Radeon X800 Non-Pro/X800 XL/X850 Voltmods

[spyskank]

Dead x850xt Pe, Yes.... Another one, 0V vcore

ok ok...

Event: PENCIL VCore & IGPU Vmods as described

Question: ON VGPU , What is the reading your suppose to get between the two soldering
pad before applyin pencil When you set you DMM to 20k ?? It isnt clear

about ??? 420 ohms

R1596 gave me correctly about 40k whiyh DMM at 200k

No change in vcore after mod , still 1.4.

So, moved to VID Vmods Set Vcore to 1.6(WAter) 01000 with conductive pen on Vid2

Noticed that Smr about 1k, am i right ?

IF so, shouldnt we solder an 1k REs between the pad instead ?

Vcore was indeed 1.6; GOOd

Then decided to Solder variable Res instead of graphite to be able to change Res Ezly in the
Future.

Cleanned first job, Pen graphyte; Vcore back to 1.4. all Res back to stock


SOLDERED VGPU with 500k Vr, dropped Res so total was about 20k after

SOLDERED IGPU with 50k Vr set to 10k,
What is suppose to be the total Res Reading by then, Vgpu about 20k, but this Igpu ??
;but since had follow the instructions,,, went on

Vcore was about 1.5, boot was ok, test ok, good, but wantee 1.6,

Removed cards , lowered the res a bit on VGPU SET it 8k, card was offboard.

wasnt to much...


AND this is how i uderstand it, lower res higher voltage;

system wouldnt boot;

tought it might be IGPU safety short so lowered from 20k to 15 k since its just to increase limits.

restarted, system woudlnt boot, VCORE IS 0 VOLTS..

cleared all the mods... no boot sigh...

Is gpu dead ?? or when desolderin did i messup res so that it when so hi that all the vcore is cut ??

tried to resolder VGPU VR (board trace Is ok) to get the 420 ohm reading from the start.. no go...


x300 workin good,


ny help would b apprciated

 

[spyskank]

More Info;

VDD & Vddq reading are still good 2.05; 2.10
Fan still spinning,

dfi nf4 stuck at 3led on top Pciex, No Video

Goes to 1 led on bottom pciex, but no go ..

 

[Mussels]

Spyskank: Its possible you toasted the card. if you can get IGPU and VGPU to their orignal ohm readings, and still no go, you got a toasted card.

 

[funkflix]

?

 

[Unregistered]

hmm.... this looks promising but I have a few questions. I am using a zalman vf700-cu, which seems to be one of the best air coolers available. At 540/590 at default voltage on my x850 pro (flashed to 16 pipes and xt PE bios), my load temp is around 50 degrees. My idle temp is around 30 degrees. My default voltage is 1.4v and I don't have a multimeter. ( it also sounds annoying to do it the pencil way). I want to mod the VID to 1.6v. From what I've seen on this forum and on google, its kind of dangerous to put the voltage that high. But someone on this forum (I think it was Viper John) said that 1.55 volts is nothing on air as long as it's good air. Can my zalman handle the 1.6 volts? Another quick question. I hear of people easily getting 600+mhz on the core with stock voltage. I can not even get to 570 before it freezes. I keep it at 540 for maximum stability. What can cause such a bad overclock?

 

[HousERaT]

1.6v with a zalman only if you want to fry your card. I suppose if you're really serious about this you can try it but generally speaking 1.45 is about as you want to go with just a Zalman.

 

[Unregistered]

Hi, it's the person with the zalman again.

I modded my card from the following stock voltages:

vcore 1.450V
vdd 2.12V
vddq 2.16V

to the following new voltages:

vgpu 1.411V
vddq 2.09V
vdd 2.09V

I measured these voltages before and after with my meter. All three were done via pencil mod. but i can't get the vore much past 545. I know for a fact its unstable at 550 and if I set it to 560 and click scan for artifacts in ati tray tools, the test image will freeze, and then 5 seconds later the screen will go black and I will have to do a hard shut down to reboot. Interestingly, when the test image freezes, I still have a working mouse pointer. It's just that nothing on the computer will respond. It behaved exactly like this before I voltmodded my card. So far I have gained nothing from the mod except my card runs about 5 degrees hotter. I still keep my clocks at 540/587 to ensure no crashing. This is a BBA x850 pro flashed to an xt PE for 16 pipes. Anyone have any idea what the problem could be?

