I am having a strange issue. I am running a Duron 900Mhz on an ABIT KT Motherboard and I am not overclocking any aspect of my system because i know that in the summer my house is an oven (living in Texas)
I've not changed anything in my system lately, though I had some issues with my ATI video card (Rage Fury Pro, yeah i need to upgrade).
Now, if I leave my system on continually, when it goes idle my CPU begins overheating, until the system locks up. This does not happen when I am actively playing a game on the computer, only when it is idle.
Can anyone give a suggestion as to why? I have turned off all power-scheme items in both the BIOS and Windows which would shut down any system components, and I am at a loss as to what is going on.
are you running any Seti or similar app to eat up spare cpu cycles? Possibly under certian conditions an application will use CPU, but not 'stress' it as hard, then you quit the app, Seti or similar kicks in and really stresses the cpu out - heating it up!
Just a thought.
-* <font color=red> Under Offer </font color=red> *-
email for application details
Sounds like the fan is stopping during standby/idle. If it's on mb header, this might be the case.
Hell, try hooking the fan up to a powerd 12v line, and see if it stays on.
Rich is the nation that has many war heroes. Long since forgotten...
November 7, 2001 3:52:47 PM
Thanks for the suggestions. I really have a gut feeling that it is something with a program running. No I'm not using SETI, in fact I'm not even using one of the 3D screensavers that comes with windows, just the plain old flying windows screensaver.
The fan still seems to be running when the system has gone idle, I can't check the RPMs on it because of the system locking up. When I reboot the system the ViaHM program will tell me that everything is fine (except for the CPU temperature, which is overheating but it starts dropping in Temperature immediately on reboot until finally it is in the upper 20s) with the fan on the MB/CPU. It doesn't seem to be aware of the case fan, but that is on constantly since I'd hear the difference.
Still doing it with the case off.... Not that I was expecting it to be any different.
Thanks for all the suggestions, and please keep em coming. This is pretty frustrating.
November 7, 2001 4:44:39 PM
On the Abit kt's you have to have the fan connected to the FAN1 on the mobo unless you have flashed the bios with the latest bios upgrade and then in the bios set the CPU FAN OFF. If not, do it.
Sounds like the fan may be winding down on the way to shutoff. That would explain you hearing it, if you only stayed around for a little while after going idle. Either that, or the fan just slows a bit for a reason I am unsure of.
:tongue: Have you ever tried cooking an egg on your HSF? Tasty. :tongue:
Only at idle? Hmmm, sounds familiar. Do you happen to be using Wpcredit/Wpcrset register offset 52 hack to lower idle temperatures? (I think one or two of the beta 3R's had the hack built-in) Anyway, the hack didn't work on my KT7, causing crashes at idle as you described. That was until I obtained a BIOS version that contains the <b>K7 Clk Control</b> option in BIOS. To make the hack work set the above option to "Optimal" as opposed to "Default".
Personally, I always found power management makes my KT7 completely unstable. I never acheived any semblance of stability until I obtained a BIOS that allowed the complete deactivation of ACPI. I think WZ was the only factory BIOS with that option. Later I learned that the capability is in all Abit BIOSes but disabled. One needs to obtain a program to mod the BIOS or optionally use a BIOS that is already mod'd. Both choices are available at <A HREF="http://www.biosmods.com/" target="_new">Binary's BIOS Mods</A> (mods for other brand mobos also available). There is a downside to not having ACPI, besides the lack of power management. The downside is there is no longer any CPU cooling software which makes your idle temperatures only about 5 degrees lower than your max load temperatures.
If the above description doesn't fit your situation then perhaps this will help. Motherboard Monitor 5.x is a much better program than the VIA hardware monitor. One feature that I think will help you is the log feature. MBM can be set to keep a running log of your fan speeds, voltages, temperatures, etc. Really should help tracking down some of your problems. You can also set alarms, shutdown the system (you need a third party program for this) or run programs when something gets out of sorts.
You didn't say what you were using for a CPU cooler or what your load temperature is but if you happen to be looking for a good, cheap cooler that doesn't break your air drums. The Evercool ND-8 keeps my Duron 600@1007mhz at 48 degrees running Toast (nice Athlon/Duron burnin program), below 46 degrees running Seti@home, and at about 45 degrees looping 3DMark2001. (These are good temps because my core voltage is at 1.91 volts and 53 degrees is my Duron's fail point under said conditions). My idle is 41-42. (In case you're wondering, just as a test I re-enabled ACPI and used the Offset 52 hack. My idle temperature was 27 degrees with motherboard temperature at 25). The ND-8 is a d*mn good cooler and it's only $10 at <A HREF="http://www.nexfan.com" target="_new">www.nexfan.com</A>.
