ASUS A7V333 Athlon XP 2200+ difficulties...

[SpaceNinja]

First off, the A7V333 is getting massively inaccurate temp readings off the on die temp monitor on the Thoroughbred core. It claims that the system is anywhere between 45C - 75C at idle while in the BIOS. My Koolance case puts it at 30C, which I beleive to be more accurate, as the brackets on the Koolance waterblock are only warm. Things seem to have stabilized in terms of temps it would seem, as prior to now I couldn't get the machine past the BIOS, and sometimes not even into the BIOS. It's still a concern though.

So here's the REAL problem at this point:
The damned thing won't run at it's native clock speed. If I try to put it to 1.8ghz (either at 13.5x, 133/33, or 14.0x, 133/33), the power on the machine cuts out after about 2 seconds of my having turning it on. I then have to hard power it off, then turn it back on again, so that the mobo will reset the clock speed back to its preferred 1.35ghz. I have the latest BIOS revision (1008), and am not sure what else to be done. I'm thinking that maybe it's not feeding the processor enough voltage for it to actually run at it's actual clock speed. The mobo says it's feeding it 1.65v, and I recall reading somewhere that this is adequate.

Any help on this matter would be greatly appreciated, thanks.

 

[DAPUNISHER]

Presuming your water block is installed correctly, what PSU are you using? Have you tried manually increasing the Vcore? BTW, It helps to list full system specs

 

[mechBgon]

Number 1 thing to keep in mind: the only thing that sees the on-CPU temperature is the C.O.P. circuit. The temp you see in the BIOS and in Asus Probe is from a surface-mounted thermistor on the surface of the motherboard. Typical temperatures are 50C to 70C depending on the quality of your cooling solution.

As for your troubles with the immediate shutdown, here are some initial brainstorms to bounce off of you:

• Is your power supply up to the challenge of a 2200+? Of course, if you're running watercooling, I'm guessing you got a quality PSU also.
• What brand/model is your RAM? Have you tried manually setting it to PC2100 speeds with high timings as a troubleshooting step? Of the people I've helped with A7V333 problems, RAM timing seems to be the source of many woes.
• Could the CPU be overheating right away and triggering the C.O.P. into cutting power? The block needs to be clamped against the CPU hard enough to compress the cornerpads on the CPU, or you won't get good thermal transfer through the "pillow" of thermal compound between the block and the CPU core. Given the wide range of temps you're seeing, this could be the problem. Additionally, the base of the block must not touch anything but the cornerpads and the CPU core, or it won't get square contact.

Hope something comes to light for you... let us know how it goes.

 

[SpaceNinja]

First things first, my psu is an Enermax 530w Whisper series, so there's no problem on that end of the spectrum.

As for the inability to get the clock speed up, I updated to the 1009.001 beta BIOS revision, and that did the trick. Also, it no longer shuts down right after turning it on.

However, overheating still seems to be an issue. The thermistor on the waterblock thinks the cpu is at 31C, whereas the BIOS still thinks it's very high (about the same as before). I'm not sure what to do. I double checked to make sure that the thermistor on the block was properly in place, and still no dice. I've also gone through and tried both Arctic Silver II and the thermal grease that comes with the Koolance. I've made sure several (ie: many) times that the waterblock was firmly in place, that the screw was situated over the core (on the cpu-200 cooler), and that I had tightened it as far as the screw would permit (if too much stress is applied, it "clicks" and won't tighten any further, thus to prevent you from overtightening and inadverently crushing the core).

I'm positive that the motherboard is right about the CPU temperature, as it shut down again but a few minutes ago, before I even managed to make it to WinXP. I immediately shut everything down, unplugged the power from the motherboard, and took the waterblock off. Sure enough, the CPU was quite warm.

Edited to say: Thanks for all the help

 

[dwifuzi]

I am using a7333-raid. I had that same problem when I tried to overclock with stupid coolermaster hsf. The BIOS detects that the cpu is too. that is why it shuts down in two sec, because this board has overheat protection. However, I read somewhere saying that the board reading of the cpu temperature is off by about 10 degree celcius. I also heard that there is a BIOS 1008(by lumberjacK) that is good. Also, I think there is a Asus BIOS 1011 that has already solved that problem. I suggest you go and read the newsgroup in amdb.com. ohyea, 1.65v is definitely not enough. Don't be shy about jacking the vcore voltage up to 1.8-1.85 or even more. This shouldn't affect the cpu temp by much. try to clock the FSB and lower multiplier to match find your correct speed, but usually the mobo can auto detect the chip speed. I don't know in xp2200 case since it's new chip. ohyea, also, make sure all the jumpers are on the right place. Btw, if you want to use usb 2.0 port, you have to set the "+5v" instead of the "default". I just setup my usb2.0 burner today. took me an hour to figure it out. hehe... Anyway, good luck. LMK how it turns out or need some more helps, I would be willing to help troubleshoot. I want to know more about this board too. I just got it for a 3 weeks. it's up and running great so far.

 

[SpaceNinja]

The 1011 BIOS works... kind of. It allows me to properly set the CPU speed, but on the downside has some major issues such as the RAID not working. So far 1009.001 seems to work fine, but if I experience more problems, I've already got the Lumberjack 1008 version of the BIOS ready on a floppy.

As for the USB 2.0 issue, I already figured that out on my own, but thanks for the advice anyways ^_^

As for the multiplier and what have you: I've got the multiplier at 13.5x, and the frequency set to 133/33. It seems to work fine.

I'm just still worried about overheating.

 

[mechBgon]

If the CPU temp reads high, and the waterblock temp reads low, that tells me that there is poor thermal transfer between them. Make sense?

The two reasons I'd suspect are that either the clamp is still not applying enough pressure to overcome the cornerpads, or the waterblock is touching the cambox of the CPU socket and not making square contact with the CPU die. There is so much free space around the CPU socket that I doubt it could be hung up on a capacitor or any other obstruction.

 

[SpaceNinja]

Hrm... I'll have to check to see if it's being obstructed by the cambox or not. It could very well be, now that you mention it. I've noticed on several accounts after having pulled off the block that only about half of the thermal paste has made marks on the waterblock.

EDIT:
Alright, having had checked, the block is indeed making quite sufficient contact with the core. The Arctic Silver left evidence on the waterblock to prove that the core is definitely making firm contact.

 

[mechBgon]

If that ends up being the issue, then it's kind of reassuring to know that C.O.P. is actually working, and keeping the CPU temperature from getting out of control.

 

[SpaceNinja]

I suppose the C.O.P. is working, but I still need to get this cooling thing under control. I'm very concerned about this now, as it's still having problems. Maybe I'm being stupid and am missing some very important step somehwere, as this shouldn't be happening. Sigh...