8800GT voltage mods

[gugle]

here with 1.50v was bad, now with 1.40v was good

WC
gpu 600 @ 810
shader 1500 @ 1944
memo 700 @ 738



1 Hour OK
600 @ 833
1500 @ 2052
700 @ 738


error 43 min
600 @ 842
1500 @ 2106
700 @ 738

[Luka_Aveiro]

here with 1.50v was bad, now with 1.40v was good

WC
gpu 600 @ 810
shader 1500 @ 1944
memo 700 @ 738



1 Hour OK
600 @ 833
1500 @ 2052
700 @ 738


error 43 min
600 @ 842
1500 @ 2106
700 @ 738

Bom overclock amigo. Saudações lusófonas

Nice overclock pal

[datdood916]

hey, i wanted to say that i did this mod, and i could not get my computer stable no matter what. Yes, synthetic benchmarks/stress tests like ATITOOL worked for as long as i had them running, but in the middle of playing F.E.A.R. 2, i get a red screen followed by a shutdown computer. I even put my video card back to stock speeds with a 1.18 vGPU (yes i realize that i slightly over on volt) and it still freezes and turns the screen red/noise/then crashes. I played cryostasis for about 10mins or so and it worked fine, it seems that the game is more cpu dependant then F.E.A.R. 2 is.

My setup is:
AMD x64 4800+ x2, evga 8800gt sc w/ vf1000 sink, 3gb ram, winXP
801/1873/stock @ 1.27v (with ATITOOL worked for 72mins before i aborted) played fear 2 for about 2 mins and red screen
801/stock/stock @ 1.27v, same as above
650/stock/stock @ 1.18v card works with ATITOOLS and plays fear 2 for a *little* longer before it gives me red screen

Question, if i take off the mod, will it correct my problem, or is the card fried? Also, has anyone in here played games for an extended period (hours) of time after this mod and if so, has anyone encountered red screens with noise in them, and crashes?

[Demo]

@datdood916

Had a similar problem but on stock speed and its not a red screen, its any color on the spectrum. all i did was up the PCIe freq to 105Mhz and it ever returned for over a year.

[datdood916]

I tried your suggestion with clockgen and put my PCIE up to 105mhz but it still crashes. Well, not shutdown computer crashes, but sometimes it will be a BSOD- happened probably 2 or 3 times out of the 10-15 times the computer the computer froze. Its actually not a solid color when it freezes, more like i dunno- a weird fog effect with black pixels in it. It looks like some weird shader effect maybe? Anyone else have this problem, or maybe know of any other solutions i could try?

[tenebre]

I have a Gigabyte 8800GT 512MB Zalman. Default clocks 700/1700/1980.
Is a software voltage mod possible? How much would it be considered safe? Right now it stays @ default clocks, 1.15V with a max of 71*C in Furmark maximum stress.

[natedog420]

datdood: ive had my 8800gt volt modded for over a year now and i have 100S of hours of gaming on mine at as high as 1.51 volts mostly around 1.45 volts. i have no problems at all. the problem your describing i would say undo the mod clean up all the soldier and try again. you might have some soldier were you dont want it so try that.

[datdood916]

lol, all of my soldiers are clean. But yeah, im done with the mod, got a 1gb 4870 (189.00 on newegg) instead of the headache. I figured the price of all the tools i needed to get the mod to work was too expensive for the gains that i got (almost $80 in tools/solder) so i returned all the tools. I see that you (natedog420) only overclocked your core to 700mhz (i understand why) to get your system stable, which converts to about 1.75 per 1mhz gain- i couldnt justify the expense for that kinda gain. Not talking down to you guys or anything, i love to see mods like this, but for me it just didnt work out. Best to luck to all you guys and your gpus.

[whe3ls]

Hey thanks.This is about as good as my camera gets.If you need macros of specific areas i can do that or i can borrow a camera for better pics of the whole pcb.

Help guys, my XFX 88GT has a different layout and only I need do to it the vgpu vmod, here some pics:









>Here the 500ohms VR



Thx in advance!

i think this is the mod. im wanting to mod my gt's. would like to know if this is the correct volt mod?

[natedog420]

lol, all of my soldiers are clean. But yeah, im done with the mod, got a 1gb 4870 (189.00 on newegg) instead of the headache. I figured the price of all the tools i needed to get the mod to work was too expensive for the gains that i got (almost $80 in tools/solder) so i returned all the tools. I see that you (natedog420) only overclocked your core to 700mhz (i understand why) to get your system stable, which converts to about 1.75 per 1mhz gain- i couldnt justify the expense for that kinda gain. Not talking down to you guys or anything, i love to see mods like this, but for me it just didnt work out. Best to luck to all you guys and your gpus.

in running low volts to keep it cool right now but i can run 830 core crysis stable with good temps. same with all other games i just choose to lower it some my sig isnt always right i im running 750 on low volts but cant crank it up and have no issues.

for benching i can do 900 core and over 2100 shader and have no issues but its not 100% stable and only good for benchmarks and all of this is on really crappy water

[whe3ls]

would a 10k trimmer be enough to volt mod a 8800gt? there all i can find locally

[natedog420]

would a 10k trimmer be enough to volt mod a 8800gt? there all i can find locally

i have a 15k on mine and its bad the volts just spike up at a certain point i would say at least a 20-50k more towards the 50k. you would have smaller increments in voltage so you can dial it it better.

