Sudden increase of CPU voltage during undervolting

[LordIT]

I undervolted my CPU. It worked fine down to 0,9375V. :

http://img221.imageshack.us/img221/6820/voltage1qu5.jpg (http://imageshack.us)
http://img221.imageshack.us/img221/voltage1qu5.jpg/1/w770.png (http://g.imageshack.us/img221/voltage1qu5.jpg/1/)

Then I adjusted 0,925V and the CPU voltage increased up to 1,62V!!!
Seems that it is a reduced CPU ratio voltage (two times 0,81Vx2).
The Intel VID of C2D E4400 is 0,85-1,5V.

http://img221.imageshack.us/img221/9826/voltage2jv9.jpg (http://imageshack.us)
http://img221.imageshack.us/img221/voltage2jv9.jpg/1/w770.png (http://g.imageshack.us/img221/voltage2jv9.jpg/1/)
That is very strange. :confused:
It has nothing to do with Aegis Panel. The same will happen after adjusting 0.925V in BIOS.
Why is that?

[HDCHOPPER]

dont know but ..... why isnt my aegis panel like yours with the same version ??

http://img4.glowfoto.com/images/2008/09/18-1004076400T.jpg (http://www.glowfoto.com/viewimage.php?img=18-100407L&y=2008&m=09&t=jpg&rand=6400&srv=img4)

[LordIT]

I dont think that your Aegis Panel is different. Please, press the OC button.

[HDCHOPPER]

bahahahahahahahaha after trying all them aegis panels totaly missed that

thanks .... time for more coffee !!!

[LordIT]

Ah, do you mean your Aegis Panel works different?

[HDCHOPPER]

:up::rofl:

[LordIT]

Many thanks for telling your detailed experiences with Aegis Panel! :ROTF:

[LordIT]

Can somebody explain the sudden increase of CPU voltage during undervolting in posting #1? Why doesn’t the computer turn off instead of given the CPU more than 1,60V –could be damaged very quickly!

[HDCHOPPER]

ya messing around with my aegis panel oc function last night while dropping the cpu voltage ( cpu speed was at 4.6) to see the least needed for 4.6 the readings on the cpu volt panel went up to like 1.8 volt :eek: quickly hit the off button on comp :D

think in future eye will only change voltages in bios :cool:

[ol'norton]

ya messing around with my aegis panel oc function last night while dropping the cpu voltage ( cpu speed was at 4.6) to see the least needed for 4.6 the readings on the cpu volt panel went up to like 1.8 volt :eek: quickly hit the off button on comp :D

think in future eye will only change voltages in bios :cool:



Keep your CPU voltage Multiplier (The top one) on X1.000. :up:

[HDCHOPPER]

normally eye do but ya at that time it was on X1.036

[LordIT]

ya messing around with my aegis panel oc function last night while dropping the cpu voltage ( cpu speed was at 4.6) to see the least needed for 4.6 the readings on the cpu volt panel went up to like 1.8 volt :eek: quickly hit the off button on comp :D

think in future eye will only change voltages in bios :cool:

That is strange and dangerous, too. Your E8400 will die very quickly @1,8V. But imagine you will get this high voltage after adjusting your low settings in BIOS!

Back to my problem:
Either my E4400 cant read very low voltages or the Foxconn Mars` voltage supply doesn’t work properly.
Could anybody test it with 0,9V on a Foxconn Mars, too? Perhaps a 45nm CPU works better.

[koda]

Are you really sure about core voltage ? :shrug:
(not reporting Aegis panel one in this case ;), it could be a bug)

Could you try to look cpu temps with a soft like realtemp/coretemp at the same time ?

My bet is :
If you really jump from 0.83v to 1.62v, you will see a very significative change in the reporting temps :up:

[LordIT]

Are you really sure about core voltage ? :shrug:
(not reporting Aegis panel one in this case ;), it could be a bug)

Could you try to look cpu temps with a soft like realtemp/coretemp at the same time ?

