs if one is careless.

But anyway, from a practical PoV, to us as plain end users it shouldn't matter if the chip is a Coppermine or a Coppermine-T, same performance, practically same thermal properties, same compatibility afaik.
FC-PGA and FC-PGA2 makes more of a practical difference due to the heatspreader thingy.

What feature?

https://tweakers.net/nieuws/15480/intel-cance … -processor.html

Relevant text (because x-bit labs original link is now defunct and this is the only bit of info regarding this remaining on tweakers)

According to the latest Intel’s plans update we saw here, the company decided to cancel Coppermine-T processor. At first this CPU was planned to be a transitional solution between the regular Pentium III processors and new Pentium III built on a 0.13 micron Tualatin core. It was supposed to be distinguished from the ordinary 0.18 micron Coppermine by the support of 1.2V AGTL+ bus used in Tualatin as well. So, Coppermine-T was the youngest and the cheapest processor supported by Almador, which Intel aimed particularly for use with Tualatin processors.

However, now that Almador is canceled Intel has changed its plans concerning Coppermine-T processor. since Tualatin will now be supported by a new i815 chipset version (B-step), which will be compatible with Coppermine, there will be no real need in Coppermine-T CPU. Instead Intel will simply launch a regular Pentium III Coppermine with new D0 core stepping. The CPUs with a renewed core will feature higher Vcore (up to 1.75V) and hence will be able to work at 1.13GHz, the top rate for Coppermine processors.

Apparently Intel calcelled Almador (cancelled chipset?) for which Coppermine-T was supposed to be the slowest supported CPU. I'd guess Intel wanted to purposefully make Coppermine incompatible with it?

EDIT:
https://tweakers.net/nieuws/15982/intel-almad … t-plaatjes.html
Apparently i830 chipset with some kind of special RDRAM-based module which was supposedly gonna be called a M-RIMM or a "GFX Performance Module".

Tetrium wrote on 2022-06-03, 07:21:

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... 🙁 Here's what I was able to do today:

The base test system is a Abit BE6-II board / Pentium II 333 @ 500MHz / 256MB PC133 SDRAM @ 100MHz CL2 / Creative 3D Blaster Annihilator 2 PRO / 20Gb Quantum Fireball HDD / 350W HP PSU with -5V / Creative AWE64. What I did first is swap the P2-333 for a slot 1 P3-1000 (FSB133 unfortunatly) I had in my CPU box just to check if the board would run correctly with a coppermine CPU - so far so good, the system posted, and I was able to run at 750MHz (7.5x100) stable (tested in 3dmark 2000 and quake3 - several runs). I did try 1GHz (133x7.5) but the Geforce 2 didn't like that and would bsod when loading windows. Then I got sidetracked and tried my new Chinese Geforce FX5500 PCI witch worked a treat.

After loosing an hour on that, I started going trough my slotket box, but it's pretty slim pickings in there. I don't have anything as fancy as a Tualatin compatible slotket, but I do have a couple of coppermine ready slotkets - specifically something silkscreened with AA370TS. Don't know the manufacturer, but one was in an AT form factor slot 1 board with a VIA 694 chipset running a 850MHz SL5Q Coppermine Celeron while the other came with my Asus P3B and a 750MHz / FSB 100 SL462 Pentium 3.

I tried the slotket with the 1100Mhz P3 and no dice. The BE6 would not post. So I reset CMOS and installed the 750Mhz SL462 in the slotket. Still no POST. I took the slotket out, washed the goldfinger with IPA and ran it over with a bit of 1000 grit sand paper, then washed it again with IPA and the BE6 poseted. I was able to boot into windows and run Quake 3 and 3DMark 2000. Results were in line with the slot 1 P3 1000 running at 750Mhz.

I then moved on to the SL5QW, and the system still refused to POST. Only after swapping the RAM stick for a double sided PC133 stick I was able to post.

I was able to boot into windows... and the PC froze. Tired it a couple of times. The second time it froze before loading desktop icons and the 3rd time it froze at the win98 loading animation. I rebooted, went into bios and checked voltage and temps.... the CPU was already at 55 That's a bit hot for a pentium 3 as far as i know 🙁 . I then bumped up the voltage from 1.75v to 1.8, and was able to load windows, but as soon as I launched 3DMark it would freeze again. This time BIOS was showing 58-60C.... So I stopped testing.

