Is it bad to not have a 24-pin connector to my motherboard?

[Navid]

Run Prime95. Click on "Options".
Click on "Torture Test".
Click on "OK".
Now, it is running.

 

[w00t]

No, It's not bad

 

[Josh7289]

Okay. I ran Prime95 Torture Test for 7 hours and 53 minutes with 0 errors and 0 warnings. My CPU temperature was, at the end of the test, 49 degrees Celsius. I don't think that's bad, right?

 

[Navid]

No, that's fine. Did you run it with the case open or closed?
What is the voltage when Prime is running? If you did not measure, you can run it again. Just let it run for 5 minutes and measure the voltage while Prime is still running.

 

[Josh7289]

I ran it with the case open for about a half hour then I slid the side of the case back on -- so it was closed. I'll try to check voltage now.

 

[Navid]

Originally posted by: Josh7289
I ran it with the case open for about a half hour then I slid the side of the case back on -- so it was closed. I'll try to check voltage now.

That's good.

Measure the voltage while Prime is running and measure it while nothing is running at idle in Windows. Be careful, don't short any wires please.

 

[Navid]

Don't forget to perform the measurements that Zepper described.
Use this as a reference for those measurements.
http://pinouts.ru/Power/atx_v2_pinout.shtml

The picture at the top left is the picture of your motherboard connector. Make sure to disconnect the PSU from the wall first. Measure the following on the motherboard connector after disconnecting the PSU from it.

Measure the resistance between pins 10 and 11.
Measure the resistance between pins 22 and 23.
Measure the resistance between pins 12 and 13.
Measure the resistance between pins 24 and 19.

 

[Josh7289]

Okay, I just found my DMM, so my current voltage on the 12V rail is 12.06 V, sometimes 12.05 V.

I'm going to run Prime95 now and see what happens. I'll edit this post when I'm done.

EDIT: I ran Prime95 for 12 minutes flat and got the 12V rail fluctuating between 12.03 V - 12.05 V and the rare 12.06 V.

EDIT 2:

I set the DMM to the Ohm symbol and this is what it told me:

Originally posted by: Navid
Measure the resistance between pins 10 and 11. 00.6
Measure the resistance between pins 22 and 23. 00.6
Measure the resistance between pins 12 and 13. 0.L
Measure the resistance between pins 24 and 19. 00.5 - 00.6

 

[piasabird]

Even if you use an adapter the power supply has to be powerful enough to support the adapter.

 

[Navid]

When you measure those resistances, you should put the meter on the lowest Ohm setting. If you did, that's fine. If not, redo them.

If you are convinced that those four measurements are below 1 Ohm, that is proof that for your motherboard, the adapter will not acomplish anything.


Now, run the graphics program and measure the voltage again. Let it run for a minute or two and measure the voltage.

Then, close the case and let the program run and log the GPU and CPU temperatures.

What is the CPU and GPU temperature at idle?
What are they after running the program for 5 minutes?
How about 20 minutes?

 

[Josh7289]

You mean measure the voltage on the 12 V rail, right? Alright, I'll get to it. ^^

EDIT: Running RTHDRIBL at 640 x 480 my voltage rating kept fluctuating between 12.00 V and 12.01 V. Right now idling it is at 12.06 V and the occasional 12.05 V.

Should I continue to run RTHDRIBL at 640 x 480 for the rest of the temperature tests?

 

[Navid]

What are your CPU and GPU temperatures at idle?

Close the case and run the utility. It will raise the GPU and CPU temperatures both. Monitor and write down both temperatures as the utility runs. What are their maximums?

 

[Navid]

Originally posted by: Josh7289
Actually, using my 256 MB 6600 GT, whenever I play games, the game will lock up and eventually give me a BSOD for nv4disp.dll then crash and restart my PC.

This never happens when using my Radeon Xpress 200 integrated graphics.

How did you upgrade from the integrated Radeon graphics card to your Nvidia 6600GT?
Did you use Driver Cleaner pro to completely remove the Radeon drivers before installing the Nvidia drivers?

 

[Josh7289]

Here's what happened with temperatures with the case closed (which made my GPU temperatures raise by about 5 degrees Celsius):

Idle: CPU @ 31C -- GPU @ 49C
Five minutes into RTHDRIBL: CPU @ 43 C -- GPU @ 88 C

About 16 minutes into RTHDRIBL, my PC blacked out then came back with lower clock speeds on the GPU. My GPU was running before this happened at 89 C and about 47 FPS in RTHDRIBL. After the black out, it was running at 72 C and about 32 FPS. So...

20 minutes into RTHDRIBL: CPU @ 42 C -- GPU @ 72 C

EDIT:

Originally posted by: Navid

Originally posted by: Josh7289
Actually, using my 256 MB 6600 GT, whenever I play games, the game will lock up and eventually give me a BSOD for nv4disp.dll then crash and restart my PC.

