• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

P6T PCI-e going out?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

muddocktor

Retired
Joined
Nov 1, 2001
Location
New Iberia, LA
Hey guys, I ran into a problem last night with my test system. I just rigged up to start testing a new heatsink for review and after installing it looks like I might have had a PCI-e bus failure on the board. I went to boot after the installation and got 1 long and 3 short beeps and no video. I took the vid card out and reseated it, no good. I then replaced it with a 6600GT card was was working when I took it out of service and no good. I then went and tried installing in the second PCI-e slot and still the same beep code and no boot with both video cards. Finally, I removed the PCI-e vid cards and installed an old PCI Radeon 9250 vid card I keep for testing and booted and runs fine. Before going to the new heatsink, the 7900GTX vid card that was installed was doing just fine and I didn't remove the mobo or any connectors off it when taking the old heatsink off and installing the new one. I even cleared the cmos and no good either.

The motherboard is a vanilla Asus P6T. Anyone else have this board and have it doing weird stuff? Also, I am wondering if somehow the new heatsink's mount could be doing something to the PCI-e interface on this mobo too. After I make my first set of runs on the board with this heatsink and go for mount #2, I am planning to reinstall the heatsink I had on it that everything was running fine with and boot with the 7900GTX back in and see if everything comes up.

So what do you think is going on?
 
:facepalm: sorry, I just REALLY wanted to use that smiley :), I imagine that's what you were doing when you had this problem.

I'm sure you have seen it, but below I found the beep code list from your mobo's manual.

From further googling, it looks like most signs point towards a video card failure, or VRAM failure, leading me to believe it would be the card itself, not the slot. Since you tried multiple video cards with the same result, that may not be your issue.

I also found several posts indicating that people had discovered the problem was with their system RAM--and we all know how finicky RAM can be at times. Before you RMA or scrap the board, I would try taking out all your RAM and see if the code reproduces. If it does not, just start trying different combinations of sticks until you find the bad one. If it does reproduce the same beep, maybe you do have a bad PCI slot--worth a try either way. It seems like if you had a physical issue with the board's hardware you would see the last beep code in the list; 1 long, followed by 4 short.

For what it's worth, I have a very similar board (see sig) and I have not had this issue. My board is about 2.5 yrs old, and I have had 2 PCIe slots occupied the entire time.
 

Attachments

  • BeepCodesP6X58Prem.png
    BeepCodesP6X58Prem.png
    18 KB · Views: 58
Last edited:
You plugged in the power connector?? :D j/k

It's possible that the heatsink mounting is tweaking the motherboard enough to cause some NB issues.
 
possibility that you could just reseat it a few times to make sure its not a connection. has worked for me several times before.

has this board ever been cold?
 
The board has never been cold; been my test rig board with heatsinks it's whole life. And I did try reseating it several times with 2 different PCI-e vid cards in the top 2 PCI-e slots (the blue ones). Every time has brought up the same 1 long 3 short beep code and no boot. But, if I throw the old PCI Radeon 9250 vid card in a PCI slot it boots and runs just fine. I also finished my first run with the heatsink I'm testing and then installed the True Spirit I had installed on it before I took her down for the new heatsink tests, which the vid card was working just fine with. Now, I get the same problem with it as with the new heatsink I'm testing. I guess I can take another system down and throw both of those older vid cards in to see if they will boot on a different mobo though. That is a real PITA though, since that machine is running just fine as it sits.:rain:

I guess I'll try that and post back the results of the testing on the other board. BTW, the other system I will try it on is a Biostar board P67 board.
 
OK, just threw both cards into the P67 system. The 6600GT ran with no problems at all. The 7900gtx looks like it's toast. I got a post beep but no picture and heard 8 faint short beeps. After that, I threw the 6600GT into the test rig and it booted fine. The only difference between now and when I tried it before was that I uninstalled the nvidia drivers. That shouldn't have affected it's operation during post though.:rain: Oh well, at least my board isn't screwed up; that's a relief. Because it's hard to find a vanilla P6T for a decent price nowadays and I would rather stick with the same board to minimize differences in testing.

Now for the million dollar question: What vid card to a I replace my old 7900gtx with? I want something with the same general layout (dual slot config, venting heat from the back of the case) and approximately the same thermals? Anyone got any suggestions? And I'm not shy about buying from the Classies either, if you have seen something that would make a good candidate.

Just to help figure thermals on this old dead card, it needed 1-6 pin pci-e plug for power. And the cheaper the better too, as I don't have unlimited funds.;)
 
nice to see that it is the lesser of 2 evils. thats for sure. and we appreciate all you do in the testing world for us.

thanks Jim, you rock
 
Thanks dejo, I appreciate that.

Any suggestions on a video card perhaps?

BTW, I believe my old vid card is still under warranty too. I just checked at eVGA and I bought it a few months before they started requiring registration for the lifetime warranty. :D I wonder what they will offer as a replacement?
 
Well, I just order an eVGA GTX 550 Ti from the Egg as a replacement. It looks like it should fit the bill pretty well. 1 6 pin power plug and has a full length shroud to vent air out the back of the case, pretty much like my old 7900GTX was set up. Plus, I imagine it can be used as a backup rig in a pinch, since the graphics won't be totally crappy.
 