I'm guessing either my Abit doesn't like the chip, the chip is overheating with that dingy cooler, or my slotket is crap. After testing the SL5QW in my Asus P3B and getting similar results I was leaning towards the latter... since the chip worked fine in my MSI master dual socket 370 board and in my Asus CUVX-M. But it did so with MUCH BEEFEER cooling.... in the CUVX-M at least I used an 80mm socket A cooler, otherwise it would climb up to 68C while gaming...

As to not de-rail this thread, I'll open a new one after I have time to do a bit more investigation.

Socket3 wrote on 2022-06-04, 19:21:

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... : […]
Show full quote

Tetrium wrote on 2022-06-03, 07:21:

Socket3 wrote on 2022-06-02, 18:39:

It seems like my memory isn't what it used to be. The two "1133MHz" pentium 3 coppermine chips I claimed to have are in fact 1100MHz SL5QW Coppermine Pentium 3's. 100Mhz FSB too. Weird chips.

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... 🙁 Here's what I was able to do today:

The base test system is a Abit BE6-II board / Pentium II 333 @ 500MHz / 256MB PC133 SDRAM @ 100MHz CL2 / Creative 3D Blaster Annihilator 2 PRO / 20Gb Quantum Fireball HDD / 350W HP PSU with -5V / Creative AWE64. What I did first is swap the P2-333 for a slot 1 P3-1000 (FSB133 unfortunatly) I had in my CPU box just to check if the board would run correctly with a coppermine CPU - so far so good, the system posted, and I was able to run at 750MHz (7.5x100) stable (tested in 3dmark 2000 and quake3 - several runs). I did try 1GHz (133x7.5) but the Geforce 2 didn't like that and would bsod when loading windows. Then I got sidetracked and tried my new Chinese Geforce FX5500 PCI witch worked a treat.

sl5q.jpeg

After loosing an hour on that, I started going trough my slotket box, but it's pretty slim pickings in there. I don't have anything as fancy as a Tualatin compatible slotket, but I do have a couple of coppermine ready slotkets - specifically something silkscreened with AA370TS. Don't know the manufacturer, but one was in an AT form factor slot 1 board with a VIA 694 chipset running a 850MHz SL5Q Coppermine Celeron while the other came with my Asus P3B and a 750MHz / FSB 100 SL462 Pentium 3.

slotket n cooler.jpeg

I tried the slotket with the 1100Mhz P3 and no dice. The BE6 would not post. So I reset CMOS and installed the 750Mhz SL462 in the slotket. Still no POST. I took the slotket out, washed the goldfinger with IPA and ran it over with a bit of 1000 grit sand paper, then washed it again with IPA and the BE6 poseted. I was able to boot into windows and run Quake 3 and 3DMark 2000. Results were in line with the slot 1 P3 1000 running at 750Mhz.

I then moved on to the SL5QW, and the system still refused to POST. Only after swapping the RAM stick for a double sided PC133 stick I was able to post.

post 1100.jpeg

I was able to boot into windows... and the PC froze. Tired it a couple of times. The second time it froze before loading desktop icons and the 3rd time it froze at the win98 loading animation. I rebooted, went into bios and checked voltage and temps.... the CPU was already at 55 That's a bit hot for a pentium 3 as far as i know 🙁 . I then bumped up the voltage from 1.75v to 1.8, and was able to load windows, but as soon as I launched 3DMark it would freeze again. This time BIOS was showing 58-60C.... So I stopped testing.

I'm guessing either my Abit doesn't like the chip, the chip is overheating with that dingy cooler, or my slotket is crap. After testing the SL5QW in my Asus P3B and getting similar results I was leaning towards the latter... since the chip worked fine in my MSI master dual socket 370 board and in my Asus CUVX-M. But it did so with MUCH BEEFEER cooling.... in the CUVX-M at least I used an 80mm socket A cooler, otherwise it would climb up to 68C while gaming...

As to not de-rail this thread, I'll open a new one after I have time to do a bit more investigation.

Just a quick reply first, if you are trying with that black Intel stock heatsink please stop as (if this HSF is what I think it is) this HSF is insufficient for the TDP of a 1100MHz Coppermine.

I will edit this reply as I read the rest but important stuff first of course! 😜

EDIT: Yes, please open a new thread and do let me know 😀

I think you're correct in also assuming that HSF is insufficient. It's still a nice HSF for ss7 (provided it is beefy enough for the particular CPU used) though due to the shape of the heatsink.

It could also be the slotket, it may not be able to cope with the combination of a fast CPU along with the board perhaps having trouble with the power delivery. For all we know it could be the caps or due to the design of the board (it may not be build for such a CPU).

