View Full Version : My 6800 ultra is voltage starved!
I'm currently running a modded XFX bios with 450/600 2D/3D clocks and the power save and performance 3D Vcores set to 1.5V and the 2D core at 1.1V. I've got the die water cooled and load temps are in the low 50sºC using rthdribl in a window.
However, I cannot get any decent stability over the 470-475 MHz range, as the core is dying for a litte bump. The Vr-zone mod doesn't seem to be correct and I've yet to see any sort of reasonable datasheet for the Volterra PWM controller to verify it. Anyone come up with any documentation on this part? Or maybe a way to control all of the Vid bits in the bios?
LardArse
07-28-2004, 06:32 AM
I've tried, couldn't get anything out from Volterra.
Hey, is that picture upside down cos the chip on mine face the other way.
Anyways, I might have gotten the pic wrong, you know the one where I pointed to is just beside the solder pads without resistor? It could the other resistor just beside the solder pads. Its definitely one of the two. Can you measure the resistance of that other one, the one connected to pin7?
Funny how come a few people said it works for them and they got some great scores out from their 6800s :confused:
Hey LardArse, I actually didn't try the mod yet as I wanted a datasheet before I started soldering just to verify it for myself. However, I measured a few of the resistances to gnd and got the following:
pin6 249 Ohms
pin7 8.06 KOhms
pin8 249 Ohms
pin9 6.84 KOhms
pin10 3.25 KOhms
So given your comment I think the line on your pic may be wrong as a 100K pot in parallel with 250 Ohms isn't going to do a whole hell of a lot :) Does that jog your memory any? Thanks for the reply.
Edit: Also I assume the Vid bits are up top? (pins 39-42) Maybe it would be easier to just do it digitally?
Edit2: Okay I'm a moron, I just reread your post on vr-zone and clearly see "Original resistance is around 8K ohms". So yes your arrow is pointing to the wrong side of the open 0402 footprint, and it is indeed pin7. My guess is you're bringing it into current limit as you crank up the voltage, so there is likely a work around for that as well.
LardArse
07-28-2004, 03:31 PM
Yeah, muist be pin7's resistor then.
Damn must be getting old! :D
Hey lard quick question. If you scale the voltage up with a resistor, are the bios increments sill 0.1V? I assume they would scale the same amount the resistor increases the Vcore if you're actually increasing the gain on the error amplifier.
LardArse
07-28-2004, 09:03 PM
Its still the same 0.1v
drcrawn
08-02-2004, 09:14 PM
Originally posted by 101
I'm currently running a modded XFX bios with 450/600 2D/3D clocks and the power save and performance 3D Vcores set to 1.5V and the 2D core at 1.1V. I've got the die water cooled and load temps are in the low 50sºC using rthdribl in a window.
However, I cannot get any decent stability over the 470-475 MHz range, as the core is dying for a litte bump. The Vr-zone mod doesn't seem to be correct and I've yet to see any sort of reasonable datasheet for the Volterra PWM controller to verify it. Anyone come up with any documentation on this part? Or maybe a way to control all of the Vid bits in the bios?
are you sure you got the extra .1v? the only reason I ask, is b/c 475 ish was my ceiling @1.43 when water cooled, when I flashed for 1.5v I can get 495 core now, clear difference. your temps r clearly good, so I cant see why you are topping out at 475. Good luck..
Maxym
07-23-2006, 11:08 PM
so you can link up the soft mod with the hard mod?
my 400/1100 6800U Gainward, @ 475/1260 using only a custom bios. Want to do the hardware mod with the volterra chip as well. Will reflash original bios for mod stage. But i might have to throw my bios back on cause i like the sound of 495 core's :)