3.3V on the GPU doen't work ???

[YoupY]

I've just tryed putting the 3.3V direct on the gpu.

In stead of soldering the 3.3V and the cap to the tab I put it on the middle pin (Vout, according to the datasheet), cause the wire wouldn't stick to the tab, but since there both the same voltage I thought it shouldn't matter, but I'm not sure....

I cut all the legs of the mosfets, on both sides of the card, cause both sides are connected. Maybe I shouldn't have done this, and should I had to cut only the legs of the mosfets the mod is on.

I used a 2200 uF cap, and put the negative side on the middle pin. I'm also not quite sure if it should have reversed this.

Alltogether I can't say if the mod works since I'm not sure if I've done everthing right. I'm hoping someone can give me some answers on this.

If the conclusion is that all these things are right, then I'm sorry to say the mod doesn't work, but I'm thinking I did something wrong, so I could use some help on this one.

[Holst]

Take the cap off.....

It should work the way you have done it.. providing that 3.3v isnt just too much for the Xabre GPU..

Im stuck trying to get much more than 2.35v out of my VGPU... I want 2.5v+

I think its down to the way the Vreg works... and my 3.3v line isnt very high..

Ive given up on my further Vmem mod for the moment as my lab PSU wont run a low enough voltage... ill nick another one from work for a week.

So im now thinking about other ways of getting the VGPU higher... my GPU is my main limitation as you guys are at 340 and I cant even do 300

[Jeff]

Originally posted by Holst
Ive given up on my further Vmem mod for the moment as my lab PSU wont run a low enough voltage... ill nick another one from work for a week.

Lab PSU... Wonder where you planned on using that.

[Holst]

An unusual memory mod... problem is that im not sure ill be able to put the card back together once I try it... and feeding the memory 4.5v (the lowest my home made botch PSU would reliably go) was too dangerous...

[MrIcee]

Originally posted by Holst
I think its down to the way the Vreg works... and my 3.3v line isnt very high..

Ive given up on my further Vmem mod for the moment as my lab PSU wont run a low enough voltage... ill nick another one from work for a week.

So im now thinking about other ways of getting the VGPU higher... my GPU is my main limitation as you guys are at 340 and I cant even do 300

Matt

I'll tell ya, unless you got super duper phase change cooling 2.5v+ won't help. With a stock Mach II I can't run over 2.4v without it freezing up in the tests. Cooling is everything as this card runs hotter than hell. I've managed 345Mhz at 2.4v but 350 is not reliable even with more volts, it will randomly freeze.

Wishing I had a 404A baker block conversion on this Mach II right now.

With your originally developed vGPU volt mod, tweaking up the 3.3v line to 3.5v yields 2.5v+ vGPU on the Xabre.

Remember, some of us hitting 300Mhz in the beginning were using phase change cooling(liquid or direct die), and no volt mod to that point. You have done a great job on air alone indeed.

Can't wait for the Vmem mods, I'm basically stuck at a plateau now in the A64 class. I'll be switching out boards and hardware and give it a go in the 2300Mhz and under AXP class next. Now we'll see what the DFI Infinity will do in 3d

Randi

[Holst]

Im on watercooling although its gained me very little (15-20mhz on the core)

Ive done a VGPU mod MK2 that lets me have upto 3.3v on the GPU... ill post details in a new thread.

Ill be very interested in your results in the AMD class.

[YoupY]

Gonna give it a try without the cap tomorrow

I hope I don't kill my Xabre

I think it could be an option to build some kind of baybus to regulate the 3.3V and put that direct to the core without using the Vregs that are on the card

[Holst]

Read my high VGPU thread, you should be able to get what your after if you use my mod...

Just solder some thickish wire back over where the left hand pins are.

I think 3.3v is just too much for the card to handle..