Ok I've got an old (Well I don't know how old) nVidia card. That I took out of my dad's 'gaming' computer because his new MOBO wouldn't support it, so I threw in my old Radeo 9600SE (That's where that card went Rampy ;D) anyways...
I've got this card in front of me(Same one I tried for HTPC a few weeks ago. And I thought is was a FX5200 or 5x00 because of the nice HS's and RAMsinks. But I'm not leaning toward GF3 or 4 dunno.
Anyone know of a site that can help you determine nVidia cards by looking at them, or even after they are installed on the machine, I would like to know what card it is because I'm dumping it in a system that is forsale. All it says on it is, nVidia Model:P50 (I think that's the GPU) Anyways it doesn't have a lot of caps on it like the newer cards so this is why I'm leaning pre GF4
Thanks for the help in adavance!
Typically a card will identify itself on boot, this is seen before the memory count screen. Some systems have a logo display instead, but you can disable that in CMOS so you can see the memory count and post of the video card. It'll have the model # and such in that post info on boot.
I generally put the card on my desk and bring a desk lamp really close to it and start grilling it!
"Who do you work for!"
"What's your model number!"
"Where are the secret plans!"
"I'm going to start breaking off capacitors one by one until you start talking mister..." ;D
Use the tips above (except breaking off caps :o) and if you want a little more information than what you see by opening up Device Manager, try downloading Adia with the video card in the slot. It?s quick and easy.
Twenty bucks says it a Geforce3 model.
I always use my own method-download and run AIDA32 or Everest.
Both programs actually query hardware, rather than relying on less accurate
information from PnP driver routines. Either one should tell you
which card you have. They've never failed me yet.
Although this thread is older than moses, so is this graphics card which I also own.
iT is a GeForce 3 ti200/500. I have an OEM version, which most people have, and is simply an exact duplicate of Nvidia's reference card. The reason is simple... Designing a high-speed graphics card is no easy task. The only way ANY graphics GPU vendor can sell chips, is to sell them to vendors who typically all duplicate the reference design.
Now, that being said, depending upon the age of your GeForce 3 card, the RAM speed will differ. The one I have with the same "model: P50" mark, is simply the same reference design, and in my case, is the ti200... But look at the RAM... Mine is 4ns RAM, the same as used in the ti500. If you enable the Nvidia "cool bits" registry key, it enables the full user i/f with a clocking tab. You can then overclock the ti200 to above the ti500 speeds which at the time, sold for a huge premium.
However, again, since these cards are alll based on the reference design, it's totally up to the company selling/making them, how they want to clock them, heatsink them, and sell them as...
Like I said, if your RAM is under 6ns, then it's used on both the ti200 and ti500.... Thus, the only difference is the clock speeds set by the mfr., and, the heatsinking on the RAM and forced cooling on the GPU.
Since mine was a ti-200 OEM from Dell, the RAM had no heatsinks, and the GPU had a passive heatsink. I enabled the OC settings and after a stern warning from Nvidea's driver about "possible HW damage", you can then set the clocks way higher than the HW can manage.
Also, there's space for TV-out HW, but it's deleted from the OEM version. It does have Both analog and digital video outputs however.
By adding a CPU fan to the GPU, and glueing on makeshift RAM heatsinks, I could crank up both the GPU and RAM clocks to above the ti-500 premium card.
WARNING... If you do decide to OC, walk it up slowly, and play some graphics 3d game to heat it up... At best. you will hit the edge and begin to see sparkles or similar artifacts... This means you're maxing out the chipset thermally. Depending upon your cooling, you can max out both the RAM, and GPU but if you go too far and "crank it", you WILL lock up the system.
I forget the numbers I can crank too, but it's above the ti-500 using an add-on GPU fan, and passive heatsinks on the RAM.
Like I said, this is an OEM cheap-ass version... The $$$ Geforce 3's have major heatsinks, and fans!
Again, the GPU's are both the same... It's the RAM speed which was different between the ti-200 and ti500. Later cards were sold as ti-200s, but have the 4ns RAM which you can clock up to 250MHz compared to the stock 175MHz. I've maxed out at 255 for the RAM and things start to artifact (malfunction). The GPU can take 450MHz which puts the ti-200 card above the ti-500. The cooler you keep the chips, the higher you can clock them.
For the braver soles, you can also up the GPU core Vcc and crank it more, but you would definately need some major cooling for it to function... The newer Nvidea chipsets have thermal monitors... AFAIK, the Geforce 3 chips have no thermal reporting ability though that could simply be a BIOS limitation, a realm I desire not to walk into without a backup card...! I do have a BIOS editor for Nvidea, but again, w/o a backup card, I ain't walking that road!
Finally, there are several revision numbers on the card:
The PCB, as mentioned prior is "model P50".
The main M/N is MS-8851, V120.
The assy build version is "Rev A03".
And the FCC reg. number.. (the actual mfr) is TW-07j 062-69700-1AD-1494.
(that would be Thaiwan)
This is the number all cards must have and can be looked up at the FCC web site to determine who actually builds the cards.
If you want to overclock yer GeForce 3, lookup "coolbits". it's a single registry key which you add, and set the value to 0x01 (I think). Look it up...
Good luck and regards,
P.S. OK, Here'sa a HUGE possible issue. I just puled my system apart to un-dust it and alarmingly, some PS caps on the GeForce are failing and you MUST check them before they fail!
right side, component side of card, are 5 electrolytic caps, two small, three large. Of the three large, two are bulging and one is close to blowing it's seal and shows signs of leaking electrolyte. A second one is buldging, but the seal is intact. The third looks normal.
These caps are on the main power planes. Years ago I read about how many Aibit MOBOs had defective caps from the same vendor and people had caps blowing up. Never heard of any issue with Nvidia cards, but I haven't fully researched it and the visual inspection shows a clear problem which I must correct so the card does not fail completely.
Each capacitor spot has 4 to 5 component ID's which is kind of unusual. But these are the three large 470µF/16VDC caps at the edge of the PCB. Other model cards may differ. but look at the tops of the larger caps. The cans have several "score marks" which are there to alow the can to vent if the internal pressure gets to high. Like a soda can, with a weakened pop tab, so it can blow...
Anyway, hope some people still have these now ancient cards, got some info from this post, and maybe caught those caps before they failed... Looks like I need to get out the soldering iron...
I do some computer service work for locals. Strange, I worked on two machines this week that had video problems. Both are Dell 4600 machines from around 2004 with an Nvidia GEForce fX5200 card in them. They both have bad capacitors. There are four Skywell SHT 6.3V 1000uf caps on both boards all are bulged or leaking. I can only assume what the SHT stands for... maybe they missed a letter.
Both boards were displaying the same problem.... the monitor goes into standby mode. One was doing it right after the initial WinXP boot screen, the other was more random. I removed the cards, and plugged the monitor into the onboard video connector and all is well.
I'm surprised there wasn't a recall on these cards after they were put into so many Dell Machines. Shame on Nvidia and especially shame on Dell!