[whe3ls]

its for a non reference card pictured above.

[whe3ls]

i got some gt's in the mail the other day. some older gt's. there were a couple bfg oc. both reference. so far one of them i ve gotten to. 756/1890/900 havent touch the memory. it has hynix ic's

[3dsag3]

Heres my V-moded EVGA 8800GT
870-2300(mem)-2160(shader)

[gugle]

very good.
which vgpu?

[celemine1Gig]

i have a 15k on mine and its bad the volts just spike up at a certain point i would say at least a 20-50k more towards the 50k. you would have smaller increments in voltage so you can dial it it better.

Just read that statement and it is plain WRONG! Don't follow this advice!!!

To correct that:
If the voltage shoots up quite badly in the lower resistance ranges with a 15k potentiometer it will even get worse with a potentiometer with higher resistance rating! Don't ever recommend things that you don't know anything about. You'll end up with people killing their hardware due to this wrong hint.

The problem is the following: Each and everey trimmer potentiometer, that I know of, gets highly non-linear resistance wise, in the lower resistance ranges. That means if you have a 15 turn 15k poti, you would expect every turn to be equal to 1k. But that is not correct. In the upper range you'll have more precision, i.e. less resistance change per turn, while in the lower range, the resistance change per turn does indeed increase.
That's why the voltage suddenly shoots up in lower resistance ranges: The resistance just drops faster --> higher voltage aand that even faster than before.

This behaviour either indicates choosing the wrong potentiometer value from the beginning (too high resistance for the desired end voltage), or just not enough precision.
The solution is either choosing a LOWER value potentiometer, which could result in quite a big default voltage bump, even when set to its max resistance.
The best solution however is to use 2 or more lower value trimmers/potentiometers, connected in series. This way you get a higher resulting resistance, but also with higher precision due to using more than one poti --> more turn per resistance change or to put it the other way round: less resistance change per turn --> higher precision.

[mike047]

Just read that statement and it is plain WRONG! Don't follow this advice!!!

To correct that:
If the voltage shoots up quite badly in the lower resistance ranges with a 15k potentiometer it will even get worse with a potentiometer with higher resistance rating! Don't ever recommend things that you don't know anything about. You'll end up with people killing their hardware due to this wrong hint.

The problem is the following: Each and everey trimmer potentiometer, that I know of, gets highly non-linear resistance wise, in the lower resistance ranges. That means if you have a 15 turn 15k poti, you would expect every turn to be equal to 1k. But that is not correct. In the upper range you'll have more precision, i.e. less resistance change per turn, while in the lower range, the resistance change per turn does indeed increase.
That's why the voltage suddenly shoots up in lower resistance ranges: The resistance just drops faster --> higher voltage aand that even faster than before.

This behaviour either indicates choosing the wrong potentiometer value from the beginning (too high resistance for the desired end voltage), or just not enough precision.
The solution is either choosing a LOWER value potentiometer, which could result in quite a big default voltage bump, even when set to its max resistance.
The best solution however is to use 2 or more lower value trimmers/potentiometers, connected in series. This way you get a higher resulting resistance, but also with higher precision due to using more than one poti --> more turn per resistance change or to put it the other way round: less resistance change per turn --> higher precision.

I have always used a 100 or 200 ohm trimmer in series with the larger size. You can really fine tune the voltage with the 200.

[celemine1Gig]

I have always used a 100 or 200 ohm trimmer in series with the larger size. You can really fine tune the voltage with the 200.

It all depends on the feedback resistance. You can't really say that in general 100-200 Ohms are optimal. For example if you are dealing with very low feedback resistances, i.e. the parallel resistance (the one for the vmod) has got to be quite big. In this situation the huge poti (say 500k-1M) will get highly non-linear in the k-Ohm range already, so your 200 Ohms trimmer won't help a thing, as the voltage would already be way too high. In this situation you'd better hope that the card has a funtioning OVP and OCP.

Essentially you are right, using a smaller poti, to be able to finetune the voltage, but you can't say to generally use value XY. It's not that easy.

[Noiz_TnT]

I attempted to do this mod but I somehow unsoldered the soldering point where the VR should be.
Anyway, my card doesn't boot anymore so is there any fix to get it going again ? Can i somehow replace the soldering point with something ?

[05_03_1992]

hi
do you have instructions for vgpu mod pcb galaxy 88gt?

[Renegade_!]

I have recently been given a Zotac 8800GT 512mb.
Model number ZT-88TES3P-FSP

one of the resistors on the 2nd up from bottom left memory chip, (lower left hand side, rear of the card, PCI-E pins pointing towards you)
has fallen off, it looks like the two resistors are in parrallel.

A diagram, the two first dashes being the two parralel resistors, the following dashes and slashes being larger white blocks, caps I think:

-_ _ _ \
-

EDIT:
Pictures:









can someone tell me the value, so I can add another resistor?

[largon]

^You could desolder equivalent resistor off from the back of another chip and measure it. Those are not too difficult nor risky to solder if you have a small enough tip on the iron.

[Renegade_!]

^You could desolder equivalent resistor off from the back of another chip and measure it. Those are not too difficult nor risky to solder if you have a small enough tip on the iron.

Would it be the same value as the resistor next to it, for the parrallel divider?
eg
22Kohm/22Kohm = 1ohm

I have no idea which resistor is equivalent as the layout on all of the memory chips is different.

[Renegade_!]

*Bump*

Anyone?