 

[Unregistered]

oops i posted those voltages in the wrong order


These are new voltages

vcore 1.450V
vdd 2.12V
vddq 2.16V

These are stock:

vgpu 1.411V
vddq 2.09V
vdd 2.09V

 

[Andre_Santarell]

hi, if i want do only vmod for 1.45 v wich paramters do i use ?
thanks

than can i change in a permanent mode the fan speed of stock cooler ? for example for set at 100% with no ati tool o riva turner ?
thanks

 

[louster35]

Solved The Pencil Vcore Problem

**SOLVED THE PENCIL VCORE PROBLEM**

Hey guys just wanted to share my solution for the pencilmod vcore problem

the problem isnt that the resistance is too low and the graphite wont change it. its that the graphite wont stick to the pcb to make a dense enough layer. anyways:

after even covering the gap in dust graphite nothing was happenning. cleaning it over and over with alcohol actually raised the resistance in my experiment.

so what i did was take a piece of normal tape and press it against the area where we want to write on with the pencil. this in turn got the pcb and the mountain points sticky enough to write on as if it were paper... i could lower resistance as much as i wanted.

on my sapphire X800GTO2@X850XT PE flashed bios and unlocked i gained 20 mhz with a 0,4 volt mod and much greater stability.


from Greece, lots of love, Over and out.

 

[SteveH6479]

No answer?

Hi, I'm the unregistered guy that posted about getting really bad clocks about 2 posts ago. Did I just get bad luck on my GPU? Or am I missing something? I even modded my vcore to 1.47V and upped my Vdd a little more. I got to 560 core, which creates artifacts and is unstable. 555 creates less artifacts but is more stable so I keep it at 555. But EVERY single post I've seen about x850s has been much higher overclocks even with the stock cooler and no vmod. Lowest max oc I have seen on stock cooler and no vmod is 560-570 mhz. Most I've seen go over 600. I am really dissappointed . My max oc on stock cooler was about 536. At 540 it was too unstable. I thought my vf700-cu would fix it but it only allowed my core to go to about 540 from 536. My core temp maxes at 57 Celsius at 1.47V. I guess the best I can do is up the voltage as much as I can. So, how much hotter can I allow my card to get when vmodding it?

 

[ViperJohn]

Hi, I'm the unregistered guy that posted about getting really bad clocks about 2 posts ago. Did I just get bad luck on my GPU? Or am I missing something? I even modded my vcore to 1.47V and upped my Vdd a little more. I got to 560 core, which creates artifacts and is unstable. 555 creates less artifacts but is more stable so I keep it at 555. But EVERY single post I've seen about x850s has been much higher overclocks even with the stock cooler and no vmod. Lowest max oc I have seen on stock cooler and no vmod is 560-570 mhz. Most I've seen go over 600. I am really dissappointed . My max oc on stock cooler was about 536. At 540 it was too unstable. I thought my vf700-cu would fix it but it only allowed my core to go to about 540 from 536. My core temp maxes at 57 Celsius at 1.47V. I guess the best I can do is up the voltage as much as I can. So, how much hotter can I allow my card to get when vmodding it?

Sounds like you just got on the bad side in the luck of the draw. It happens.
The vf700-cu is a quieter replacement for the stock x850 cooler but it is not
enought GPU cooler for volt modding by any stretch of the imagination.

Viper

 

[SteveH6479]

oh well

Thats too bad. But at least you ended my search for something that isn't there. So, thank you viper. I'm just curious about one more thing. I know that you can't go very far with a vf-700cu but I'd like to get the most out of my unlucky card. So, how far would you allow the temperature of an x850 on 1.5 volts to go? A small suggestion for that guide would be to add a table of recommended max temperatures at certain voltages. So I could look up things like the recommended max temp for 1.5V.

 

[ViperJohn]

Thats too bad. But at least you ended my search for something that isn't there. So, thank you viper. I'm just curious about one more thing. I know that you can't go very far with a vf-700cu but I'd like to get the most out of my unlucky card. So, how far would you allow the temperature of an x850 on 1.5 volts to go? A small suggestion for that guide would be to add a table of recommended max temperatures at certain voltages. So I could look up things like the recommended max temp for 1.5V.

Well do not feel like the Lone Ranger. While i've had the pleasure of working with many
cards having really superior cores i've had my fair share of dogs with three broken legs
tool. You just can't get around having some luck of the draw.

You could probably run 1.50 loaded Vcore with VF-700cu if you can keep the loaded die
temps below about 57-60C. Typically you go above that temperature, and it does
vary from core to core, the higher OC that more Vcore would normally bring is just taken
back from the higher die temperatures the higher Vcore produces. It's classic Catch 22
in action lol.

Viper

 

[SteveH6479]

Yea I kinda figured it was somehing like that. But want I want to know is if its safe to try it. I don't want to fry my card. It's pretty stable at 57 degrees at 1.47. At 1.4 it maxes at 55. So it's still worth it to be at 1.47V. but I'm already .02V over the recommended max. As I up the clocks when its at 1.5V, can I fry my card? Or will it just get unstable? When does voltage and heat become a bigger factor than just monitoring the card's stability?

