8RDA+ power off issue

[Yitch]

Hi everybody,

I just bought the following system:

EPOX 8RDA+ rev 1.1
Athlon XP 2100+ (Tbred B)
Thermaltake Volcano 9 HSF
512 MB PC3200 Kingston DDR RAM
60 GB Maxtor Hard Drive
48x24x48x Lite-on burner
300W PSU

After assembling everything, the computer worked fine (CPU temp was 35C). I managed to get into the BIOS, set the FSB to 166 MHz and the DDR to 200 MHz. I turned off CPU throttling, because as far as I know, the AXP 2100 doesn't use it. My next step was to install the OS so I decided to attach the burner to the mobo. After turning the computer on again, it shut off automatically after 10 seconds (I'm sure this is unrelated to the burner)! Now, every time I try booting up, the computer shuts off after 3-10 seconds! I am barely able to get into the BIOS before the comp shuts off.

I've stripped the computer down to its bare essentials (only RAM, CPU/HSF, keyboard connected) outside of the case and I still get the problem. The power on and reset sw connectors are attached to the case, which lets me power on the mobo. I've tried swapping PSUs/memory but that didn't help.

My first thought was perhaps the mobo is shutting off because the CPU is too hot. At POST (and in the BIOS when I have enough time to get there), 35C is reported for the CPU temp. I read somewhere that the EPOX boards can't display the on-die thermal diode temp in the BIOS (or mobo manager) but can in fact read it for the purposes of the temp protection circuit. It seems impossible that the CPU is hot enough (110C??) to trigger the protection circuit, since the mobo reports 35C.

Any ideas? I'm going to try swapping the CPU to see if that's the problem but I'm 90% sure I've got a bad mobo. I guess it's back to ASUS for me

Thanks,
Yitch

[Yitch]

I solved the problem (knock on wood)! I removed my heatsink and noticed that perhaps I installed it the wrong way (the CPU barely had made contact with the copper core of the base). After switching the heatsink around, my computer no longer shuts off automatically.

I'd have to assume that the EPOX protection circuit kicked in and shut off the computer. I have to admit that I'm very surprised that having the heatsink the "wrong way" would have a drastic effect on the temp. I'm also worried about the EPOX since it reported 35C for the CPU temp!

Yitch

[timbob2469]

Yitch , your not the first person to do this and probably won't be the last . Do a search on this forum for ,"epox saved my bacon". That auto-thermal shutdown is a great feature. I don't know why the temp was registered so low, maybe there are 2 probes, one for bios reading , and one for shut-down.
bob

[Áedán]

Originally posted by timbob2469
That auto-thermal shutdown is a great feature. I don't know why the temp was registered so low, maybe there are 2 probes, one for bios reading , and one for shut-down.
bob

There are two probes! If you have a CPU with the thermal diode onboard, then the thermal protection uses that diode. This is a *FAST* system, that can react in under a millisecond to overheating.

The displayed readings just use a thermistor in the middle of the socket. Hence, it takes a while to track the changes on the CPU, due to thermal inertia. It can take a good number of seconds to track the changes.

Áedán

[Flymo]

Originally posted by Áedán


There are two probes! If you have a CPU with the thermal diode onboard, then the thermal protection uses that diode. This is a *FAST* system, that can react in under a millisecond to overheating.

The displayed readings just use a thermistor in the middle of the socket. Hence, it takes a while to track the changes on the CPU, due to thermal inertia. It can take a good number of seconds to track the changes.

Áedán

I wonder why, if the boards are able to use this sensor at all, that they dont use it for temp readings in bios?????
makes a LOT of sense


Flymo

[Áedán]

Originally posted by Flymo
I wonder why, if the boards are able to use this sensor at all, that they dont use it for temp readings in bios?????
makes a LOT of sense

In the case of the EPoX board, the thermal protection chip is completely seperate to the rest of the system. This means even if the processor has locked up completely, the thermal protection will continue to operate. This is the correct way of implementing thermal protection, as it's not possible to disable it in any way, and has no reliance on any other part of the system. It's also blazingly fast at cutting the power should the CPU overheat.

Unfortunatly, it's not possible to have two devices reading the same sensor. So, EPoX implemented a second sensor, which is what you see in the BIOS/MBM/whatever.

Additionally, not all AMD chips have this thermal sensor embedded in them. For example, my chip doesn't! If the only way to read the temperature was via the thermal sensor, then some users would not have any temperature reading at all.

Áedán

[Louofm1]

I made the same bonehead mistake while building my first PC (a few weeks ago). After the first boot, I noticed that the temperatures rose into the 40's quickly. I wasn't sure if it was normal or not. I continued to add the remaining drives. A few minutes after the second boot, my machine shutdown automatically and set off the alarm. Since the CPU was adequately cooled and the case was still open, I realized that I did something wrong.

I decided to remove the Volcano 7 heatsink and fan assembly. I quickly realized that the silver 3 paste had smeared around (even though I was careful to not shift the heatsink while seating it to the chip). Where I messed up was that I didnt properly match up the notch (on the bottom of the heatsink) to the chip base. The notch im speaking of is used to lock the CPU to the board. Remember, the heatsink must sit on the chip perfectly flush (for proper heat transfer). Therefore, the heatsink was not sitting flush on the chip.

I removed the chip and used the appropriate cleaning material to remove all of the old thermal paste. I reapplied a fresh coat to both the chip and the heatsink (per the instructions on the thermal paste website). I installed the heatsink/fan in the "correct" manner, rebooted and cured my overheating issue. Thank goodness the 8RDA+ MB was developed with the overheat protection circuits.

I was more interested in keeping the Thermaltake Logo upright (for the clear plexiglass window) that I didnt pay attention to the fact that there was a notch on the copper heatsink. After installing the heatsink properly, I was able to remove the cover of the fan and flip it over (so it reads upright).

Hope this makes sense and hope this helps. Learn from your mistakes (and others mistakes).

Lou