Tetrium wrote on 2022-06-04, 19:38:

Just a quick reply first, if you are trying with that black Intel stock heatsink please stop as (if this HSF is what I think it […]
Show full quote

Socket3 wrote on 2022-06-04, 19:21:

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... : […]
Show full quote

Tetrium wrote on 2022-06-03, 07:21:

And perhaps even better chips in a way. These would make for some awesome 440BX upgrades 😋
I think the fastest P3 @ 100MHz FSB I have is like 850MHz or so?

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... 🙁 Here's what I was able to do today:

The base test system is a Abit BE6-II board / Pentium II 333 @ 500MHz / 256MB PC133 SDRAM @ 100MHz CL2 / Creative 3D Blaster Annihilator 2 PRO / 20Gb Quantum Fireball HDD / 350W HP PSU with -5V / Creative AWE64. What I did first is swap the P2-333 for a slot 1 P3-1000 (FSB133 unfortunatly) I had in my CPU box just to check if the board would run correctly with a coppermine CPU - so far so good, the system posted, and I was able to run at 750MHz (7.5x100) stable (tested in 3dmark 2000 and quake3 - several runs). I did try 1GHz (133x7.5) but the Geforce 2 didn't like that and would bsod when loading windows. Then I got sidetracked and tried my new Chinese Geforce FX5500 PCI witch worked a treat.

sl5q.jpeg

After loosing an hour on that, I started going trough my slotket box, but it's pretty slim pickings in there. I don't have anything as fancy as a Tualatin compatible slotket, but I do have a couple of coppermine ready slotkets - specifically something silkscreened with AA370TS. Don't know the manufacturer, but one was in an AT form factor slot 1 board with a VIA 694 chipset running a 850MHz SL5Q Coppermine Celeron while the other came with my Asus P3B and a 750MHz / FSB 100 SL462 Pentium 3.

slotket n cooler.jpeg

I tried the slotket with the 1100Mhz P3 and no dice. The BE6 would not post. So I reset CMOS and installed the 750Mhz SL462 in the slotket. Still no POST. I took the slotket out, washed the goldfinger with IPA and ran it over with a bit of 1000 grit sand paper, then washed it again with IPA and the BE6 poseted. I was able to boot into windows and run Quake 3 and 3DMark 2000. Results were in line with the slot 1 P3 1000 running at 750Mhz.

I then moved on to the SL5QW, and the system still refused to POST. Only after swapping the RAM stick for a double sided PC133 stick I was able to post.

post 1100.jpeg

I was able to boot into windows... and the PC froze. Tired it a couple of times. The second time it froze before loading desktop icons and the 3rd time it froze at the win98 loading animation. I rebooted, went into bios and checked voltage and temps.... the CPU was already at 55 That's a bit hot for a pentium 3 as far as i know 🙁 . I then bumped up the voltage from 1.75v to 1.8, and was able to load windows, but as soon as I launched 3DMark it would freeze again. This time BIOS was showing 58-60C.... So I stopped testing.

I'm guessing either my Abit doesn't like the chip, the chip is overheating with that dingy cooler, or my slotket is crap. After testing the SL5QW in my Asus P3B and getting similar results I was leaning towards the latter... since the chip worked fine in my MSI master dual socket 370 board and in my Asus CUVX-M. But it did so with MUCH BEEFEER cooling.... in the CUVX-M at least I used an 80mm socket A cooler, otherwise it would climb up to 68C while gaming...

As to not de-rail this thread, I'll open a new one after I have time to do a bit more investigation.

Just a quick reply first, if you are trying with that black Intel stock heatsink please stop as (if this HSF is what I think it is) this HSF is insufficient for the TDP of a 1100MHz Coppermine.

I will edit this reply as I read the rest but important stuff first of course! 😜

EDIT: Yes, please open a new thread and do let me know 😀

I think you're correct in also assuming that HSF is insufficient. It's still a nice HSF for ss7 (provided it is beefy enough for the particular CPU used) though due to the shape of the heatsink.

It could also be the slotket, it may not be able to cope with the combination of a fast CPU along with the board perhaps having trouble with the power delivery. For all we know it could be the caps or due to the design of the board (it may not be build for such a CPU).

I read somewhere that the 1100Mhz adn 1133Mhz Coppermine CPUs are 49 to 55W. That's a lot of power for socket 370. I agree about he cooler, but I don't have anything bigger that will actually fit a slotket. On the BE6 a taller heatsink will hit the ram slots, and on the P3B a wider heatsink will touch the northbridge heatsink causing the slotket to sit at an angle.... I have a pentium 3 1000 box cooler, but that won't fit either - it's too wide and it overlaps the edge of the slotket. Most socket A coolers won't fit either, being either too tall or too wide.... Another reason I dislike slotkets.