 

[ViperJohn]

Yea I kinda figured it was somehing like that. But want I want to know is if its safe to try it. I don't want to fry my card. It's pretty stable at 57 degrees at 1.47. At 1.4 it maxes at 55. So it's still worth it to be at 1.47V. but I'm already .02V over the recommended max. As I up the clocks when its at 1.5V, can I fry my card? Or will it just get unstable? When does voltage and heat become a bigger factor than just monitoring the card's stability?

Actually if you really want to know how hot your core will run use the 3DM01se Nature test only set to repeat (not loop) 25 times with the title screens disabled. If you lockup then lower the core clock as most R420 cores will lock before they start to artifact. If you use RivaTuner's background temperature graphing you will have a nice graph to refer to when the 3DM01se Nature "torture loop" is exited.

On air cooling you want to stay lower than around 56-58C or less loaded in the Nature torture loop as that is about the point where increasing the Vcore produces no gain in core OC due to increased heat. You will not fry the card running say 60C at 1.50 Vcore but you may and probably will be above the point where running 1.50 Vcore at that die temp is doing anything for your core OC you so what is the point.

The 1.450 max Vcore on air cooling is pure fiction. It all depends on what air cooling you are using and how warm the die runs. 1.620 to 1.650 loaded Vcore is regularly run 24/7 with my Viper Magnum (Turbo Magnum) air coolers producing true loaded die temps in the 53-56C range with core clocks up to 690Mhz. The heat dissapated by a given GPU (CPU's too) at a given Vcore and clock speed can vary quite a bit core to core and is one of the factors that can cause a core to be binned for use on a slower clock speed card, like an XT or Pro Vivo, even though it actually speed bins for XT-PE use.

Two things effect core power draw in watts and the resultant heat dissapation with those being operating voltage and frequency. The power the core is drawing goes up at the square of the ratio of the voltage change times the ratio of the frequency change so things can happen very quickly...especially on the voltage side of the equation.

Now with that in mind the better the air cooling you use the more Vcore you can run and the higher you can push the core clock without pushing the die temp to high. As you push ever higher on the core clock the tendency is for the die temp where the point of no OC gain with a Vcore increase begins to occur, happens at ever lower die temperatures. You then have to switch to more aggressive cooling means, like water, to lower the die temps so the increased Vcore will be effective at producing higher core clocks.

Viper

 

[SteveH6479]

Hmm, ok then in that case it it seems relatively safe to experiment. I know the stock cooler on my card let it run over 70 C easily at stock clocks and no unlocked pipelines. So theres no way its gonna fry. It just won't work properly. Maybe I'll get lucky and have a GPU that runs at higher than average temperatures. I'll post some results when I get the chance to give the mod another try.

 

[SteveH6479]

not too bad

I'm still dissapointed but I managed to get to 570 with 1.514 volts on the core. My card maxes at 61 degrees C. It's better than nothing. Thanks viper.

 

[Althalus]

lol

lol, if you think 570 is bad you need you head checked!

I'm only getting 500 on core before the system hangs, my memory goes to 600.

Check this thread http://forums.overclockers.co.uk/showthread.php?p=5902548 there are quite a few people topping out at 500.

It is very annoying. 505 and the system locks in anything 3d. but I can set it to 550 in 2d and it is fine! any ideas?

 

[SteveH6479]

but mines an r480

What kind of card do you have. If it is an x800gto or gto2 i believe it is randomized whether or not you get an r430 or r480 core. I have an x850 pro which is definately an r480 core. You might have an r430 core which is an x800, not an x850. x800's tend to crap out quicker on the core. Also, what voltages are you running at?

 

[Althalus]

It is a gto2 running at stock volts. I thought all the gto2 cores were the r480?

 

[Mussels]

all GTO2 are, the sapphire ones. but normal GT/GTO can vary core types.

 

[SteveH6479]

oh, really? So GTO2's are all R480's? Hmm, I kinda feel bad for ya then. That clock sucks .

 

[Mussels]

that was sapphires claim, that the GTO2 had a better GPU on it than the normal GTOs.

bad clocks are bad clocks i had a X800PRO ViVo that did XTPE speeds, and an RMA turned up a PCI-E X800XL that gets a whole 15Mhz before it BSOD's on me. you win some you lose some.

 

[Super XP]

Sapphire Radeon X800GTO2

Ya, the Sapphire Radeon X800GTO2 all have the R480 Core. And when you flash the bios to gain the extra 4 pipes for a total of 16, then you can OC to @ least X850XT PE speeds. And even take it further to over 600 MHz on the core with water cooling.

They beat the new X1800 XL's by far in terms of performance.

I ordered a GTO2, should get it in 2 days or so