GTX285 various mods.. which may work for GTX260/280/295 also

[Hondacity]

NapalmV5....there are many here who appreciate your efforts, please continue.

+1 you're work is uber good...i got one card coming...

[NapalmV5]

Did you compare the memory timing table between the two BIOSes? They may have loosened the timings to allow a better memory overclock in the FTW BIOS and as result running the same memory speed as original BIOS you may have lower bandwidth/higher latency.

dont have nibitor 4.9 which supports 285

i wasnt sure its the ftw bios untill flash

its probably just the drivers.. didnt have time to check 181.22/182.05 on the ftw bios.. maybe ill do it later this morning

[NapalmV5]

I had taken them all off because I had no clock gains so I am holding off til I get better cooling. I saw no need in putting extra voltage to it if its running fine now.

I probably did not clean the resistor enough because even with everything removed I am getting a reading that is .02 higher than when I started.

without the vid your not gonna get any higher clocks on stock cooling

[NapalmV5]

Ok, I've found 6B pencil, also 21/2=HB. I'll try both of them, I hope one of these pencils will help this time.

EDIT: Between I've flashed my card with FTW BIOS, but nothing changed with FW 182.05.

EDIT2: I've managed to reduce the resistance with 21/2=HB pencil to 330ohm, but it gives me only 1.24v (before only with 1/2/3/4/11/12 it was 1.21v). I'll clear everything and make the hall mod again, maybe some bridges are not connected very well

1.24v @ load ?

you should be @ 1.3v load

[NapalmV5]

man, i give you credit for being able to do this. i wish i could recap a GPU - my hands shake tooo much and my eye sight is VERY bad.

Please link caps you used and *begs like child* please post pics of what you re-capped.

i have a trusted person in mind for re-capping.

lol i just deleted the pix of the recaps ive been doing

later today i should get the 1000uf 2.5v 4mohm esr kemet caps.. cant wait to getem on

but as for posting pic of that.. idk.. some will attempt (someone already has) the recap using different caps available to them and may get worse results than default clocks like i did at first..

the pic is not gonna look any diff than what i had up before.. the caps are exactly the same just diff values

ill think about it

[NapalmV5]

What would you recommend on leads, I managed to find 6B at office depot. Seems the HB are very common, all i see everywhere else..

I am just awaiting the HR 03 GTX with xtra VRM heatsinks to get in stock, Thermalright shipped them out to resellers this week... But I figure when I have it apart I want to shade it as well as possible and see if I can get any clock gains with cooler temps...

no.2

one no.2 got it down to 320ohm max the other got it down to 100ohm

[NapalmV5]

I also managed to reduce the resistance with some 21/2=HB pencil to 290 ohm. I've cleared everything and made the hall mod again, but it gives me only 1.24v again. It looks like 1.33v is a dream for may card without hard vmod

1.24v load ?

[NapalmV5]

Sorry to offend. From here on, I won't post in this thread and I'll keep all mod and cooling results to myself. If everyone does the same, maybe we'll all learn nothing from each other. So much for constructive criticsm.

well sorry to offend you.. what have i been doing all this time..

how about before constructive criticism you actually proceed with the mods correctly/surely

[NapalmV5]

anyone not getting 1.30v @ load from doing vid bridging and vdiff mod (step #1 and #2 @ guide) suggests a redo of the vmods until 1.3 is achieved

as for clock gains i cant guarantee anything there

my ssc was affected more from capping C561-4 than .03v increment

the amp on the other hand was affected more from higher vcore/than the caps further helped

[CryptiK]

Napalm no need to thank me mate, you have been doing all the work I have just put my 2c in here and there. I and I'm sure a few others do appreciate your work and time. Thank you.

I did see an increase in vgpu and vmem after doing the mods, and as a result in clocks, as I reported earlier in the thread. I measured at the measure points I showed earlier in the thread and also at the caps behind the gpu socket, and there is definitely a gain.

My card wouldn't run crysis at 1620 shader at stock voltage and auto fan, 1512 was max, but now with 1.24 vgpu it does 1620 no problem with temps hitting 86*C on a hot day with the fan on auto never going over 72% duty cycle.

Before the mods I could not run 756 core on auto fan in crysis for 30 seconds, now it lasts 5 mins before locking, its almost stable, with another 0.02v I think it would be totally stable. If I crank the fan up to 100%, I can run 756c/1656s/1512m in crysis no problems at all.

Guys as Napalm said, do not attempt things that are out of your ability level. Re-capping/adding caps is relatively high level as far as modding goes, and takes some good soldering skills, which Napalm and myself have. Also the caps characteristics need to be selected carefully, adding caps/coils & other components can have deleterious affects due to resonance and other factors.

[NapalmV5]

Napalm no need to thank me mate, you have been doing all the work I have just put my 2c in here and there. I and I'm sure a few others do appreciate your work and time. Thank you.

I did see an increase in vgpu and vmem after doing the mods, and as a result in clocks, as I reported earlier in the thread. I measured at the measure points I showed earlier in the thread and also at the caps behind the gpu socket, and there is definitely a gain.

My card wouldn't run crysis at 1620 shader at stock voltage and auto fan, 1512 was max, but now with 1.24 vgpu it does 1620 no problem with temps hitting 86*C on a hot day with the fan on auto never going over 72% duty cycle.

Before the mods I could not run 756 core on auto fan in crysis for 30 seconds, now it lasts 5 mins before locking, its almost stable, with another 0.02v I think it would be totally stable. If I crank the fan up to 100%, I can run 756c/1656s/1512m in crysis no problems at all.

