24 Mar 2003, 18:48, JIM HOLSONBACK (1:123/140), wrote to MATT MC_CARTHY:

Hi JIM.

 MM>> Over the years, I _have_ repeated this with different CPUs at 
 MM>> different speeds, on different brand MBs, all with the same 
 MM>> results.  I _must_ start with a faster CPU to be able to get back 
 MM>> into the system to set things back to normal.

 JH> Have you noted whether the same problem occurs on boards which 
 JH> have jumpers for setting the clock-multiplier, or have you seen it 
 JH> only on the boards which "autodetect" the CPU?  Dang! 
What good is 
 JH> an autodetect function if it doesn't work, or only works _one_ time?

Only on the 'autodetect' boards.  Honestly though, I never tried on the
manual jumper boards, as I was pretty careful to try getting all the
jumpers right _before_ inserting the CPU.  Any time we were upgrading, I
had to dig out a manual and find all the settings first.

This phenomena only cropped up when the 'autodetect' boards came out, and I
had to plug in CPUs "blind" (and hope the board did it's thing
before the CPU cooked [never happened - personal phobia]).

Once the board detected and set itself for a '400' CPU, and later that CPU
was replaced with a '300', the board would not reset and the '300' couldn't
take the overclock, so in effect you had a 'dead' board. 

The current batches of Asus P4PE boards _do_ specify that they recognize a
CPU change, and 'reset' to a low speed.  I've not tested that feature at
all though, as it is now a real pain to replace P-4 CPUs just for the fun
of it, without destroying the heatsink mating materials.

 JH> I've seen what you described a time or two, where I put a PII-266 
 JH> in a board and it showed up in POST as a 233.  I guess I could get 
 JH> video since it was an underclock, and then I got into SETUP and 
 JH> toggled it back up to 266.

That might have been a board which had the 'reset when CPU changed' feature.
...................

 JH> Speaking of overclocking - - I have a Celeron 300a here which 
 JH> I've seen get video on a board jumpered for 100 fsb and clockmul of 
 JH> 4.0. AFAIK, this was the chip release that finally got Intel's 
 JH> attention about overclocking, and they subsequently took 
 JH> countermeasures. I guess maybe I should take it downtown with me, and 
 JH> give it a try in some of these boards before we finally give up on 
 JH> them.

Having a CPU that _could_ take the overclocking just might do the trick.

 JH> The big power users were the early PII thru 300 MHz.  The PII333 
 JH> and 350 data looks kinda funny.  But, looks like maybe boards in the 
 JH> 400-500 range which may be killing capacitors could have a prob in the
 JH> way they support those chips with core voltages of 2.0?

 JH> Just FWIW -  what do you think?

Close to the same as you.  Just a WAG, but in most cases a board is
designed for what CPU speeds are _presently_ available, based on sample
lots from Intel (or whomever), and the manufacturers KNOW the speeds will
go up in a short time, so instead of having obsolete boards as soon as they
roll out of the factory, they just "add a few more speed
settings", for chips that have not yet been designed.  

When those faster CPUs actually hit the market, their designs are likely
different in small ways that won't prevent their use in the boards, but are
likely to stress different parts of the circuitry in ways the board makers
hadn't thought of at the earlier time.  We KNOW the boards were not tested
at those higher speeds, as the CPUs didn't exist back then.  They can
simulate it at the factory, but nothing beats the real thing.


     Good luck...  M.

--- Msged/386 TE 06 (pre)