1.4 GHz Tualatin Celeron runs, but is unstable in Prime95

[bdwilcox]

I stacked a Tualatin 1.4 GHz Tualatin Celeron on top of an Upgradeware 370GU on top of a Powerleap NeoS370. I tried it in both CPU sockets and with voltages ranging from 1.5 to 1.6 volts. I also tried the RU and RV BIOSes as well as two different sticks of RAM.

In all instances the system boots and runs fine in both Windows 98se and 2000 SP4, but when I try to run the Prime95 Torture Test it runs for about a minute and then fails with a rounding error.

I was wondering if anyone else has experienced this instability with a similar setup and if it possibly relates to the VRMs on my board's revision. I can't find the specific article, but I remember there being some info about bad VRMs on a specific revision of BP-6 boards.

The markings on my board are such:

Back of Board:
V0.41

9935
TEAN TN2 94V-0
FP-236-C41-4B

On ISA slot:
BP6-OKv1.0 937094353-273X
P6 F14 QA
BP6-TAJ004021 MX D

On board near memory slots
3A37 T18

Thanks.

[BCN]

try it on lower FSB and not on 100MHz, it might be just your BX cooling, it might be RAM if it is not PC133 or at least PC100. And then of course it might be VRM...

[bdwilcox]

I'll try it on 66. It ran rock-solid stable with two Celeron 466s at default speed and voltages, so I know the board is OK as far as "in-spec" usage is concerned.

The northbridge chip has active cooling on it (first thing I did when I got the board) so I doubt that's the problem. RAM is PC133 spec and I have it running at CAS 3 even though its spec'd at CAS 2 at 100 MHz.

I'll run it at 66 and see if it's stable. If not, I guess I can deduce that the VRM or caps are at fault. I've watched Motherboard Monitor and voltage rarely dips below 1.49 V at default 1.5V setting, so I don't know what to make of it.

[BCN]

I think it is BX, it is not sufficient in many cases only active cooling when it is at 100+ FSB - then you need an agressive cooling

[hyperspace]

It could be related to the EC10 capacitor problem.

[bdwilcox]

EC10 capacitor problem? I've never heard of it before. Could you elaborate on what this problem is? Do you have any information on how to fix it?

In the meantime, I'll also try BCN's recommendation to make sure the adapters are totally mated flush and report back if stability has improved.

[hyperspace]

Check out this link:

viewtopic.php?t=957

[bdwilcox]

I have a 100uF EC10 cap so that is probably the source of my woe. I don't know how to solder and I don't know anyone who can do soldering this fine. And its just not worth the investment to have a pro do it, especially since it won't guarantee stability. I guess I'll cut my losses and buy a motherboard that can take a Tualatin Celeron 1.4 and eat the cost of the Powerleap and Upgradeware adapters. In the long run it will probably be cheaper and much less risky.

[BCN]

would you leave and not climb mount everest only because it seems high enough?

[InactiveX]

bdwilcox: It would be a shame to give up on the BP6.

If you do decide to go no further, consider putting it up for sale on Classifieds section of this board.

[Dave Rave]

I have 5 bp6 boards, and 4 have the 100uf EC10.
only one of them needed the upgrade to something better, and that one had a damaged 2nd cpu socket, only runs one cpu.

[Holodeck2]

did someone say agressive cooling?

you're pushing the limits of the board, it wasn't built for this, the engineers didn't think there were people weird enough to try to do this. so, you're gonna have to make some power modifications.
It's like climbing everest with a pickax and sneekers

[BCN]

I had no problem with them at 3400m...

though did not have the posibility for higher yet...