Guys as Napalm said, do not attempt things that are out of your ability level. Re-capping/adding caps is relatively high level as far as modding goes, and takes some good soldering skills, which Napalm and myself have. Also the caps characteristics need to be selected carefully, adding caps/coils & other components can have deleterious affects due to resonance and other factors.

youve helped me greatly and i appreciate it

if you got any idea youd like to try out.. im here ill try it on my 285s.. i got amp and ssc to sacrifice for whatever mods



edit: (d18 d15 d14) d18 increased 2d voltage to 1.22v (330 ohm vdiff resistor) but 3d voltage remained the same 1.33v

3d shouldve went up .10-.14v (that wud be ~1.47vcore) as well.. what u think its holding it back ?

[NapalmV5]

updated guide to 1.39375vgpu max

before attempting any vcore mod please measure your current vcore

courtesy: CryptiK

red:vgpu/green:vmem


step #1
- for better understanding please check dengyongs post #80 cryptiks post #112 and #118
- simple way to increase vcore from 1.15v reference

1.26vcore @ load

step #2
- the lower the resistance at this resistor the higher the vcore allowed
- pencil this resistor to bring resistance down to around 350ohm

1.29-1.33vcore @ load

step #3
- this step is more for the water/more xtreme cooling
- this point increases vcore according to the resistor in above pic ^
- solder a 150ohm resistor to ground from this red point

1.36vcore @ load

2.20vmem @ load

so far ive gotten 1.38vcore

unfortuanately no matter how low the resistance gets @ step #2

all is limited to the reference voltage: 1.39375

past this voltage further mods are needed


thanks goes to Cryptik for all the help/confirmations

dengyong and everyone else contributing @ thread

[10acjed]

no.2

one no.2 got it down to 320ohm max the other got it down to 100ohm

Ok, I had read a few posts back that said the no.2 HB arent good for this and out of the 3 diff stores I stopped at, they all carried a vast array of HB no.2 pencils... I found a 6B no.2 lead refill pack that i picked up so I will probably use that....

[NapalmV5]

the no.2 HB i got barely did anything

dont go below 200ohm.. if you can bring it down to 300ohm keep it there

and with the vid bridging you should get ~1.3 @ load.. while 3D runs measure vgpu

[tranceaddict]

I'm about to get a vanilla XFX GTX 285 and want to use "step 1" to get 1.26vcore @ load. The PCB layout should be the same on XFX as on the picture above, can anyone confirm that please?

[G4h4o8s6T]

can anyone confirm that please?

Newegg has some good highres. pics of most of their cards, always a good place to check

http://www.newegg.com/Product/ShowIm...d%20-%20Retail

1,3,11 look bridged the same as other vanilla cards.

[tranceaddict]

Just one more question; in the above pic (2),(4),(12) are bridged in addition to the already vanilla bridged ones, but in post #118 it says to bridge (2), (4) & (5) in addition to the alreaddy bridged ones. I assume that the (2),(4),(12) + (already bridged 1,3,11) is the correct method?

[10acjed]

Just one more question; in the above pic (2),(4),(12) are bridged in addition to the already vanilla bridged ones, but in post #118 it says to bridge (2), (4) & (5) in addition to the alreaddy bridged ones. I assume that the (2),(4),(12) + (already bridged 1,3,11) is the correct method?

that is correct. I dont think bridging 5 did anything. But I did 2/4/12 and got the desired voltage...

[joshua7]

Do you think, that a fully stock GTX285 can do 850Mhz on the gpu? Without mods, aftermarket cooling, anything? Read a test on a site and the writer clamis it is true.

[LexDiamonds]

What are people using to connect the bridges? Can it be done with a small blob of solder?

[SAE]

Has to be a VERY tiny blob of solder then. I just looked at my card to confirm that. I am gonna try to connect them using silver ink and a hair of a brush to apply it.

[10acjed]

Has to be a VERY tiny blob of solder then. I just looked at my card to confirm that. I am gonna try to connect them using silver ink and a hair of a brush to apply it.

I keep the alcohol and q-tip handy... i think i removed one of them 3 different times before i got it just right

if it can be done with ink rather than a solder, i always recommend the ink as you can remove it.... just my logic

[G4h4o8s6T]

So I did the mods today, bridged 2,4,12(in addition to 1,3,11) and penciled the resistor down to 300ohm, but dont think Im getting the results I should.

Core- 1.23v Load

Mem- 2.11 Load

The bridges look good, and measured the resistance 5x, so I know its not that. But I kinded jumped the gun and did them all at once instead of trying each at a time. Do you guys recommend starting over from scratch? Or any other ideas?

[10acjed]

So I did the mods today, bridged 2,4,12(in addition to 1,3,11) and penciled the resistor down to 300ohm, but dont think Im getting the results I should.

Core- 1.23v Load

Mem- 2.11 Load

The bridges look good, and measured the resistance 5x, so I know its not that. But I kinded jumped the gun and did them all at once instead of trying each at a time. Do you guys recommend starting over from scratch? Or any other ideas?

something sounds wrong. I did all 3 resistors at once and got results. you should be seeing at least clocser to 1.28 - 1.29 with all that..

I would double check all the resistors. Are you using conductive ink?

[nickless]

So I did the mods today, bridged 2,4,12(in addition to 1,3,11) and penciled the resistor down to 300ohm, but dont think Im getting the results I should.

Core- 1.23v Load

Mem- 2.11 Load

The bridges look good, and measured the resistance 5x, so I know its not that. But I kinded jumped the gun and did them all at once instead of trying each at a time. Do you guys recommend starting over from scratch? Or any other ideas?

Same here, I did the mods three times already, everything looks good, penciled the resistor down to 304ohm, but my vcore and vmem are the same like yours.