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

[G4h4o8s6T]

EDIT: nevermind, vdiff mod is the shading of the resistor..I'll try to get it lower. as explained by Napalm, don't go lower than 200ohms, currently I'm at 340...I'll try to get it down to 300 and retest to see if it makes a difference. If I hit 1.27 then I know I'm on to something.

At 300ohm I got 1.28v load with Furmark, but other less intesive apps. around 1.25v. I wanted to try and get 1.28v when running something other than Furmark, so I brought it down to 200ohm.........no difference. Not sure why some of us are having issues with this, but its a little strange.

I have a similar question - anyone know what the max safe voltage is for the 285? I would assume its less than what the voltage tuner gave for the 280 (1.3v) since a 55nm core is more sensitive to voltage?

Is running 1.25 idle and 1.28 load 24/7 going to reduce the life of the card, or is this within limits? I may just go ahead and do the pencil mod to try and get 800 core.

Only time will tell, we are the guinea pigs

[Orangeblast]

well Now I'm around 290-300 ohms...still no difference.

Looks like I really need to revisit the bridging.

1/2/3/4/5/11 are bridged right now...

should I do 12 and get rid of 5? From what I was told in this thread, doing 5 or 12 will give the same results....it doesn't matter doing 5 or 12....any results that prove otherwise?

EDIT: Can someone post a picture of their bridging work? I'm also curious to see how the conductive ink was used...

EDIT: Redid my conductive ink...I got rid of 5 and did 12 instead....so now I'm at: 1/2/3/4/11/12 with my resistor at 290 omhs. I'll test when the ink dries.

EDIT: Okay, getting better results now...Got rid of bridge 5 and bridged 12....now using 1/2/3/4/11/12 and I'm getting 1.26 on load...about 1.263 to 1.268 The only thing I can think of is to better the bridge connections...i'm using conductive ink....i'll try an improve the connection with the ink, may later switch over to solder and a tiny lead. I still have yet to see 1.28

[NapalmV5]

ive stepped away from the thread to let you guys figure out on your own what your doing wrong/right

seems its not working

[NapalmV5]

Does the FTW have the same VID layout as SSC from the factory ?

Anyone know if the 285 FTW bios has higher voltage ?

yeh same vid

i dont have nibitor 4.9 to check on the bios voltages

but ive just measured my ftw: 1.26v @ bootup/1.29v @ load

ssc: 1.23v @ bootup/1.26v @ load

i will measure the components on the ssc vs ftw and hopefully ill find what i seek

[NapalmV5]

the FTW is just a better binned card...its then flashed with a BIOS that reflects the higher clocks of the FTW. The volt mod remains the same from the factory as the SSC cards.

if they bin their ssc/ftw then i must be binning my poop aswell

they just use random vanilla 285s for their overvolted ssc/ftw

[NapalmV5]

lets make it very clear why 1/2/3/4/11/12 are the ones to be bridged

1/2/3/4/11/12 are the back of D14/D15/D19 which are dioded

thus any other bridging/more than 6 bridges will not provide more voltage

[NapalmV5]

Does this give more vdimm to ?

yes.. its more of a vgpu+vmem mod

its a big plus.. when you overvolt/overclock youd want to increase vgpu+vmem accordingly

[Demo]

If i only need to increase the vGPU, w/c dones do i bridge only?

[NapalmV5]

Well, tonight I'm going to go ahead an get this a try...

I'll show some results and stuff. If you don't hear from me, something bad happened. lol


EDIT: Okay, made the mod to just 1 of my SSC cards. 1/2/3/4/5/11 are now bridged, however to my dismay, I'm only getting around 1.228 volts...which is a bit far from the expected 1.26

I hard modded it with solder and a tiny lead...

I noticed that some people did bridging on 1/2/3/4/11/12 ...is this what I should have done? Any when I solder it, should I run solder straight across to bridge it and get rid of my tiny lead?

Okay, well I was able to get the resistor down to 340 ohms...both of my SSCs.....and I went ahead and got rid of the leads and used conductive ink..so hopefully I get the full Voltage and hopefully get a boost.

I haven't tested this new setup yet, but from last night during my first test I did see a slight boost from 735/1656/2800 to 745/1656/2850...nice boost, temps were still good as well.

Once I get my Koolance water blocks, I should be seeing some great gains.

Yup, this is EXACTLY what I'm experiencing as well. I've got the resistor down to 340 ohms......I've done the bridging on 4/5 along with the already 1/2/3/11 bridges.

I'm only getting 1.26 volts during furmark testing....

I can also confirm that I too am getting different voltages depending on what application I'm running....

During Vantage testing, I get around 1.24....during Furmark testing I get 1.26....
Wonder why its not consistent.

Any suggestions for us guy running in to the anomalies?

Yea, its strange, almost seems like we missed a step or a step wasn't mentioned. I dunno......