This never happens when using my Radeon Xpress 200 integrated graphics.

How did you upgrade from the integrated Radeon graphics card to your Nvidia 6600GT?
Did you use Driver Cleaner pro to completely remove the Radeon drivers before installing the Nvidia drivers?

Definitely. Here's what I did, exactly:

Used the Catalyst Uninstaller Utility to remove ATI drivers.
Restarted to make sure there was nothing left in the Control Panel related to ATI.
Restarted again, but into safe mode.
Ran the newest version of Driver Cleaner Pro: Ran Cab Cleaner on both driver.cab and SP2.cab, then removed every single thing related to ATI in Driver Cleaner, and finally went through Program Files to check for and delete (if found, though I found nothing) anything ATI-related and deleted all ATI-related registry keys.
Restarted into normal mode.
Downloaded my motherboard south bridge drivers from ATI's website and tried to reinstall those to make sure my drivers for that were the newest.
Restarted and installed the Nvidia beta ForceWare 91.33 drivers.

All of these things were done with most if not all task bar icons closed and my network cable removed, unless I say I downloaded something.

 

[Navid]

89C is a little bit too high!

Now, reboot to bring the graphics clocks back to default.
Open the case to improve cooling.
Run the utility again, with the case open, and log the temperatures again?
Does it crash again?
What is the highest the GPU temperature reaches?

 

[Josh7289]

Well, I rebooted and tested with the case off, but I realized something weird. The temperatures I have given so far are from Everest. This time, I ran both Everest and the Nvidia control panel's temperature monitor and got different results, so here they are:

Idle: CPU @ 32 C -- GPU @ 49 C Everest -- GPU @ 46 C Nvidia
Five minutes into RTHDRIBL: CPU @ 43 C -- GPU @ 88 C and sometimes 83 C Everest -- GPU @ 73 C and sometimes 74 C Nvidia

Again, at about 15 or 16 minutes, my PC blacked out and came back with the GPU running slower and cooler:

20 minutes into RTHDRIBL: CPU @ 43 C -- GPU @ 71 C and sometimes 72 C Everest -- GPU @ 62 C Nvidia

 

[Navid]

Are you overclocking the graphics card?
Do you have the graphics card with its stock cooler or have you replaced it with an after market cooler?

 

[Josh7289]

I am not overclocking my 256 MB 6600 GT PCIe at all, and it still has the stock eVGA cooler.

 

[Navid]

Do you have a fan that you can manually hold to keep it cool for a few minutes?
You need to confirm that this is a heat problem. Then, there are ways to fix it. But, not before confirming.

You can even use a big fan you may have for cooling the room!

Either way, just open the case. Place the PC so that there is room for the fan. Place the fan such that air blows right over the graphics card.

By the way, after the PC freezes, there is no point in continuing the measurements. Just reboot.

 

[Josh7289]

Do you want me to blow air over the top or the bottom of the video card (the side with the HSF or the other side)? I'll look for something I can use...

 

[Navid]

Just put the fan so that it blows air horizontally towards the card right directed at the card. So, some of the air will go over the card and some of it under. The idea is to move the air around.

Do you have a case exhaust fan? How about an intake fan? It is quite possible that you may need to improve the cooling on the card as well as the case cooling.

The idea of the big fan with the open case is to remove the case cooling out of the picture. But, that is not always easy if you do not have a big fan!

I will be installing a new motherboard for the next few hours. So, I will not be around. But, will be back.

What is your graphics card part number? Does the cooler cover the memory also?

 

[Josh7289]

I have a 120 mm case exhaust fan, but that is it for case fans.

As for my video card, it is this:

http://www.evga.com/products/moreinfo.asp?pn=256-P2-N420-TX&family=17

256-P2-N420-TX

As you can see, the RAM chips on the card have heatsinks on them.

EDIT: Alright, here's what I did with the fan idea. I took the box fan from my window and shoved it up against the side of my PC with the case side off and started taking down temperatures under RTHDRIBL.

Idle: CPU @ 33 C -- GPU (Everest) @ 49 C -- GPU (Nvidia) @ 46 C
Five minutes in: CPU @ 44 C -- GPU (Everest) @ 87 C -- GPU (Nvidia) @ 72 C and sometimes 73 C
20 minutes in (no crash yet): CPU @ 44 C -- GPU (Everest) @ 88 C -- GPU (Nvidia) @ 74 C and sometimes 73 C

From here, I just walked away and kept the box fan against my PC. I checked back periodically and eventually saw the temps around this with the box fan, finally taking effect, on the side:

40 minutes in: CPU @ 41 C -- GPU (Everest) <=85 C -- GPU (Nvidia) @ ~65 C

I then took the fan away and waited about three minutes and got this:

CPU @ 42 C -- GPU (Everest) @ 84 C -- GPU (Nvidia) @ 70 C

I then closed RTHDRIBL and put the box fan back on the side and felt daring, so I opened up Battlefield 2 (Demo ) and tried playing that. It lasted a few minutes until it quickly blacked out and came back and quickly cut my FPS in half, and created artifacts and whatnot. I Alt+F4'd that and am now writing this.

