PIII 1400s has twice the TDP of PIII 600EB. If you are using a passive slotket it probably won't get enough power. You need an active slotket, Powerleap made some.

CPU and board are running out of spec. Often this works, sometimes it doesn't.
With what little knowledge I have I will assume signal levels from the Tualatin aren't high enough for the board at 100MHz FSB.

RaverX wrote:

PIII 1400s has twice the TDP of PIII 600EB. If you are using a passive slotket it probably won't get enough power. You need an active slotket, Powerleap made some.

While I do not know enough about the boards to confidently rule out what you're saying, the Celeron 600EB has a TDP of 22.1W while the Tualatin-S 1400 has a TDP of 32.2W (it will be higher with increased voltages, probably about 35-36W @1.6V). Both boards support Klamath as well as Katmai cores that go up to ~43W. Amperage on the PIII 600B and Tualatin-S 1400@1,6V are about equal. This leads me to believe the boards should be able to deliver sufficient power to the processor. Please correct me if I'm wrong here.

RaverX wrote:

PIII 1400s has twice the TDP of PIII 600EB. If you are using a passive slotket it probably won't get enough power. You need an active slotket, Powerleap made some.

The CPU behaves same way on both motherboards, one being S370 and one Slot 1. So for that reason it seems the slotket is not the issue.

Doornkaat wrote:

RaverX wrote:

PIII 1400s has twice the TDP of PIII 600EB. If you are using a passive slotket it probably won't get enough power. You need an active slotket, Powerleap made some.

While I do not know enough about the boards to confidently rule out what you're saying, the Celeron 600EB has a TDP of 22.1W while the Tualatin-S 1400 has a TDP of 32.2W (it will be higher with increased voltages, probably about 35-36W @1.6V). Both boards support Klamath as well as Katmai cores that go up to ~43W. Amperage on the PIII 600B and Tualatin-S 1400@1,6V are about equal. This leads me to believe the boards should be able to deliver sufficient power to the processor. Please correct me if I'm wrong here.

Could it be so that both these boards are struggling to give sufficient amperage at the CPU:s native voltage (1,475V). Maybe I should try to override the voltage on the slotket to 1,5 or 1,6V and try again on the Slot 1 board?

SubZero wrote:

Maybe I should try to override the voltage on the slotket to 1,5 or 1,6V and try again on the Slot 1 board?

I would assume the boards can't supply 1,45V anyway so they deliver their minimum voltage but since you can set VID with the slocket it's worth a shot. Maybe increasing core voltage is going to increase signal levels as well. Good idea! 😀

Doornkaat wrote:

SubZero wrote:

Maybe I should try to override the voltage on the slotket to 1,5 or 1,6V and try again on the Slot 1 board?

I would assume the boards can't supply 1,45V anyway so they deliver their minimum voltage but since you can set VID with the slocket it's worth a shot. Maybe increasing core voltage is going to increase signal levels as well. Good idea! 😀

OK, I tested to manually set voltage on the slotket, from 1,5V up to 1,65V in 0,05V increments. However it did not solve the issue. Interesting is that the CPU seemed to be less stable the more I increased the voltage. At 1,65V it hung up already before loading all icons on Windows desktop.

Well, I guess next step is to to test the CPU on a "real" 133MHz board with I815 or Via Apollo Pro 133A chipset. Unfortunately got none of these at hand at the moment.