My bet is :
If you really jump from 0.83v to 1.62v, you will see a very significative change in the reporting temps :up:

Yes, I am sure. I made a short stress test, too. The core temperatures increased rapidly.
http://img217.imageshack.us/img217/729/vcoreng3.jpg (http://imageshack.us)
http://img217.imageshack.us/img217/vcoreng3.jpg/1/w996.png (http://g.imageshack.us/img217/vcoreng3.jpg/1/)

[HDCHOPPER]

That is strange and dangerous, too. Your E8400 will die very quickly @1,8V. But imagine you will get this high voltage after adjusting your low settings in BIOS!

Back to my problem:
Either my E4400 cant read very low voltages or the Foxconn Mars` voltage supply doesn’t work properly.
Could anybody test it with 0,9V on a Foxconn Mars, too? Perhaps a 45nm CPU works better.

have never had the bios do this since day 1 with this board :clap:

only happened with the oc function and this aegis panel
figger just a bug with the oc function on this release

every setting in bios has been exaly correct in windows to date & eye check every time once in windows ;)

[LordIT]

every setting in bios has been exaly correct in windows to date & eye check every time once in windows ;)


have you ever tried to adjust the VCore in BIOS less than 0,938V?

[HDCHOPPER]

no thats why my readings on realtemp arnt calibrated ;)

verrrryyyy hard for me to down clock :rofl:

[saaya]

That is strange and dangerous, too. Your E8400 will die very quickly @1,8V.actually i know somebody who is trying to kill an 8400 for months and even with 2v vcore and 2v pll voltage for hours he didnt manage to kill it :lol:

but yes, its not good that the voltage jumps up like that :confused:

so regardless of bios or windows, when you decrease the voltage below a certain point itactually jumps up?
sounds like a bios bug to me... since aegis panel doesnt do anything else but talk to the bios which talks to the vreg chips (iirc)

ill check with the engineers! :toast:

[negev]

Surely undervolting the blackops is like trying to stop an erupting volcano with balsa wood. No wonder it blasted through it :P

[saaya]

heheheh :D

[LordIT]

actually i know somebody who is trying to kill an 8400 for months and even with 2v vcore and 2v pll voltage for hours he didnt manage to kill it :lol:

Perhaps he is not good enough in his job? :lol:



so regardless of bios or windows, when you decrease the voltage below a certain point itactually jumps up?

Yes exactly, after adjusting below 0,9375V (BIOS) or below 0,938 (Aegis Panel) it immediately jumps up to maximum CPU Voltage 1,60V (CPU multipler X1).


sounds like a bios bug to me... since aegis panel doesnt do anything else but talk to the bios which talks to the vreg chips (iirc)
ill check with the engineers! :toast:

Would be great to fix that with a new BIOS. :up:

[LordIT]

actually i know somebody who is trying to kill an 8400 for months and even with 2v vcore and 2v pll voltage for hours he didnt manage to kill it :lol:

but yes, its not good that the voltage jumps up like that :confused:

so regardless of bios or windows, when you decrease the voltage below a certain point itactually jumps up?
sounds like a bios bug to me... since aegis panel doesnt do anything else but talk to the bios which talks to the vreg chips (iirc)

ill check with the engineers! :toast:

What do the engineers say?

[Ryan12]

What do the engineers say?
I remember in another thread Saaya saying that the lowest voltage or something in aegis panel doesnt reference properly to the bios... Causing the sudden high voltage increase.

[LordIT]

I remember in another thread Saaya saying that the lowest voltage or something in aegis panel doesnt reference properly to the bios... Causing the sudden high voltage increase.

Yes, but I have that problem though Aegis Panel is not started. I would like to know if its is either my Mars only or if its a problem generally. So you could clock down your CPU very low and adjust 0,9375V in BIOS to get started your computer. Then open Aegis Panel change the voltage from 0,93V to 0,925V (posting #1) and whatch the real voltage on the right side in Aegis Panel. If its goes up to 1,6V you can change it to 0,9375V quickly.

After that we would know what kind of mistake that sudden voltage increase is. Something must be done anyway.