EDIT 2: Just removed a sentence.

EDIT 3: Spelling. ><

EDIT 4: Clarification. w00t

EDIT 5: I was just playing another game (GunZ -- much lower-spec, if that even matters), and with the fan on the side it only locked up twice in about an hour or an hour and a half or so. The first time it locked up, maybe a half hour or 45 minutes in, it recovered with, most likely, lowered clock speeds, but since GunZ is a rather low-spec game, I didn't notice and it kept running smoothly. The next time it locked up, maybe an hour or an hour and a half in, my PC restarted.

I went back into the game quickly to clear some things up, with the fan off the side of my case now, and it locked up within five minutes.

 

[Navid]

You can benefit from improving the cooling of your graphics card based on those experiments.
But, I am surprised that it still crashes with a big fan at it!

You can correct me if I am wrong.
The way I see it, you can extend the length of time it takes your computer to crash by placing a fan blowing at it. That shows that we have a cooling problem and not a PSU drive capability problem, which was the initial suspect.

You can get an after market cooler for the graphics card and hopefully that will solve the problem. The eVGA warranty is not voided by that.

What is the room temperature?

You can post in the Video forum or the Cases & Cooling forum and ask for suggestions for an after market cooler.
Or, you can ask if anybody else has similar problems with 7600GT.

Wait for others to make suggestions too.

 

[Josh7289]

Actually, I'm not too sure about it being a heat problem anymore. I just took the card out of my PC and stuck it in my sister's. She was already running a GeForce 6100 IGP 64 MB on this Acer machine, so I just stuck my 6600 GT in and disabled her IGP. I didn't reinstall the drivers or anything. It's running ForceWare version 81.95. I downloaded Everest and ran RTHDRIBL and got these temperatures (again, the Nvidia control panel gave me different temperatures than Everest):

Idle: GPU @ 53 C (Everest) -- GPU @ 49 C (Nvidia)
Five minutes in: GPU @ 88 C (Everest) -- GPU @ 73 C (Nvidia)

From this point on, the system ran completely stable, so I took the measurements again after a whole hour:

One hour in: GPU @ 88 C (Everest) -- GPU @ 73 C (Nvidia)

The temperatures reached at their max that I saw 90 C in Everest and 75 C in the Nvidia control panel. Not once did the system lock up, and it was running hotter. I also played a game of GunZ and it never locked up. I was going to install the BF2 demo on her PC to test that out, but she said I was done messing around with her PC.

I would doubt that her PSU is more powerful than mine, because the PC was $290 at Circuit City a few weeks ago, but it is a 24-pin PSU. For some reason, though, the 8th wire from the PSU is not there. There are just the other 23 pins plugged into the 24-pin motherboard.

Her system specs are weaker than mine, but pretty much only in the RAM (512 MB vs. 1042 MB) and CPU (Sempron 3200+ Socket 939 Palermo vs. Athlon 64 3000+ Socket 939 Venice) departments. Her case and motherboard are both Micro ATX-sized, too, so that should contribute to heat. (My motherboard is also Micro ATX-sized, but my case is full ATX-sized).

This is definitely making me reconsider heat being the problem, since it is running hotter but perfectly stable in her machine.

Originally posted by: Navid
What is the room temperature?

Room temperature is usually around 75 degrees Fahrenheit to 85 degrees Fahrenheit, depending on the time of day and weather outside.

Originally posted by: Navid
Or, you can ask if anybody else has similar problems with 7600GT.

Actually, it's a 6600 GT.

EDIT: I was able to install and run the BF2 demo on my sister's machine, but at lowered settings since she only has 512 MB or RAM, and the game ran flawlessly. There were no lockups or crashes resulting from the GPU. This further makes me cautious about blaming this problem on the video card itself.

Oh, and if you were wondering why I didn't include CPU temperatures from my sister's PC, it's because Everest was giving me completely random CPU temperature readings ranging from 1 degree Celsius up to 59 degrees Celsius and possibly beyond. Whether I was running something intensive like RTHDRIBL or just idling, the CPU temperature reading just kept randomly flying around everywhere. This makes me question Everest's GPU temperature readings, but at least they were consistent.

 

[Josh7289]

Since the video card itself no longer seems to be causing the problem, does anyone have any suggestions as to what it could be or more tests for me to do?