Are these 8800GT for Apple, and how can I test them

[dicky96]

Hi guys
I picked up two Geforce 8800GT graphics cards at the flea market for not a lot of money, and looking on ebay, the only 8800GT I can see that look like these are for some Apple Mac computer

On the plus side, looking at sold listings, they are worth quite a bit more than the PC version, but on the downside I have no idea how to test them!

I had a look around them, can't see any shorts or anything obvious - they are PCIe cards but can I just plug them into a PC? Are they Apple GPUs?

Is there any way to test them as I don't have any Apple computers here?

Could I reprogram the BIOS maybe, to make them work in a PC to test, then flash them back with Apple BIOS afterwards (on the basis that if they work in a PC they would almost certainly work in the Apple)?

[ChaosLegionnaire]

Re: Are these 8800GT for Apple, and how can I test them

what other labels or stickers are there on the reverse side of the video card?

[dicky96]

Re: Are these 8800GT for Apple, and how can I test them

Sure Ill send some pics. I could just stick them in a PC and see what happens I suppose.

The reason I thought they were for Apple is because when I search on ebay for 8800 GT the only ones I can see that look like that are for Apple MacPro. All the rest (for PC) have 'gamers' stickers on them with mermaids and big guns and stuff.

https://www.ebay.co.uk/itm/Apple-630...8AAOSw0sZf~yaq

And If I search ebay for 8800 GT Apple then the all look like that
https://www.ebay.co.uk/sch/i.html?_f...Apple&_sacat=0

[dicky96]

Re: Are these 8800GT for Apple, and how can I test them

Pics as requested

[R_J]

Re: Are these 8800GT for Apple, and how can I test them

I don't see why it wouldn't work in a pc, it seems to be PCI-Express 2.0 x16 interface https://www.techpowerup.com/gpu-spec...e-8800-gt.c201 You likely need the 6 pin power plug connected. I believe the apple are pci-e 1.0 bus and the card will shift speed to work at pci 1.0

[dicky96]

Re: Are these 8800GT for Apple, and how can I test them

OK thanks R_J

I'll stick them in a PC tomorrow and try them. One thing puzzles me though, if these work in a PC, then one would assume a PC 8800 GT works in an Apple - so why do people pay 3x to 4x the price for used Apple cards (which they clearly do) vs the same model used PC card?

[R_J]

Re: Are these 8800GT for Apple, and how can I test them

Maybe I am wrong:
Here is what seems like the same card https://www.ebay.ca/itm/Apple-Mac-Pr...YAAOxyHt9R0Sni

Compatible machines include:

•2006-07 (1,1/2,1) Intel Mac Pro machines only (will not work in a 2008-09 2nd Gen. (3,1) Mac Pro, Dual/Quad G5 PowerMac, or a PC)

Also here is some info https://forums.guru3d.com/threads/ho...8800gt.357612/

[piernov]

Re: Are these 8800GT for Apple, and how can I test them

Graphics card needs a specific VBIOS to work properly in Apple machines for example to be able to show boot screen. Even without the correct VBIOS it'll often still work under macOS, but you boot "blind" until it has fully booted (and some features may not work).

[ribcage]

Re: Are these 8800GT for Apple, and how can I test them

You could try making a backup of the bios with GPU-Z or the bootable version of NVflash, then try flashing a bios from a similar PC version of 8800GT. Perhaps linux based OSes would accept the card as it is?

[momaka]

Re: Are these 8800GT for Apple, and how can I test them

The reason I thought they were for Apple is because when I search on ebay for 8800 GT the only ones I can see that look like that are for Apple MacPro. All the rest (for PC) have 'gamers' stickers on them with mermaids and big guns and stuff.

They could be OEM cards - i.e. something that was included with an HP or Dell build from the factory. While it's rare to see gaming GPUs in HPs and Dells, they do exist. Or it could also be from an AlienWare PC, since those are technically OEM'd by Dell too.

In fact, this picture...
https://www.badcaps.net/forum/attach...8&d=1616580425
... confirms at least one of your cards was OEM'd for Dell, based on that little white sticker in the upper right corner.