In other news, I managed to get a 1100Mhz tualatin to post on my Abit BE6-II. I did it by pulling two jumpers off the slotket - JP1 and JP2. These are labeled as "CPU type" on the slotket, with settings for Coppermine, Mendocino and VIA C3 chips. It seems that removing them completely has the same effect as pin modding a tualatin CPU. I think... in any case, the 1100Mhz Tualeron works on my BE6-II. It POSTS with "Unknown CPU type 726Mhz", but after entering BIOS I can select FSB 100 and drop the voltage down from 1.65 to 1.45v and the system is perfectly stable.... post message also changes to "Pentium 3 1100MHz" just like the picture in my other post. I even managed to post wit a 1000mhz tualeron at 1333mhz (133x10), but it was unstable regardless of voltage and I went up to 1.65v. Similar symptoms as with the SL5QW... And this particular 1GHz tualeron has been stable at 1333Mhz stock voltage on every socket 370 board I've put it in so far. This leads me to believe the stability issues are either linked to mainboard power delivery or signaling quality when using a slotket.

The dingy black intel stock cooler is actually meant for coppermine celerons up to 700Mhz, but it works a treat with the 750Mhz coppermine P3 as well, In never went over 51C during testing witch I believe is perfectly acceptable. It is the slightly taller version of the black intel skt 370 box cooler, not the short one that came with Mendocino celerons but yeah it's clearly insufficient for the 50w 1100Mhz P3

Socket3 wrote on 2022-06-04, 20:57:

I read somewhere that the 1100Mhz adn 1133Mhz Coppermine CPUs are 49 to 55W. That's a lot of power for socket 370. I agree about […]
Show full quote

Tetrium wrote on 2022-06-04, 19:38:

Just a quick reply first, if you are trying with that black Intel stock heatsink please stop as (if this HSF is what I think it […]
Show full quote

Socket3 wrote on 2022-06-04, 19:21:

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... : […]
Show full quote

Ok so I tried to get the SL5QW to run in a 440BX board with a slotket as per your recommendation but I didn't get very far.... 🙁 Here's what I was able to do today:

The base test system is a Abit BE6-II board / Pentium II 333 @ 500MHz / 256MB PC133 SDRAM @ 100MHz CL2 / Creative 3D Blaster Annihilator 2 PRO / 20Gb Quantum Fireball HDD / 350W HP PSU with -5V / Creative AWE64. What I did first is swap the P2-333 for a slot 1 P3-1000 (FSB133 unfortunatly) I had in my CPU box just to check if the board would run correctly with a coppermine CPU - so far so good, the system posted, and I was able to run at 750MHz (7.5x100) stable (tested in 3dmark 2000 and quake3 - several runs). I did try 1GHz (133x7.5) but the Geforce 2 didn't like that and would bsod when loading windows. Then I got sidetracked and tried my new Chinese Geforce FX5500 PCI witch worked a treat.

sl5q.jpeg

After loosing an hour on that, I started going trough my slotket box, but it's pretty slim pickings in there. I don't have anything as fancy as a Tualatin compatible slotket, but I do have a couple of coppermine ready slotkets - specifically something silkscreened with AA370TS. Don't know the manufacturer, but one was in an AT form factor slot 1 board with a VIA 694 chipset running a 850MHz SL5Q Coppermine Celeron while the other came with my Asus P3B and a 750MHz / FSB 100 SL462 Pentium 3.

slotket n cooler.jpeg

I tried the slotket with the 1100Mhz P3 and no dice. The BE6 would not post. So I reset CMOS and installed the 750Mhz SL462 in the slotket. Still no POST. I took the slotket out, washed the goldfinger with IPA and ran it over with a bit of 1000 grit sand paper, then washed it again with IPA and the BE6 poseted. I was able to boot into windows and run Quake 3 and 3DMark 2000. Results were in line with the slot 1 P3 1000 running at 750Mhz.

I then moved on to the SL5QW, and the system still refused to POST. Only after swapping the RAM stick for a double sided PC133 stick I was able to post.

post 1100.jpeg

I was able to boot into windows... and the PC froze. Tired it a couple of times. The second time it froze before loading desktop icons and the 3rd time it froze at the win98 loading animation. I rebooted, went into bios and checked voltage and temps.... the CPU was already at 55 That's a bit hot for a pentium 3 as far as i know 🙁 . I then bumped up the voltage from 1.75v to 1.8, and was able to load windows, but as soon as I launched 3DMark it would freeze again. This time BIOS was showing 58-60C.... So I stopped testing.