BTW, I believe my old vid card is still under warranty too. I just checked at eVGA and I bought it a few months before they started requiring registration for the lifetime warranty. :D I wonder what they will offer as a replacement?

The TDP on the 7900GTX was around 85W, on a 550Ti I believe it's around 115W or so. So you'll be putting out a little bit more heat, but probably not a noticeable difference. Nothing an increase in the fan speed couldn't equalize.

Tell eVGA that you want a GTX580 as a replacement... it was the 580 of it's day... it should be justifiable :D

Glad it wasn't the board!
 
I would bet that the gpu wont make much difference in cpu temps from the change in ambient inside the case, as that card should exhaust all heat to the outside of said case. Worst case is that you could try and rerun one of the old setups with known temps and see if there is any negligible temp differences from the gpu change and then know what if any difference there is and adjust your findings accordingly.

We do appreciate all that you do in the heatsink testing world. I have personally used your findings to my advantage on more than one occasion.
 
The TDP on the 7900GTX was around 85W, on a 550Ti I believe it's around 115W or so. So you'll be putting out a little bit more heat, but probably not a noticeable difference. Nothing an increase in the fan speed couldn't equalize.

Tell eVGA that you want a GTX580 as a replacement... it was the 580 of it's day... it should be justifiable :D

Glad it wasn't the board!

Heh, I just retrieved the invoice from newegg on it so I could upload it to eVGA and I paid $450 for that vid card back in the day and that price looks to be comparable to a 580GTX for sure. It would be nice if they would send me one, but I'm not holding my breath for a 580 either.:eek:

I would bet that the gpu wont make much difference in cpu temps from the change in ambient inside the case, as that card should exhaust all heat to the outside of said case. Worst case is that you could try and rerun one of the old setups with known temps and see if there is any negligible temp differences from the gpu change and then know what if any difference there is and adjust your findings accordingly.

We do appreciate all that you do in the heatsink testing world. I have personally used your findings to my advantage on more than one occasion.

One reason why I went with the eVGA version of the 550Ti is due to that very fact you are talking about. The shroud on the eVGA card goes all the way back to the pci slot covers with no breaks or gaps and the other cards I looked at that were cheaper had gaps between the slot covers and the shroud. And even if the gpu does draw a tad more when loaded, I would think that at idle (like when I am testing a heatsink, the newer gpus should actually draw a bit less power if anything. They have come a long way with power management on them in the last 5 years.

One other thing; Newegg was really on the ball with my order. I just received my UPS tracking number already for the vid card and I ordered this kind of late last night. :thup:
 
I am still having some weirdness going on with this damn motherboard. Now I am having spontaneous reboots and unstable vcore. Until recently, I could set 1.2815 (or something like that) in bios and with LLC on, have a stable load voltage of 1.304, with it very occasionally bouncing down to 1.296 all according to cpu-z. Now with it set at the same settings in bios I was jumping up to around 1.38 vcore under P95 load. I now have it set at 1.26 or so and I am still seeing load voltages around 1.328-1.336 volts, with LLC enabled . With LLC disabled, I see some horrid vdroop under load, just like it used to do when new. Also, the first spontaneous reboot today I was in the computer room while it was making a run and it did the 1 long and 3 short beep vid card trouble beep code. I shut it down with the power button and then rebooted it and it came up ok then. I'm just not feeling any confidence in the damn Asus board right now.

Addendum, at 18 minutes into the present P95 run I just saw the vcore hit 1.344 volts. Something ain't right here. This present run is totally shot as far as a recordable test run, but it's interesting to see what the damn thing's doing.
 
No, no bulging caps. The board uses solid caps anyways and not conventional electrolytic caps. And after I made my last post the board spontaneiously rebooted with about 5 minutes left of a 30 minute P95 run. I then tried to make a run with LLC off and had bumped vcore up to around 1.4v or so in bios and as soon as I started P95 I had another spontaneous reboot. I then went into bios and looked at the hardware monitor page and noticed that the monitored voltages were all whack and showing in red fonts, like they were way out of line. I do know better than to trust any voltage reading the board monitors, but they used to have only 1 of those showing in red fonts. So I grabbed my dmm and stuck the leads on the psu to check actual voltage against what the mobo was saying. The mobo said the 12v rail was at 13.992v, actual measured reading was 12.03v. Board showed 5v rail at 5.959v and actual measured voltage was 5.09v and board showed 3.3v rail at 2.912v and actual measured voltage was 3.38v. After leaving it running Memtest while I bought a new mobo from the Classies (and a good deal too :D ) I went and rebooted the system and set it to crunching Rosetta while I went to bed. When I got up this morning, it was still crunching it's little heart out and I know it had no spontaneous reboot last night because I have BOINC set up to not start automatically on boot. So I really don't know what the hell is going on with this board.

Oh, and BTW, I did 2 complete passes of Memtest86 with 0 errors. I didn't think I would have any as I am actually underclocking the ram while still running at stock timings.
 
Back