Are there any 775 chips with no cold bug ?

[The Byter]

10x !

[n00b 0f l337]

Yes, presscotts, cedar mills, and in general net burst.

[The Byter]

We have a 630 and 641 cpus both locks up under -128C ~

any idea ?

[Gautam]

What board are you using?

[The Byter]

P5B and commando

[DEVIL K-ce]

This is board bug, not CPU's :)

Regards

Martin

[The Byter]

BOTH ?!

if so I need to find a "not cold bug" mobo ??

im lost... :\

[krullet]

ye its probobly the board, im having the same problems with conroe clk with xtreem cold, the board gives up @ around -90..and i now for a fact that the cpu can run under ln2 or around -130

[The Byter]

We tried P5B DLX and Commando, both locks up with 641,E6600 @ the same temps...

we had to keep the temps above -130C~

Is it the mobos or the cpus ?

[Jaan]

We tried P5B DLX and Commando, both locks up with 641,E6600 @ the same temps...
we had to keep the temps above -130C~
Is it the mobos or the cpus ?
It is a mobo bios bug, the internal cpu temp reading logic is limited to -127*C on most boards.
If you install Speedfan, you see it yourself.
After that some mobo will freeze, some going to show very high temps- this is Asus P5B-Plus for example:
http://www.oc.pri.ee/temp/asus_bios@-186.jpg
This mobo have cold bug at this showed temp- after that is lockup.
Intel Bad Axe2 worked down to -195,5*C :eek: I soft power off mobo and wait min temp.
At this 195,5 is starterd normaly and worked too.
Old bad axe is having same -127C lockup, so I have have to fool temp sensor for normal operation below this temp.
This info is getted with cedarmill 631ES.
Load temp with this cpu was -180,6*C at 1,712V Vcore.
Temp reading sensor is calibrated witw ln2 and it placed 2mm from cpu ihs, thermal interface was liquid metal and sensor hole is full of it too.
So it is quite good solution for reding temps imho.

[The Byter]

So what can we do in order to use P5B DLX or commando with LN2 low temps ?

What board are you using?


I PMed you about my LN2 prob and you say:

"Commando acts perfectly under ln2, no bugs"

so what gives ? some ppl say its the mobo, some say it's not the mobo...

[Gautam]

Yeah. That's why I'm scratching my head.

I had the Commando running at about -161 container base temp with an X6800, which would be around 140ish core.

I saw high temps in BIOS just like that, and it said "Over Temperature Error," but no troubles with freezing up or anything. Everything still worked fine. Try disabling all the thermal protection options.

However, one thing to keep in mind is that subzero is quite tricky. There have been cases were different motherboards had different cold bugs in the i955X days. It is possible that some part of your Commando doesn't like extreme cold, even though mine can...maybe. It might just be all the chips too, as going below -130 usually takes a special one.


But on second thought, are you trying to boot up at below 130? I always boot at about -50C, and then work up from there. Enter Windows maybe around 110, then go lower after I'm in Windows.

[The Byter]

I check the lockup in the bios and in windows.

I put ln2 in when im in the bios and i watch the temp, when it gets to -128C~ it jumps to +123C and if I cool it more it locks up.

same in windows, after I hit that "too cold" wall I have to worm it all up or it will not boot...

I tried 630,641,e6600,x6800,e6300 on P5B DLX and commando... no luck :confused:



:(

[Gautam]

123 exact for every chip?

That doesn't sound right...if the mobo is bugging it'd bug at 127.5.

[The Byter]

all those chip locked up with ln2 on p5b and commando.

all around -128C~

I need some other mobos ? bios ? cpu ?


:(

[Gautam]

If it's "around" -128C, then its possible all chips might be bugging there...that's a common point for 65nm chips. If they're all exactly at 128C then maybe the boards are bugged.

[harleybro]

IMHO part of the bugging is due to the caps being so close to the cpu socket and tube base. 630 and 640 should work better. Back with the P5WD2 it used to bug around -130 well wouldn't boot you had to boot with it warmer then drop the temps. once you where in windows. Although there are some killer C2D's out there I haven't seen any of them takling full LN2 temps.

[Nosfer@tu]

IMHO part of the bugging is due to the caps being so close to the cpu socket and tube base. 630 and 640 should work better. Back with the P5WD2 it used to bug around -130 well wouldn't boot you had to boot with it warmer then drop the temps. once you where in windows. Although there are some killer C2D's out there I haven't seen any of them takling full LN2 temps.

