Can a PCIe slot go 'bad' if the video card blows up?

A long while back I got a Physx card on the cheap, and installed it in the slot next to my graphics card. Not realizing that the fan on the GPU protrudes more when the case was upright, the GPU blew up months later because the fan couldn't spin properly. I promptly replaced the video card with a new one and moved the physx card (removed eventually). Ever since I get random hardlocks with display corruption with no rhyme or reason every once in a while. There is no overclocking in the machine and I've tried reseating the card and uninstalling and reinstalling the drivers just in case something funky was going on since they were both nvidia cards. In terms of debugging I can only assume its the "new" video card, the pci-e slot or the new card might be drawing to much power from low wattage power supply that came with this pre-built computer.

The situation is as follows, the machine is really old and I've been meaning to upgrade but been holding out for a broadwell or even for skylake to release (windows9 really going to go prebuilt again). I don't really want to dump money unnecessarily into the old computer but seeing as how I might start using the computer for more serious things again soon the random crashes are unacceptable. I was going to buy a video card that I could use in this machine and eventually in my future machine but I am scared to waste the money if the PCIe slot is to blame and it wind up frying the new card. So any suggestions on troubleshooting whats wrong with the system that is non destructive??? I don't have a different power supply to test my power theory, and am afaid that if the new card is boinked from being in a bad slot testing it in another computer I have at my disposal might "spread the disease". I know I'm being paranoid but money's tight can't afford to waste any.

So any ideas on how I could proceed? Many thanks in advance. Oh and the other option is to buy a haswell machine now, get a new video card and trash this one except the HD and DVD burner but I'd really prefer to wait.
 
I take it you don't have another GPU to test in order to determine if the GPU itself is bad?
To answer your question, yes, there are situations where damage to a card can lead to damage to the PCI-E bus itself. I'm not sure about this happening strictly from overheating.

The more likely case is that the newer GPU is exceeding your PSU's current capabilities. Can you tell us the GPU and the PSU?
 
I take it you don't have another GPU to test in order to determine if the GPU itself is bad?
To answer your question, yes, there are situations where damage to a card can lead to damage to the PCI-E bus itself. I'm not sure about this happening strictly from overheating.

The more likely case is that the newer GPU is exceeding your PSU's current capabilities. Can you tell us the GPU and the PSU?

I sort of do, first the motherboard has an intel IGP (945G chipset) so there's that. Second I have a ATI x1300 or x1300pro in another machine, but I didn't want to risk messing it up but again thats me being paranoid and cautious. I also avoided that because I wanted to avoid installing ati drivers on the system but if there's no other options I guess I'll do it.

The old GPU was a 8500GT 256MB the new one a GT240 1GB, I know neither draws that much power. One thing of note is I'm running Vista Home Premium 32bit with 3GB of RAM, perhaps driver issue with lack of address space... 300W power supply.

Oh and I should mention when I removed the old GPU the fan was a combination of melted and warped and I lost video completely don't remember if I got POST beeps.

Thanks for the help and prompt responses... and if you know of any web resources where they have solutions/discussions to similar problems could you share? Google failed me when I tried.
 
How old is the computer? What are the specs?

Besides the above specs, it has an Core 2 Duo at 2.00ghz an E2180. Integrated sound and all other PCI and PCIe cards have been removed for a long time now. One 7200rpm HD and one DVD/CD burner nothing much. I think the computer was from around 2005 or 2006 don't know or remember.
 
It's time to buy a new computer. No sense in waiting for Broadwell or Skylake since those will be very minor improvements compared to the gains you'll see coming from such and old system to a modern i3 or i5. And if you get an SSD I think your worldview will be altered.
 
So no advice in regards to troubleshooting this machine?