I just saw this post....I'm only getting a max of 1.26V....and you mention that "without vdiff the max allowed is 1.26v"

so......what is the vdiff...I must be missing something here...

EDIT: nevermind, vdiff mod is the shading of the resistor..I'll try to get it lower. as explained by Napalm, don't go lower than 200ohms, currently I'm at 340...I'll try to get it down to 300 and retest to see if it makes a difference. If I hit 1.27 then I know I'm on to something.

well Now I'm around 290-300 ohms...still no difference.

Looks like I really need to revisit the bridging.

1/2/3/4/5/11 are bridged right now...

should I do 12 and get rid of 5? From what I was told in this thread, doing 5 or 12 will give the same results....it doesn't matter doing 5 or 12....any results that prove otherwise?

EDIT: Can someone post a picture of their bridging work? I'm also curious to see how the conductive ink was used...

EDIT: Redid my conductive ink...I got rid of 5 and did 12 instead....so now I'm at: 1/2/3/4/11/12 with my resistor at 290 omhs. I'll test when the ink dries.

EDIT: Okay, getting better results now...Got rid of bridge 5 and bridged 12....now using 1/2/3/4/11/12 and I'm getting 1.26 on load...about 1.263 to 1.268 The only thing I can think of is to better the bridge connections...i'm using conductive ink....i'll try an improve the connection with the ink, may later switch over to solder and a tiny lead. I still have yet to see 1.28

theres no other step just step 1 -> step 2

if your not getting higher voltage from lower vdiff resistance.. you have to redo your penciling

http://www.xtremesystems.org/Forums/...&postcount=134

pencil modding is unstable.. even if you get lower resistance doesnt mean you get higher voltage

[NapalmV5]

If i only need to increase the vGPU, w/c dones do i bridge only?

when you overvolt/overclock youd want to increase vgpu+vmem accordingly

ssc/ftw are overvolted the same way.. via bridging which overvolts both gpu/mem slightly

[Orangeblast]

if they bin their ssc/ftw then i must be binning my poop aswell

they just use random vanilla 285s for their overvolted ssc/ftw

haha...well I do know this....the FTWs come from SSC cards. The FTWs are considered special and are not high on inventory, they are a short run, limited quantity. The FTWs are going to be better overclockers, there's no doubt about that. The FTWs are the best of the SSC cards. I know I talk about this stuff, but I can't really get in to how I know this.

Anyways, back to the modding, I cleaned off my resistor and I'm starting over again...I don't think that my bridging is doing too well, I cleaned off the resistor and retested and measured voltage...cleaning it off and running bridges only got me to 1.245 on load during furmark testing. As you have stated, I should be seeing 1.26 on load, my bridges probably aren't solid enough. Maybe I should just ditch the ink now and go solder. I want those 1.30 volts! My blocks are coming in soon...

[NapalmV5]

Okay I am having the same issues as some of you guys in that I have gone through all the steps and am getting only 1.26v under load. I have a XFX Black Edition and started out by using conductive ink for step 1, 2b pensil for step 2 and a 500ohm VR tuned down to 150ohm for step 3. After reading all the posts I removed the conductive ink for step 1 end replaced 2, 4 and 12 with tinned wire (thin wire). Another issue is that the SMR from step 2 starts out at 1.07kohms, that seems much higher then what it should be. Given that however I was able to pencil it down to 330ohm with my 2b. And i'm still only at 1.26v under load! What am I doing wrong? Please help!!

This question is for NapalmV5. I have followed your modding adventures with the GTX 285 in various forums and you have done many mods to your card (s?). My question is, have you tried your 3 step vmod method on a fresh card (one that you have done no other mods on)? The reason I ask is perhaps you have unlocked something along the way to allow you to hit 1.36v that is in tandem with the 3 steps you have shown all of us. Please don't take that the wrong way as I appreciate greatly the work you have done. Thank you.

theres no other step

this is all XS xclusive.. luxx forum is the only other place ive shared with

you have to pencil/repencil till you get a working resistance.. just as with unstable silver ink @ step 1

not just lower resistance.. resistance that increases your vgpu

[NapalmV5]

How much is tha max safe temp for 24/7 stock cooler use if i perform the ink mod that gives 1.26(core) and 2.15(ram) ?

Aside from Warhead, Furmark , 3dmark05/06 what other stability testing tools/methods should i consider?

I have a similar question - anyone know what the max safe voltage is for the 285? I would assume its less than what the voltage tuner gave for the 280 (1.3v) since a 55nm core is more sensitive to voltage?

Is running 1.25 idle and 1.28 load 24/7 going to reduce the life of the card, or is this within limits? I may just go ahead and do the pencil mod to try and get 800 core.

you guys are fine.. if you do worry than why would you even want to oc ?