How I know? - Whenever you see a sticker starting with "DP/N ....", that's basically Dell Part Number, and then the letters and numbers that follow are specific to that part. In your case, that card is a "0CP187" or "CN-0CP187". This is basically the same thing as Sony's internal part number format system. I think HP has one too, but I forgot what theirs looks like. I guess I can say I'm more of a Dell guy, lol.

Graphics card needs a specific VBIOS to work properly in Apple machines for example to be able to show boot screen. Even without the correct VBIOS it'll often still work under macOS, but you boot "blind" until it has fully booted (and some features may not work).

^ This.
You can use pretty much any Apple PCI-E card in a PC without problems. But if you use a regular PC card in a Mac, you loose the boot screen, and possibly fan control in BIOS (and maybe OS too??) So that's the only difference between Apple and PC video cards. People used to make a lot of money off of this by flashing a regular card to an Apple version and then re-selling it for more. But I don't think this is as popular of a thing anymore.

Nice score on those 8800 GT cards, by the way!

The only suggestion I have for you for those is to upgrade their cooler to something dual-slot, if you have any spare dead GPUs whose coolers would fit. Otherwise these 8800 GT cards run mad-hot under load (80C is not unusual at all.) Another option would be to try and flash their BIOS to 8800 GT PE (power-efficient) or 9800 GT PE version (I think PNY and a few others had these.) This will reduce the card's voltages and clocks slightly, but will also tremendously decrease the heat output as well. Hopefully ChaosLegionnaire will chime in here and say if that's possible or not with these, though. He definitely knows a lot more when it comes to these older nVidia cards than I.

[dicky96]

Re: Are these 8800GT for Apple, and how can I test them

So I tried them in a PC
There is definitely something different between the two cards because with one of them the PC bleeps to say no graphics card present - then continues to boot, but no picture, obviously. On this one something near the 12V EATX connector is getting quite warm (I can feel the heat on the back of hte PCB, I need to check what it is)

The other one does not bleep, the POST analyzer card shows the PC locks up, then after a minute tries again and locks up... and the same area of the PCB is not getting anywhere near as warm

So I need to look into what is the difference. When I checked the boards originally, without powering up, the resistances on all the buck coils matched between both cards.

[piernov]

Re: Are these 8800GT for Apple, and how can I test them

These are affected by bumpgate so I wouldn't be surprised if the GPU on the second one is dead at least.

[Dan81]

Re: Are these 8800GT for Apple, and how can I test them

Not quite - from what it looks (and the slim info I could gather), these Dell cards use the G92 core, not the original G80 core, so IIRC anything from the G9x series (except maybe a few chips that were made on the old G84/G86 cores, like the desktop 9300GS (G86), the mobile 9300MG (also G86) and the mobile 9500M GS (which is a rebranded 8600M GT) should be free of bumpgate issues?

[piernov]

Re: Are these 8800GT for Apple, and how can I test them

The early G92 were also affected by bumpgate. Almost everything manufactured before approximately the summer of 2008.
https://www.badcaps.net/forum/showthread.php?t=94383

[dicky96]

Re: Are these 8800GT for Apple, and how can I test them

OK just an update - the one that was getting hot was because one of the power rails to the GPU went to 2.3R on the bad one, but was about 200R on the other one. Previously both read about 200R.

While probing around it actually went back to 200R (like it is intermittent) and by gently flexing the PCB I could get it to do that a couple of times but then it went to 2.3R and stayed that way

The 'short' is on the load side - I couldn't get enough current into it to get anything to warm up, but I could tell by injecting 1.175V (got the voltage from the other board) and then using my meter on 1.999 millivolts range, that the problem was either the GPU or one of the MLCC on the rear side of the board behind the GPU ,as I could clearly see a voltage drop across the traces from the buck coil with about 0.6A flowing.

As I could initially change the 200R/2.3R by flexing the board I suspected a cracked MLCC so I traced all the MLCC I could find on that rail and desoldered them but it made no difference so I then removed the GPU. The 'short' was then gone, but unfortunately I used too hot a profile and damaged the GPU laminate. The PCB was not damaged.