[HDCHOPPER]

ok loard will try that for you

me thinks it just something wrong with the aegis panel .15~.34
first time eye saw panel .15 for download it was sposta have an oc button but it didnt for me
and every panel since has something wierd about it but .14 is perfect except no oc button
just gotta new E5200 to play with and 1.6 v aint gonna hurt it cause eyema gonna try to kill it anyways lol

[HDCHOPPER]

ok loard the lowest eye could boot in was 1.025
opened aegis panel and cpu v reading 0.094
droped 1 tic and panel reading was 0.093
droped 1 tic and panel reading wes 0,091
droped 1 more tic and blue screened

so my E5200 wont boot that low to start with sorry
maybe cause eye allready beat it up a bit:D last night was trying to see top
max clock with it and fed 1.7 v into it but it still does 4.0 at 1.36 and that was what it did the first time it got fired up :D it's vis was 1.250 anyways

maybe someone else can try this for you
good luck

[LordIT]

ok loard the lowest eye could boot in was 1.025
opened aegis panel and cpu v reading 0.094
droped 1 tic and panel reading was 0.093
droped 1 tic and panel reading wes 0,091
good luck

Thanks chopper. :) I would like to repeat in case of misunderstandings:
You dropped down CPU voltage to 0,912V in AegisPanel and reading was 0,91V (0,81 idle), too. And there was no sudden voltage increase to read on the right site of AegisPanel.
Is that right?

[ThugsRook]

sorry no help, but just for reference...

my e8200 on a P5Q (p45) is able to run 6x333 on only 0.85v (prime stable) :cool:
(but i think they are supposed to be able to do that :shrug: )

[LordIT]

@ThugsRook:
What will happen if you try 6x200?

[HDCHOPPER]

right every drop resulted in a drop in v on panel ( also using panel oc function to do all this )
and not once did it overvolt
thier was an overvolt event once but it was the aegis panel before the last realse but dont remember what all the facts were
come to think of it believe it in this thread if not in another

EDIT: yes it was post #9 & post #10 explained it and yes if cpu multi keept at 1.00 default it wouldnt overvolt
but when using any multi other than default it would overvolt then only .. for me anyways
eye just quit using the panel for overclocking and used bios or set fsb

[LordIT]

so my E5200 wont boot that low to start with sorry

That is strange. I thought that the FSB of the E5200 can be clocked down a lot - because of 45nm. :confused:


right every drop resulted in a drop in v on panel
If you got 0,91V adjusted then there is a mistake on my Mars only. Did you check it with CPUz, too?

[ThugsRook]

@ThugsRook:
What will happen if you try 6x200?
6x200 ran fine, then tried 266 and 333, both passed. 400 failed and 366 failed. 350 might pass, havent tried it.

temps were extremely low, barely moved above idle when loaded :cool:



http://i224.photobucket.com/albums/dd164/ThugsRook/P5Q/6x200.jpg

[LordIT]

@ThugsRook:
Was the multipler 6 fixed @0,81V or would it go up on stress?

[ThugsRook]

@ThugsRook:
Was the multipler 6 fixed @0,81V or would it go up on stress?
C1E and speedstep was disabled, fixed on 6x multi.

i was testing for lowest clock, then how far i could go on the boards minimum voltage... with a 6x multi.

:)

[LordIT]

Yes, thanks for that ThugsRook. It is nice to know that the E8200 can go down to 0,816V. :) So many users could test it on Mars, too.

But I think it is not possible to enable EIST/C1E on PQ5/P45 anyway if FSB and VCore is adjusted, is it?

[ThugsRook]

Yes, thanks for that ThugsRook. It is nice to know that the E8200 can go down to 0,816V. :) So many users could test it on Mars, too.

But I think it is not possible to enable EIST/C1E on PQ5/P45 anyway if FSB and VCore is adjusted, is it?
youre welcome :)

C1E works, it just lowers the multi to 6x at idle if its set above 6x. i dont know about EIST, never bothered to try it.

:)

[HDCHOPPER]

DO'OH didnt lower fsb lol
will try again later for ya must of been the med's :rofl: