Troubles running 133 MHz on ASUS P2B-B.

Hi!

I bought 733/256k/133MHz slot 1 CPU.
Unfortunately, it doesn't want to work with 133 MHz (PCI 33.3 MHz) on my P2B-B rev. 1.02 mb.

In fact, at the beginning I had even problems with 100 MHz: when windows
xp started I always had blue screens with core dumped. I suspected memory
problems, so from the three DIMMs I have I left only the best one 256 MB Hynix DIMM, and only then there wasn't any blue screens with 100 MHz.

However with 133 MHz although there was initial testing of memory in graphic
card (32 MB), after that cursor was blinking on the EPA Pollution Preventer
emblem, and that was it.

When all three DIMMs were in and it was set to 133 MHz, then computer
didn't start at all, it was only longly beeping instead, beeeep, beeeep,
beeeep... :/

I added now one more 256 MB VM DIMM and it seems working with 100 MHz
without blue screens, so I suspect that the 128 MB Winward DIMM was
causing the problems before.
However, that 128 MB DIMM also has 133 MHz sticker on it (as the two other
DIMMs), so it is strange that it doesn't want to work even with 100 MHz.

So what's the problem that I cannot run the computer with 133 MHz? How to find out what is causing the problem?

Is that Leadtek WinFast NVidia GeForce2 MX graphic card problem or some other problem?
Has anyone run it 133 MHz with that particular card or nvidia chip?

I have to run this processor now with 100 MHz, so I have 550 MHz, and not 733 MHz as it should be.

All my three DIMMs data collected using ctspd software is available on the
page:
http://an.feedle.com/memory/

1. That's what's written on the first DIMM:
56 MB Hynix 134A
HY67V28820HCT - H Korea.
Hynix 56
0134
HYM71V32635HCT8-H
PC133U-333-542
256MB Sync 133MHz CL3

2. Second:
256MB VM VT28SD16M8PC-7

3. Third:
128MB Winward WN57V658020T-7

There is now 3-3-3-10 setting by SPD in Bios.
Bios is the newest 1014 beta3.

And I was planning to upgrade it for tualatin 1.4/133MHz... :::...

Please, HELP!

Cheers!
andy

I did tests overnight with memtest86+ v. 1.60 for the two 256MB
memories at 100 MHz.
6 passes, and no errors.

I done some more testing with interesting results.

In chronological order.
I started the testing with the two 256 MB DIMMs and VIO 1.5V

1. FSB 124 MHz, PCI 31.0 MHz: beeep, beeep, beeep
2. 140 , 35 : beeep, but on the second try it started and crashed after
displaying "PCI device listening".
3. 150 , 37.5 : beeep
4. 133 33.3 : first try: hanged up on EPA sign, second time: beep
5. 105 35 : running fine
6. 110 36.7 : works well again
7. 112 37.3 : all fine
8. 115 38.3 : still works

Now I changed VIO to 3.66 V

9. 133 33.3 : beeep
10. 124 31 : works fine!

Now I removed Hynix memory from the first DIMM socket, and moved VM
memory from second DIMM socket into the 1st DIMM Socket. VIO still 3.66 V

11. 133 33.3 : beep
12. 124 31 : beep ; I moved then memory in the socket as sometimes it
happens that there is no electrical contact and... it helped... it worked
fine after that.
13. 133 33.3 : now it works too!

Back to the VIO 3.5V. Still only one 256 MB VM DIMM in the first DIMM
Socket

14. 133 33.3 : it still works!

Now I added second 256 MB Hynix memory to the second Socked.

15. 133 33.3 : blue screen, core dumped, windows restarting, then again
the same blue screen etc.

Now I removed 256 MB Hynix, and put 128 MB Winward in the second DIMM
Socket, in the first socket there was still 256 MB VM memory.

16. 133 33.3 : blue screen or beep

Swiching to VIO 3.66 V again

17. 133 33.3 : still blue screen

Removed Winward 128 MB

18. 133 33.3 : works well

Inserted Hynix to the third socket, leaving second socket empty, in the
first socket still the VM memory.

19. 133 33.3 : system crashed during startup and restarted, next time a
message displayed that file in System32\Drivers\Ntfs.sys not found or
damaged, and that I need to use rescue operation using Win xp
installation disk. :/

I removed Hynix and left only VM in the first Socket

20. 133 33.3 : working fine ; system didn't report missing ntfs.sys again

Well, that's all for now. I didn't try to insert either Winward or Hynix
alone in the first socket.
Initially it seemed that changing to VIO to 3.66 V helped, but since I
changed it later back to 3.5V and it still worked with one DIMM only it
seems that it's not that crutial after all.

So what can I do now? Buy better memory? Which one is good? From the
unbuffered one, because I guess I might find it difficult to find the
buffered one, and besides for some reason in the mb manual they say that
I should not use buffered:
http://an.feedle.com/memory/p2bb-102_ASUSa.pdf

Even when it was 66 MHz with the old celeron I had quite often electrical
contact problems. 3 DIMMs were inserted, and for example only one or two
were detected initially, I had to move memories a bit, or remove them and
insert again untill eventually all three were seen by bios when computer
was starting.
So perhaps in this case sometimes it has been also contact problems.
However, there is no visible dirt in the sockets, so I don't know why
these contact problem appear. I tried to clean it in the past with
alcohol C2H5OH, but it didn't help much.

Have anyone had similar problems?

One more thing. Apart from VIO jumpers (marked as vcore), next to it
there is also another set of jumpers marked in one position "test", and
in the other "normal" - what is that for? AFAIR there is nothing about it
in the mb's manual:
http://an.feedle.com/memory/p2bb-102_ASUSa.pdf

I've done some more testing.
Those two 256 MB DIMMs (one Hynix, one VM) which didn't show any errors
with 6 passes in memtest86+ with FSB set to 100 MHz, did not behave that
nicely with 133 MHz FSB.

With VIO set to 1.66 V VM memory showed around 315 errors after 1 pass.
Errors were in most of the types of testes performed by the program,
particularly many in the test no 6 and 5. However, when VIO was set to
1.5 V the same memory caused only 2 errors (both in test no 6) in one
pass.

The Hynix memory with VIO 1.5 V FSB 133 MHz demostrated 68 errors, most
of them in test no. 5.

The 128 MHz Winward memory didn't want to work at all with 133 MHz.

It's curious that they didn't show any errors at all at 100 MHz, and show
them now with 133 MHz. All those memories are allegedly 133 MHz (have
even stickers on them saying that). You've seen the ctspd screenshots.
However, because the memtest86+ implicitly tests not only memory but also
the CPU, L1 and L2 caches as well as the motherboard it is impossible for
the test to determine what causes the failure to occur, so perhaps those
errors are not due to bad memory, but mb failing to work correctly with
133 MHz?
What do you think?