[NapalmV5]

At 300ohm I got 1.28v load with Furmark, but other less intesive apps. around 1.25v. I wanted to try and get 1.28v when running something other than Furmark, so I brought it down to 200ohm.........no difference. Not sure why some of us are having issues with this, but its a little strange.
Only time will tell, we are the guinea pigs

http://www.xtremesystems.org/Forums/...&postcount=134

vdiff mod is not working

you gotta pencil/repencil till you get an increase in vpgu

ive guinea piged vdiff thoroughly

btw nice clocks

[NapalmV5]

if you guys want accurate/actual results from these mods please dont use silver ink

i dont and i believe neither cryptik recommends it.. and do one step at a time that way you can move forward to the next step confidently

step 1 ^

if you guys want accurate/actual results from step 2: remove/replace the 770 ohm vdiff resistor with a 300 ohm resistor

ive done just that on the amp 285 and i dont have to pencil/repencil to get working resistance

[Orangeblast]

Retesting....repenciled to 250 ohm...just to see if I'll get any difference....boot up is around 1.23v, idle is: 1.07v load: 1.27v (after 1 min of furmark testing)

hrm...I suppose its my bridging then..move to solder tonight.

[Demo]

@NapalmV5

Thank you for the information =)

[KingK76]

step 1 ^

if you guys want accurate/actual results from step 2: remove/replace the 770 ohm vdiff resistor with a 300 ohm resistor

ive done just that on the amp 285 and i dont have to pencil/repencil to get working resistance

That is the resistor that on my XFX Black Edition reads 1.07Kohms not 770ohms. Does anyone else have an XFX card to test this on? Perhaps this is why i'm not able to achieve the voltage I want. It's very frustrating as this is not my first time doing these type of mods and even bridging 2,4 and 12 it's only giving me 2.10v on the memory! I'm going kinda nuts here! With that being said I was still able to pensil that resistor down to 330ohm but I made no difference.

[Demo]

I just got my GTX285 now, gonna post my voltages later =)

[Orangeblast]

Soldered leads...still getting a max of 1.27v on load and resistor is at 250ohms....no change.

I have no idea what else to do. I guess it has something to do with the PCB build.

[Demo]

Can anyone tell me w/c map on the crysis warhead benchmark is the most demanding for the GPU stability ?


EDIT:

I noticed the card makes a very audible buzzing when in operation, my PSU is a silverstone Zues 750ZF and the card is connected on the 12v4 rail.

EDIT2:

at 30c ambient my card runs at 87c at furmark.. a little hotter than i expected.. fan speed is at auto

EDIT3:

vGPU seems to be at 1.12-1.13 during load and idle,and vMEM at 2.09 - 2.10 at both idle and load, seems a little bit low for stock as it was mentioned that the stock is 1.15 ~ 1.16

my DMM reads 12.01-12.02 on my 12v line.

[nickless]

That is the resistor that on my XFX Black Edition reads 1.07Kohms not 770ohms. Does anyone else have an XFX card to test this on? Perhaps this is why i'm not able to achieve the voltage I want. It's very frustrating as this is not my first time doing these type of mods and even bridging 2,4 and 12 it's only giving me 2.10v on the memory! I'm going kinda nuts here! With that being said I was still able to pensil that resistor down to 330ohm but I made no difference.

I'm also with XFX Black Edition (with FTW bios) and my DMM show ~750ohm vdiff resistor before pencil mod.

[NapalmV5]

That is the resistor that on my XFX Black Edition reads 1.07Kohms not 770ohms. Does anyone else have an XFX card to test this on? Perhaps this is why i'm not able to achieve the voltage I want. It's very frustrating as this is not my first time doing these type of mods and even bridging 2,4 and 12 it's only giving me 2.10v on the memory! I'm going kinda nuts here! With that being said I was still able to pensil that resistor down to 330ohm but I made no difference.

Soldered leads...still getting a max of 1.27v on load and resistor is at 250ohms....no change.

I have no idea what else to do. I guess it has something to do with the PCB build.

you guys are making it the hardest mod ever.. have guys tried diff #2 pencils ?

ive tested/retested/spent a lot of time on the vdiff mod @ ssc and amp.. why does it work for me ?

when penciling the resistor make sure you dont get any pencil on the other components.. it will lower your vgpu not increase it

also when you guys measure the resistance after penciling make sure you dont remove the resistance with your dmm when measuring

[Demo]

That penciling step is for allowing higher vGPU right, dont you guys get temps like 90c over 1.26v loaded? as im getting 87c loaded with furmark and those are at stock volts.

[NapalmV5]

^ yes even @ 2D temp rises

i just bridged 4 and 12 on my ftw for a total of 6 bridges 1/2/3/4/11/12

bootup/load increased by .01

then #2 penciled the vdiff resistor down to 300 ohm and bam

bootup @ 1.29v and load @ 1.34v (more like ~1345mV) @ load (first pencil mod attemp)

come on guys this mod doesnt only work on my 285s