Problem with 2 2.26B CPU's with TH7II-RAID

Hi guys,

Am hoping that someone can help me with this problem!

I've tried 2 2.26B Cpu's now

1 was a Malaysia 06/06/02
1 is a Philippines 06/19/02

Am using a newer TH7II-RAID Board (have actually tried 2 newer boards)

Using 512mb Samsung PC800 RDRAM

Now I know the RAM is good for up to 180 x 3 because I used to have a brilliant 1.6A which did 2.8 Ghz and a little more stable

The CPU does 150 x 17 fine on around 1.75 actual and 155 x 17 on around 1.80 - 1.85 actual and is stable in SANDRA and 3dMark2001SE

But here is where the problem starts

When going to say 160, the system boots fine and goes into Windows but won't complete 3dMark2001SE - it gets stuck on Dragothic (High Detail) - it just freezes
Or the Message is Timer Sanity Check Failed
Also sometimes it just hangs when running the SANDRA CPU Arithmetic Benchmark

I've VID Pinned the CPU to get upto 1.95 in the BIOS which is around 1.90 actual on the board and it still happens

Now the 06/06/02 pack date from Malay is apparently brilliant and though I also now have tried 1 Philippines 06/19/02 the same things happen

They need ridiculous amounts of voltage and even then they are not stable

The max stable I can get the CPU is 155 x 17 which is 2635mhz at around 1.8 actual but anything from there and the problems start with both CPU's

As I said before I know the RAMBUS can do more as it did 175 comfortably with my 1.6A which did 2.8Ghz @1.85 actual

Can anyone help me out here cos i've tried 2 CPU's and 2 mobos and it happens with both - these 2.26B's are apparently brilliant overclockers but there seems to be a problem somewhere and i hope someone can give me a hand

Thanx

Kiean

BTW am using Mr N's 7F Bios

Sounds EXACTLY like what all of my TH7II's have done. They will get up to 150, maybe a little bit more, and from there on they're completely unstable. It doesn't matter what the memory is running at, what the core voltage is, or what the rdram voltage is. Even using the fix pci/agp option doesn't help. Right now I've got a 2.4B that was pre-tested for 3GHz at 1.75 volts. I can get into windows all the way up to 166, but nothing in between 148 and 166 is stable. I've put a fan on the northbridge, and even a custom heatsink on the drcg's - I just can't get this board to do anything beyond 150 stable, regardless of any combination of parts I use with it. Bottom line I think it's the chipset.

Hi there

Thanks for the reply

Yeah you could be right.....I did have a 1.6A which did 2.8ghz and I ran that on a TH7II-Raid and it was perfectly stable and no problems at all
Perhaps it varies between boards but even that explanation wears thin as i've tried 2 different 2.26B's on 2 different TH7II-R and it happens on both of them
Upto 150/155 totally stable and then regardless of how many volts the CPU gets it is totally unstable
I know it isn't the RAM because i've had it running with the 1.6A on a TH7II-R @ 2.8Ghz with no probs so it might have something to do with the board I reckon
One of the 2.26B's is from a very good batch and even giving it 1.90 actual does not make it stable above 160mhz

Kiean

sounds like the chips are bummers. i would give the chips a go with a ddr setup.

Hiya JD,

Thanks for the reply

The 1st chip which was a Malay 06/06/02 is apparently from a very good batch and lots of people were getting 3.0ghz with very low voltage and atleast 2.9ghz which as little as 1.75v

I'm not sure how the Philippines 06/19/02 batch is

It is really odd that the exact same behaviour occurs with both chips

Are there any others ideas or suggestions from anyone? If not can someone suggest a good ddr mobo to try the chips on?

Thanks

The first socket 478 board I bought was the Epox 4SDA. It had the best bios imo at the time - voltage adjust for everything, very tweakable, etc. Now I guess I would look at the P4S533 or P4B533 from Asus, the latter allowing pci/agp lock like the TH7II which is important when you need those high fsb's to overclock the chips with lower multipliers.

Check your voltage on your 12v line

My rig with an enermax EP431 sucked the 12v line dry

I did the wire pin wrap and was able to get to 3068 (17 x180) but 12v line was at like 11.4v

I have ordered the psu in my sig waiting to get it will resolve all my power issues

Thanks TC and also Steven

What is the lowest the 12v line should be to maintain stability?

Also excuse the silly question but should I be looking at the +12v or -12v?

Thanks

Look at both of those lines and if they are way off - you have discovered your problem

The more I have read on different web boards I am clearly seeing that the wire pin wrapped P4 & Abit MB suck big power from the 12v line

Have to have a psu that can handle it

Steven,

Thanks very much

Apologies for yet another stupid question but what is defined as way off with the +12v and -12v?

Thanks

Well I am not an expert on this

But I would say that if you are seeing big swings like my 12v line with enermax was rapidly swinging between 11.4 & 11.75 and if it is dipping below 11.75 could cause problems

Thanks very much for your help Steven, it is greatly appreciated

With board set at 1.85v actual i've noticed that underload +12v goes to about 11.6 to 11.5 and -12v goes to -12.52
On idle they are +12v = 11.8 and 12v = 12.11

This is with the CPU clocked to 2635mhz (obviously do not need 1.85 actual to run this speed but just goes to show that if I try 2.8ghz at 1.85 then it is v unstable)

So could this be the reason why I am having problems?

I'm using a Sparkle 400Watt PSU btw

Thanks

Again I am not an expert

But I would say definately - yes

I also have 6 - 80mm case fans and a high speed 80mm 84 cfm fan on my swiftech and when my system goes under load you can hear all the fans slow down

There just isnt enough 12v power

Thanks again Steven
Eventhough you say you aren't an expert you obviously know your stuff and you are helping me out a lot, thanks

Are they both too low?

Also what is the solution? A bigger PSU or?

Thanks

Yes I would say both low

You need a psu of at least 400w but it must have adjustable pots

I just ordered the following psu

PC Power & Cooling - Turbo Cool 425 ATX with external fine adjustable pots on both the 5v and 12v lines

Since the 5 & 12v line pots are external can adjust the psu while the system is running so very easy to tweak the voltages while running MBM and watching the result at idle & load

Wont have it till July 11

Take at look at this page & you will see why their psu are so good

PC Power & Cooling

I have also seen where the Antec PP412X psu has adjustable pots - it is a low end cheaper power supply but can be adjusted

Heres a link

Antec PP412

Steven, you're a star, thanks ever so much for all the help, it is greatly appreciated

Thanks