So a heater in betwean the socket area and the cpu would help?

Do you reasky think it is those smd condensators ?
That would explayin why xbx2 works better. It is fild with condensators.

[Jaan]

Do you reasky think it is those smd condensators ?
That would explayin why xbx2 works better. It is fild with condensators.
SMD caps [small brown one] are ceramic- they dont fear LN2 temps,- I test ceramic caps putting it in LN2- they work flawlessly@-196*C.
SMD bigger caps [tantalum] also work nicely in cold, but loose ca. 10 percent of capasity.
-40 and down is problem to electrolyte caps [cylinder style]- they loose capasity in cold and turn into passive element on mobo@-100 or more cold.
Result is bad power to CPU-NB-etc.
I test only small number of capasitor in ln2, it is possible, that some other ceramic-tantalum-electrolute caps are performing other way in cold.
Electrolyte@-190*C: bad thing
http://www.oc.pri.ee/temp/Vcore_caps_s.jpg

DMM show 551 nanoFarads@-70*C, cap real electrical capacity is 4700 mikroFarads, or capasity is 8530X smaller, :eek: than in +20*C.

http://www.oc.pri.ee/temp/cap@cold.jpg

[The Byter]

pro info !

10x alot...

so those caps could be the "cold bug" ?

but when I think of it, ppl have very long ln2 sessions when holding the container temp at a certain temp for the cpu not to bug... the other parts around the cpu don't put out much heat and work for all this time under ln2 temp...

:confused:

[Nosfer@tu]

Thx for the info :)

That is why you have fans around the cpu area :D

to heat the capasitors up :D

[vampire338]

both fo the cpu's you where using ive seen overclocked very high using liquid nitrogen. so its your mb. by the way any of you know a good overclocking motherboard that would work good for ln2/di?

[G H Z]

No problems for me booting @ -150 on 965-S

[Nosfer@tu]

No problems for me booting @ -150 on 965-S

WHAT ?

How do you mesure?
What Cpu and what mobo ?

[Gautam]

His -150 is not your -150 nor is it mine...every container will read very different temperatures at the base with respect to what the core/ihs is actually at. Base temperatures simply cannot be compared like this.

[Lord_RTKK]

I boot at around -130ºC with my Commando and E6700 (container base temp)

BTW, I had an old CedarMill 631 that had very agressive cold boot. When you went colder than -60 it wont boot, but you can boot a -50 or so and when you are in Windows freeze it up to -155 or so... (tested on the same commando that boots @ -130 with E6700, so no mobo bug)

[Nosfer@tu]

His -150 is not your -150 nor is it mine...every container will read very different temperatures at the base with respect to what the core/ihs is actually at. Base temperatures simply cannot be compared like this.

I know they will deviate, but one should be able to compare them.

[harleybro]

I know they will deviate, but one should be able to compare them.

It actually could be a big differance depending on the method used to check temperatures and the probe location. My thermocouple is only into the base about 1/4 inch but I can tell when it will bug within 2-3c. If someone else has there thermocouple drilled all the way into the center of the base it will be a much differant reading then one where it is taped to the edge of the base. Also a low end temperature measuring device may have a +/-5c varience where a high end unit could be as little +/- 1c.

[Jaan]

Exactly. The heatpoint of usual CPU is rather small, so sensor placement in pipe is important thing, if you want to know what is cpu/ihs real temp.
And the first thing is you have to check/calibrate your sensor and thermometer- if you use LN2- put sensor in it- it have to show
- 195.79 °C/ -320.42 °F, if not, you have to take delta into consideration of adjust ofset in your termometer (if you playing near this temp),
in dry ice you have -78.5 °C (-109.3 °F) in ideal. Of course if you have opportunity to check-calibrate your stuff with other temp is better,
becouse sensor delta is not always linear.

It actually could be a big differance depending on the method used to check temperatures and the probe location. My thermocouple is only into the base about 1/4 inch but I can tell when it will bug within 2-3c. If someone else has there thermocouple drilled all the way into the center of the base it will be a much differant reading then one where it is taped to the edge of the base. Also a low end temperature measuring device may have a +/-5c varience where a high end unit could be as little +/- 1c.