I'm guessing either my Abit doesn't like the chip, the chip is overheating with that dingy cooler, or my slotket is crap. After testing the SL5QW in my Asus P3B and getting similar results I was leaning towards the latter... since the chip worked fine in my MSI master dual socket 370 board and in my Asus CUVX-M. But it did so with MUCH BEEFEER cooling.... in the CUVX-M at least I used an 80mm socket A cooler, otherwise it would climb up to 68C while gaming...

As to not de-rail this thread, I'll open a new one after I have time to do a bit more investigation.

Just a quick reply first, if you are trying with that black Intel stock heatsink please stop as (if this HSF is what I think it is) this HSF is insufficient for the TDP of a 1100MHz Coppermine.

I will edit this reply as I read the rest but important stuff first of course! 😜

EDIT: Yes, please open a new thread and do let me know 😀

I think you're correct in also assuming that HSF is insufficient. It's still a nice HSF for ss7 (provided it is beefy enough for the particular CPU used) though due to the shape of the heatsink.

It could also be the slotket, it may not be able to cope with the combination of a fast CPU along with the board perhaps having trouble with the power delivery. For all we know it could be the caps or due to the design of the board (it may not be build for such a CPU).

I read somewhere that the 1100Mhz adn 1133Mhz Coppermine CPUs are 49 to 55W. That's a lot of power for socket 370. I agree about he cooler, but I don't have anything bigger that will actually fit a slotket. On the BE6 a taller heatsink will hit the ram slots, and on the P3B a wider heatsink will touch the northbridge heatsink causing the slotket to sit at an angle.... I have a pentium 3 1000 box cooler, but that won't fit either - it's too wide and it overlaps the edge of the slotket. Most socket A coolers won't fit either, being either too tall or too wide.... Another reason I dislike slotkets.

In other news, I managed to get a 1100Mhz tualatin to post on my Abit BE6-II. I did it by pulling two jumpers off the slotket - JP1 and JP2. These are labeled as "CPU type" on the slotket, with settings for Coppermine, Mendocino and VIA C3 chips. It seems that removing them completely has the same effect as pin modding a tualatin CPU. I think... in any case, the 1100Mhz Tualeron works on my BE6-II. It POSTS with "Unknown CPU type 726Mhz", but after entering BIOS I can select FSB 100 and drop the voltage down from 1.65 to 1.45v and the system is perfectly stable.... post message also changes to "Pentium 3 1100MHz" just like the picture in my other post. I even managed to post wit a 1000mhz tualeron at 1333mhz (133x10), but it was unstable regardless of voltage and I went up to 1.65v. Similar symptoms as with the SL5QW... And this particular 1GHz tualeron has been stable at 1333Mhz stock voltage on every socket 370 board I've put it in so far. This leads me to believe the stability issues are either linked to mainboard power delivery or signaling quality when using a slotket.

chips (no fish).jpeg

test rig.jpeg

The dingy black intel stock cooler is actually meant for coppermine celerons up to 700Mhz, but it works a treat with the 750Mhz coppermine P3 as well, In never went over 51C during testing witch I believe is perfectly acceptable. It is the slightly taller version of the black intel skt 370 box cooler, not the short one that came with Mendocino celerons but yeah it's clearly insufficient for the 50w 1100Mhz P3

The Tualeron working is really interesting!
What slotket are you using? Got any clear pics of it?
Removing both jumpers and the Tualeron works, nice. I definitely gotta try that at some point in time ^^

There are similarly shaped (usually metal colored) heatsinks that have the same dimensions but are better because they have a thicker base from where the fins extrude.
I typically replace the stock fan with one I cannibalized from a s7 Coolermaster HSF (I bought several of those years ago NIB) but any with decent airflow for the size should work.
I also think 51c is acceptable, especially if this is after prolonged gaming for instance. Coppermine has a fairly low TDP which only starts getting high after 800MHz or so. The ones slower than 800MHz typically run quite cool, which means you can get away with rather poor cooling relatively speaking.

I don't know if the 1100MHz part has a max TDP of 49W? I'd assume it to be more around 35W, 40W tops.
This webpage lists even slightly lower, but roughly similar values for TDP
https://www.pchardwarelinks.com/elec_pentium.htm