This doesn't really have to do with QW, but QW is however the reason I need it to work.

I have the weirdest problem with my computer/GPUs.

When I bought this computer, AMD Ath 64+ dual core 4200+, it came with some ATI radeon shitcard, or well, it was alright. 512mb and ran very smooth, even with QW (i've heard ATI sucks for qw basically).

Anyway, after a few years the GPU broke, so I was trying to find a new card. In the meantime I used a 256mb shitcard, that one from ATI too.
After a few days - that one broke too. At this point I was getting a bit frustrated and started wondering if the power fried the cards or something.

However, I found another 256mb shitcard from ATI which I am using now.

Now, a friend bought a new computer, so he gave me his old GPU which is EXACTLY the same card as the one I had in the beginning. EXACTLY the same 512mb ATI Radeon, same everything.

But I won't get it to work.


Ever since I had the problems at first, most of the times when I boot my PC nothing happens with the monitor. It still says "saving power" or something and goes black, as if it had no contact whatsoever to the GPU. However, after a few tries (shutdowns+reboots), suddenly the monitor gets contact with the GPU and starts up like nothing happened. I have to do this every time I boot my PC, and it needs anything between 2 and 100 tries to have them connect to eachother.

So my thought was first; the cable from my monitor to my GPU is fucked - replaced it 3 times - problem was still there.
Then I started thinking, perhaps it's the monitor that has seen its last good days - so I replaced that as well - but the problem remained.
That left me with the GPU, so I changed between several cards - problem remained.

I have no idea whatsoever what the problem could be, but it's fucking annoying.


Anyway, back to the new card-thingy. The card that was exactly the same as this PC's first card.

When it didn't work I ofcourse thought if the card was actually intact, so I installed it on another computer. That computer however didn't even find the card, it acted like it wasn't there at all, so I started to get convinced that the card was broken. However when I installed it on a THIRD computer, everything worked all fine. The card ran smooth, didn't get hot, everything worked great!

So I installed it on this PC again, in hope that it might work this time. Booted the PC - but the monitor just won't react. Of course my first thought was that it was the same old problem that this PC has no matter which card is in it, so I rebooted a few times. After having rebooted about 50 times I gave up and realised that it's not gonna work. Besides, I was getting scared I was gonna burn the entire motherboard or something, or have the computer crash completely.

So... I installed the same old 256mb fucking shitcard back into the computer and booted -> still same problem remains, the monitor won't react. However after rebooting a couple of times - Tataah! The monitor reacted and since then I haven't shut the computer off, in fear that the monitor won't connect with the PC anymore.

So well... anyone know what the problem could be?

Note that I am waaaay too poor to get myself a new GPU. I do have other computers that I could play QW on, but somehow the sensitivity and shit like that feels different from computer to computer, even though I adjust the windows settings to the EXACT same values as on this PC. don't know why it's different so dont ask me

I'd appreciate any help =)

/Andeh

Could be caused by a few things.

The first thing that comes to my mind is that thing could be PSU related. Then we could think about BIOS version. Also, is this on AGP or PCI-E system? Also, if you have had your motherboard for a long time, it could be that some of the motherboard's capacitors (condensers) have worn off or dried out during the years and are barely functioning.

The easy way to check this is open your chassis and measure the voltage from various points of the power cabling (you have a voltmeter, right?) when your computer is running and under heavy load. Also, check the capacitors near CPU and AGP/PCI-E slot if they have expanded or even leaked (nasty brown or white shit).

Heya again,

I messed up a little in the first post. The new GPU that I was trying to install is infact a Nvidia GeForce 6600, and not an ATI. For some reason my computer no longer senses the Nvidia card but only the ATI cards. Any idea? And yes it's PCI-e, all of them.

Also, the motherboard thingy, I couldn't find any leaks and I'm not sure what to look for with the voltmeter as I don't know what values are supposed to be shown.

Another thing that concerns me is the extreme heat. Both the GPU and parts of the motherboard gets extremely hot when under some load. Basically, my motherboard should already be burned out. It's been up and running for almost 2 years WITHOUT SHUTDOWNS, meaning I haven't turned the computer off during nights for almost 2 years. Before that, it also had another 3 years running but with shutdowns, so logically my entire computer should be pretty tired.



This is a picture of the motherboard, taken from MSI's homepage.

You see the silver-plate-thingy in the lower right corner? That one almost reaches 100c and is impossible to even touch without burning your hand off. Also, even when playing quake, the GPU reaches 80+ degrees which is extremely bad. I'm getting a few fans now to put it under constant cooling but that'll only lower the degrees of the components, not erase the problem of WHY they're so hot.

What you think Renzo?

