X2 3800 temperatures problems

[wevsspot]

Ah, I understand now.

[G-Code]

Have you tried to overclock ur system? or do you suspect that these temps will stop you.

 

I do not think that these temps would prevent you from getting some overclock. So if you are just focus Just on the temp you might be missing something else that is stopping you.

 

I have found a couple of things with this board. You need to give the CPU atleast 1.45 V, you need to get all the settings for the ram in the bios right, Get the FSB/multiplyer combo right. I found you go from getting nothing at all tearing your hair out to what my overclock ceiling when do I stop this is great.

 

Bruce

[panheimer]

I've been working on these temperatures for all the last week without stop and finally i've found what's wrong!

I think i'ts simply the bios that has something wrong reading CPU temperatures and have tried different bios. Originally the board came with 711 and it works great for my configuration (very stable once I understood its bugs) but has something wrong in reading my CPU's temperatures.

So I've tried the 602 but i've got several problems getting to POST

Then i've tried the 705, temperatures were fine, but i found some limits when i've tried to overclock it.

Finally i've decided to try the beta DC19 and I found it great! Some new option, like the possibility to get over 1,8 in CPU volt...i've run it for two days and my temperatures never got over 55° having my CPU got over 2750 Mhz...at stock speed it idles at 33° and get to 49° in full load!

But finally i've decided to move back to 711, even if temperatures readings are wrong because i've found that the DC19 has something wrong with my OCZ: with super PI it takes about 2:42.953 for the 4M while with the SAME SETTINGS and the 711 bios, it takes one second less: 2:41.703.

So now i'm happy with this buggy bios and my CPU.

 

A final question: now i'm setted at 296 x 9 (roughly @2673 Mhz) and 1,550 + 0,1 v, with my OCZ (2GB) working at 243 Mhz, 3-3-3-7 1T. I've tried to move over, but i have to give too much voltage on CPU and i don't feel comfortable doing so. it's right? or may i try to go over it? to get to 2750 my cpu needs 1,675 for stability...too much?

[wevsspot]

Yes too much. Check your PWMIC temperature and see how hot it runs when powering that X2. On my Ultra-D board the limiting factor is the PWMIC temp.

[panheimer]

You're right, the PWM reaches 52° in full load, but the system is still rock solid @2750 with 59° on the CPU and 1,675-1,7 v (BIOS DC19). As I said i don't feel comfortable giving so much voltage, and obviosly getting so high temperatures.

I think i'll remain at 2673 with decent temperatures, maybe in the future i'll move to water cooling and then i'll try to get more mhz....

thanks for your assistance!

[wevsspot]

NP,

 

I agree with you. I'd be happy at 2.67ghz at fairly low voltage and good temps. Severe over-volting of your cpu for long periods of time will reduce it's life span. My general rule of thumb is never to exceed 10% of the max manufacturer's recommended voltage. i.e. 1.40X10% = .14

 

1.4 + .14 = 1.54vcore