In regards to a new computer, I was considering the following refurbished computer available rather near me, what do you think?

http://www.microcenter.com/product/423699/Envy_700-074_Desktop_Computer_Refurbished

Intel Core i5-4430 Processor 3GHz
Microsoft Windows 8
12GB DDR3-1600 SDRAM
2TB 7,200RPM Hard Drive
Intel HD Graphics 4600
SuperMulti DVD Burner
15-in-1 Media Card Reader
10/100/1000 Network
802.11b/g/n Wireless
Bluetooth 4.0

It only uses 2 of the 4 dimm sockets, and has a 460watt power supply so I should be able to upgrade it without replacing the PS or wasting RAM sticks.
 
So no advice in regards to troubleshooting this machine?
Considering how old and unimpressive it is at this stage, it's probably best to consider it a write-off. No point in spending any money replacing components in any case.

Do be aware that power supplies are very commonly the source of system flakyness, and your PC has a number of years under its belt, and it's a 300W PSU as well...? There might not be a lot of margin left in that PSU (they age), and if the caps are starting to go bad that could explain your crashes, possibly.

Just speculation. It's really impossible to diagnose computers from across the internet unless the problem's so fundamentally simple that it basically boils down to "you forgot to stick in the power cord" or something like that. ;)
 
Agree with Grall. Most likely candidate is the PSU. They go bad. I've had 2 go bad over the last 7 years and all were higher-end, high-wattage PSUs.
 
My own current monster rig almost did not want to start up from a cold boot the other day. The inrush current limiter relay kept clicking on, and then off again, requiring a number of attempts before the PSU actually turned on completely.

...Not sure WTH is up with that really. Nothing good, probably! :LOL: PC works fine when it's running though so it's just a startup glitch...for now. These issues have a tendency to grow tho, so I'll have to keep an eye on it.
 
No idea. I assume the sucker's on the verge of breaking down on me. Can't even tell if it's the relay itself that's screwy or the electronics that controls it... If I was an electro-engineer with a scope and some probes I could probably find out myself what's wrong, but alas. That is not who I am. :(
 
So no advice in regards to troubleshooting this machine?

In regards to a new computer, I was considering the following refurbished computer available rather near me, what do you think?

http://www.microcenter.com/product/423699/Envy_700-074_Desktop_Computer_Refurbished

Intel Core i5-4430 Processor 3GHz
Microsoft Windows 8
12GB DDR3-1600 SDRAM
2TB 7,200RPM Hard Drive
Intel HD Graphics 4600
SuperMulti DVD Burner
15-in-1 Media Card Reader
10/100/1000 Network
802.11b/g/n Wireless
Bluetooth 4.0

It only uses 2 of the 4 dimm sockets, and has a 460watt power supply so I should be able to upgrade it without replacing the PS or wasting RAM sticks.

That looks like a right good deal. Add an SSD I would, and a GPU if you want to do any serious gaming on it (although the IGP will still be faster than what you had).

I do wonder about having a 4GB DIMM in one slot and 8GB in the other. Will it still run in dual channel mode like that?
 
AFAIK, dual channel operation requires matched (size and speed) DIMMs.
Make that rig a 16 GB RAM, add a small SSD for the OS and a GPU and you'll have a great rig.
 
Wait. Is that the PSU you told me to buy?
Indeed, yes! :devilish: Then again, everything breaks, so if we want something that won't ever fail we can't buy anything. You could wear out even an abacus if you wanted to...

Besides, just because my PSU might be getting cantankerous/sickly doesn't mean yours will - you're not superstitious are you?

AFAIK, dual channel operation requires matched (size and speed) DIMMs.
Modern (Intel atleast) CPUs can map oddly sized RAM sticks so that interleaved addressing is supported for the common address space (IE, say 2GB for a 2GB and a 4GB stick.) Unless you're overclocking your RAM the speed and timings would be set to a common denominator of the available JEDEC SPD profiles - or so I believe anyway... Of course, it's generally somewhat inadvisable to mix memory sticks; it might work but then again it might not. RAM problems can be very dangerous, leading not just to straight crashes but also drive/data corruption, so it's best to use a matched set of memory modules, preferably from the manufacturer's verified list if one is paranoid about compatibility... :)
 
Back
Top