I can get replacement GPU for these boards from Ali, for about 8 euro, but I don't know if I will bother. Maybe I will just for educational/experience purposes.

I'll try taking the BIOS off the (now) fried one and compare it to the BIOS on the other one (the one that hangs the PC on booting). If there is any difference I will try reprogramming the BIOS. If not I will try reflowing that one and see if it then wants to work

Rich.

[ChaosLegionnaire]

Re: Are these 8800GT for Apple, and how can I test them

that is an interesting failure mode for the gpu. im guessing it ran so hot the tin whiskers grew like mad until it shorted out to an adjacent pad. it also looks like it was the bga substrate to pcb solder connections that shorted since flexing the pcb fixes the short temporarily. usually its the solder bumps between the die and substrate that short out. flexing the pcb wont fix that at all. lol!

ah well, at least u were wise and smart and bought two so u had two to compare with and against. u also have one that u can use to fix the other if one of them is a goner. good one!

[momaka]

Re: Are these 8800GT for Apple, and how can I test them

The early G92 were also affected by bumpgate. Almost everything manufactured before approximately the summer of 2008.
https://www.badcaps.net/forum/showthread.php?t=94383

Well, affected or not (they probably are ) those high temperatures will cook any GPU to death over time. At least from what I've seen from my own g92 8800/9800 GT video cards, they will *idle* around at least 55-60C with the stock cooler and fan speed. Push them a little bit, and they quickly jump in the mid-80's Celsius or even roaming in the low 90's with a higher room temperature.

As I could initially change the 200R/2.3R by flexing the board I suspected a cracked MLCC so I traced all the MLCC I could find on that rail and desoldered them but it made no difference so I then removed the GPU. The 'short' was then gone, but unfortunately I used too hot a profile and damaged the GPU laminate. The PCB was not damaged.

No big loss. That GPU chip was probably dead anyways. If you don't believe me, try finding the contacts on the GPU chip that connect to the same rail where you measured the short-circuit (by those resistance numbers when the short-circuit was gone, that looks like the GPU V_tt?), then measure that rail on the GPU chip to its ground contacts. Good chances it's still probably showing shorted or shorting out and clearing when you flex it.

I can get replacement GPU for these boards from Ali, for about 8 euro, but I don't know if I will bother. Maybe I will just for educational/experience purposes.

I actually do recall having a spare g92 chip from a recycled card, but no idea if it works or not. I got it with a scrap CPU lot, and someone had removed this GPU chip with a screw driver or something like that, as most of the solder balls were still stuck to the chip (and small scrapes could be seen on the chip substrate from the tool that was used to remove it.) Luckily, none of the pads were lifted. I cleaned it off and all pads appear good. Attempted a manual reball (since I have no stencil for it), but never finished it when I had a temporary issue with my hot air station. So I still got that chip somewhere, lol. Like you say, though, not really worth fixing these unless doing it for educational/experience purposes.

If not I will try reflowing that one and see if it then wants to work

Or a re-ball, if you feel like putting in the extra work... but probably not worth it.
And either way, if you do get the card to work, you will need to ditch that stock cooler for something dual-slot and much beefier that can handle 100 Watts of heat without letting the GPU go past 55C. Otherwise, the issue will likely come back. The stock coolers on these cards are such a joke - at least for the TDP of the card. Otherwise, they aren't actually that bad. I'm in the process of transferring/modifying one of those stock 8800/9800 GT coolers to a 7900 GS/GT (about 49W TDP)... and with any luck, perhaps that would be a good way to re-use my 8800/9800 GT stock coolers while also improving the cooling on the older 7900 GS/GT cards. The stock coolers on most GeForce 7900 cards are also a lame joke. They can handle maybe ~25W of TDP well. 30-35 is pushing it (if you want to stay under 60C, which with these bumpgated cards is a must, really.) And 45-50W is no good for them, which is what most 7900 cards will dump when loaded up.