Basically, you are looking into this kinds of parts (capacitors):

They should not have any kind of indication of leaking or expansion (like a barrel).

When measuring voltage from molex, you are taking the measurement from xx.. or ..xx and not like .xx. or x..x . So, 12v+ground next to it or 5v+ground next to it. Example picture:


That being said, how did you measure the silver heatsink temp? It sounds [b]VERY[b] high temp for north/southbridge heatsink, should be like ~50°C or so. Going above 65°C will most likely cause severe instability and shorthen the life-span of the components.

Anyway, all of this is just guessing: there's just no way of knowing sure when guessing like this and not having the actual computer infront of me. So, it's hard to say anything for sure, but at least you have some things to be checked out.

I didn't understand much of the xx..xx ..xx..x.x.x.x..x.x..x.xx. stuff, but about the heat;

The silverplate measurement isn't sure as it was measured by hand. However I'm using some hardware sensor monitor program which checks the temp of other stuff in the computer like GPU. The GPU when playing is at 80-83 degrees celcius which is crazily much. By touching the GPU and then the silverplate I must say that the silverplate is hotter, or at least feels so.

I've now put a big motherfucking fan used to cool entire rooms right infront of the chassi (opened on the side of course) so now everything stays at around 40 degrees according to the program. I have no lights in this room unfortunately so I'll check that leak-thingy out the first thing tomorrow.

Thanks for the help Renzo, if any more ideas please let me know.

/Andeh

It means the pins, there are four pins in the molex, like so:

....

So if you want to measure the voltages, you choose the pins like the pic shows you, so either:

xx..

or

..xx

Where "x" means the pin you are taking the measurement from and the . is unused.



You have to realize that hand measurement is not ok, even 40C can feel really hot while 50C can feel like burning even. So the only way to measure temps are either some measurement tool or by infrared, the latter being out of the question of course

The GPU can handle up to ~110C as long as the power phase won't get that hot, it's not a problem.

But really, if the caps seem ok to you (it should be pretty obvious if they are bloated or leaked) and the voltage measurements give you somewhat stable 12V and 5V under load, then even that should be ok.

// EDIT: ok, it is a PCI-E system. How about updating your bios for newer version if such exists?

Alright because when the program shows the temperature of the GPU it also displays if it thinks the temp is ok or not, and mine shows totally red like "OMFG UR COMPUTER IS GONNA EXPLODE"-ish. Also went to a friend's house and tried it on his and his was a stable 20-40degrees even when using maximum graphics on quake 4. I was mad. :-D

I'll try updating the bios if a newer version exists, but what makes no sense to me is that this computer came along with the EXACT SAME graph card that I want to install now. The exact same version, and it was already installed by the company that sold the computer to me. But now for some reason it won't even notice theres a fuckin card in it when I boot it. Or well, I can't know that for sure, as my monitor says it's saving power whenever I try to boot the computer with the card in it. My monitor is like "eh fuck it" and doesn't even want to show sign of signal from the card.

Point being, with the card in, I can't even get to bios, I can't even get a loading screen or anything.

And it's fucking annoying that my pc doesn't have an internal graphic card to boot from alternatively -_-

Check power supply and change if possible (480W should be fine).

Update on the matter;

The reason I have to restart the computer loads of times before it picks up the monitor with the boot = heat. I've put a big fan next to it with the chassi open to keep it cool, and it starts every time as it should as long as it's cold.

However it won't sense the GeForce card, which is the exact same card that was used when I bought the computer.

@xpr: how can I check the power supply? I'm a newb.

Also couldn't find any leaks of any kind.

I have yet to check voltages, but first I have to get a tool to be able to.

/Andeh

Just replace it, if you don't want to buy new then ask friend When i has problems with video cards 80% of time it was a bad power supply.

I already gave you a way, but since you are missing the tool for that you could use some hw monitoring program (like here) to see the voltages. HOWEVER, doing it this way will not give you too reliable results, but at least you can see something. And if you are lucky, you will get to see real values and not some garbage.

Xpr's PSU recommendation sounds like most of the different US forums users giving advice. While ~500W good power supply (enermax, seasonic) is enough for a lot of things, you don't really need that much, especially considering your computer. Xbitlabs had a test regarding this not so long time ago, and the test results are interesting.

That being said, if you had to get a new PSU (and you are the type upgrading your PC from time to time), something like Enermax modu82+ 525W would be a good choice, but it costs more than slightly weaker unit. Also, keep away fro random manufacturer's pieces of shit, they are nothing but head-ache and trouble.

You should delete all the different GFX drivers installed, and then install the latest Geforce drivers. Seems like your past ATI drivers are in conflict with the current Nvidia drivers.