Hope some of this helps.
Duron 600@1007mhz, FSB 106mhz
Evercool ND-8 cooler with Evercool 28 CFM fan (<b>updated: wrong fan listed originally</b>)
KT7 (minus chipset fan. It failed and I don't need)
OCZ Enhanced PC133, 128MB, and Enhanced PC150, 256MB, @141mhz.
Visiontek Geforce2 GTS-V (175/286 overclocked to 210/340)
Maxtor DiamondMax 40VL hard drive
ATX Mid Tower (a small mid)
Enermax EG351P-VE, 330 watt PSU
Enermax 80mm, thermally-controlled case fan (permanently shunted to max speed)
Generic 80mm case fan (exhaust)
Various ancient hardware.
<b>We are all beta testers!</b><P ID="edit"><FONT SIZE=-1><EM>Edited by phsstpok on 11/09/01 00:16 AM.</EM></FONT></P>
November 14, 2001 8:17:39 PM
Not at home yet so I can't check but I think I know where the problem lies. This may just be a dumb case of an error between chair and computer.
I believe that at some point prior to this happening I had made a change to the BIOS to go to a 'standard' defaults mode of operation. With this change I believe that it reset the CPU voltage to 1.7V which is the standard for Athlon T-Bird processors. I just read that for Duron it should be 1.5V. This of course would explain why even at Idle I was having such a high temperature problem going on with my system.
I checked out the chip and heatsinks and the pink goo of stuff on the bottom of the heatsink positioned above the CPU has flowed, completely encasing one or two of the support chips on the CPU. I have manually cleared the pink goo from the chip and re-installed in the system and it no longer seems to lock up, also it only over-heats to about 53degrees (still above my limit, but better).
I plan to check the voltage and reset it to 1.5V.
Do you think it will fix my problems? It sounds Spock-on Logical.
Thanks for the link on the Fan. I think I will end up replacing the cpu/heatsink/fan anyway, it's just too cheap and easy to get a 1.2GHz 200Mhz T-Bird these days.
<P ID="edit"><FONT SIZE=-1><EM>Edited by braincandy99 on 11/14/01 05:23 PM.</EM></FONT></P>
I thought the Duron 900 used a default voltage of 1.7volts (maybe it's the 950 and up).
My Duron 600 needs 1.91 volts to run at 1007mhz which is up from 1.5 volts. Anything lower and the system becomes unstable. This is all true with my original cooler. With the new ND-8 the CPU is about 5 degrees cooler so maybe I can use a lower voltage.
If you suspect heat truly is a problem for your system you can try leaving the case open and using a room fan to blow in lots of air. You should see much lower CPU temperatures. If your system is now stable then heat was definitely an issue.
<b>We are all beta testers!</b>
November 20, 2001 5:32:28 PM
Just for closure and completeness for anyone else having a similar problem.
I benchmarked my AMD Duron 900Mhz using 3DMark2001's CPU benchmark at core voltage settings of 1.7 and 1.55 and found no difference in the performance.
Resetting the core voltage to 1.55 lowered my standard operating temperature to 43-47C and I am no longer having the lock-up problems I was having.
I will make a broad statement that Duron processors at 900Mhz or lower require only 1.5V on the core to perform at their intended speeds. This is based on my 1 time experience.
Thanks to everyone for their help. Man I feel dumb, but at the same time, I don't see too many useful catalogs of CPU/Core Voltage/Operating Temperature tables for todays(and yesterdays and tomorrows) processors.
I think the problem you were running into is execessive heat caused by the higher voltage, not the voltage itself. Your present cooler wasn't able to deal with the extra heat.
When overclocking, higher voltage and great cooling are keys to stability. Usually the higher the voltage the more stable the CPU is but at the same time the temperature needs to be lowered. At the limits of overclocking, the CPU needs to be far below the temperature range at which a non-overclocked CPU can operate.
Glad you found your problem.
By the way, I've been overclocking my Duron 600 for just about a year now at 1.91 volts. It's never been stable at anything less and never was stable at 1.5 volts, not even at the stock speed.