8800s OVP Mod ......

[7499richard]

ok yes that makes sense now, thanks for clearing it up.

so is there any way to beef up the supply or anything to correct these horrible voltage spikes that are shutting down our cards, I really want to be able to oc my card more, as I am going to keep it for awhile longer now that it doesnt look like there will be another high end card from nvidia for a little while.

Is that X38 PWM a viable option?

[wittekakker]

ok yes that makes sense now, thanks for clearing it up.

so is there any way to beef up the supply or anything to correct these horrible voltage spikes that are shutting down our cards, I really want to be able to oc my card more, as I am going to keep it for awhile longer now that it doesnt look like there will be another high end card from nvidia for a little while.

Is that X38 PWM a viable option?

Not really, mainboard PWM is not easy to reconstruct, not like lineair controllers. You could use a high-end mainboard because of its better PWM design, but I don't really see anyboddy doing it, costs to much

[7499richard]

well have you added capacitors or anything else to your card to increase stability?

If so what did you use and where did you get it?

Also were there any real improvements in overclockability after the mod was done?

[wittekakker]

well have you added capacitors or anything else to your card to increase stability?

If so what did you use and where did you get it?

Also were there any real improvements in overclockability after the mod was done?

Look at Hipro's post, he has done cap modifications. I'm planning on doing some too, can't find the right coils, but I do have some 6,3V 1800uF caps from a dead MSI-6758 P4 mainboard.

[G H Z]

Unfortunately no.... or else I could see if I could sovle the "VDroop problem".....

hipro I think the 8800 GPU mod was located using the PX3535 datasheet, I had this saved on an old drive so maybe this will help for droop



OVP mod is connected to pin5 upper bank (from right), Fault2 - based on largon's GTS pic on page #1

[G H Z]

I have another question, can someone tell me what pin the vcore mod traces to (based on the datasheet)? It's not OUT4 (pin 6) as I had previously read somewhere.

[hipro5]

I have another question, can someone tell me what pin the vcore mod traces to (based on the datasheet)? It's not OUT4 (pin 6) as I had previously read somewhere.

Resistor's inner is connected with OUT4....

[G H Z]

Hmm, then why isn't there apparently any connection to it?

[hipro5]

Propably coz they don't use the schematic above....

Problem is that I cannot understand it.....Why we use an output to Vmod the VGPU of the VGA?....
Maybe they have connected OUT4 with Vsence directly and that is a fine explanation of the HUGE Vdroop this VGA has.......I'll investigate it more.....

Can someone measure the Vdroop at the DEFAULT Volts of the VGA?....Is it big or not?....

[wittekakker]

PX3535 is different from PX3540, just look at the amount of VID pins this chip has. Another example, the ISEN pins are different located, I've posted this somewhere on this forum. I don't think we can use the PX3535 datasheet as example to modify the PX3540.

[Oc-Ghost]

Why look at a different chip? google is your friend
PX3540

[wittekakker]

Why look at a different chip? google is your friend
PX3540

in the 8800 modding game.

We are way past that one, the problem is that the product sheet is not a full datasheet with pinouts etc.

[T_M]

The other day i went in blindly (since my resistance measurements are all over the place still for that resistor) and soldered on the 100k again, with the ground line on an aligator clip. Got the system running without the mod engaged and then clipped on the ground. Nothing burned up and the system ran 3DMark fine.

That was right before i killed the mobo of course so didnt get to test whether the OVP was actually adjusted heheh.

[Alexxx650]

Hi everyone. I need your help. There is card 8800 runaways before frequencies 729 on water with vGPU of the chip 1.490 card can work at 756 but more 1.510 put (deliver) can not since card does not work. The Question: why OVP modes do not work at 8800Ultra? Water running +8 degrees. The Card in simple 28 idle in load 33.

My system:C2D Conroe SL9S7 E6700@3900 487.5x8 1.4875 Swiftech Apogee™ GT
ASUSTeK Striker Extreme; nForce680i SLI; vDroop mod, NB/SB - Swiftech MCW30
Crucial Ballistix 2 x 1024 / Micron D9GMH 533@975 4-4-4-4-T1-16-20-6-7-3 2.3250v Dominator AirFlow
Видеокарта: PCI-E 768МБ Leadtek WinFast PX8800 Ultra 783(1998)/1260 OVP Mod, vGpu 1.489, Vmem 2.24, EK, PCI-E 100@125
320 Гб 3200YSx2 (RAID edition) NVIDIA STRIPE 596.18G
Thermaltake Toughpower 650W V2.2 (20/24+4+6+8pin, d140мм)
DVD RAM Plextor PX-760A
Logitech G15
Razer Copperhead
Razer Mantiz Speed
Technics RP-DJ1210
Microphone: SM58 Vocal
Microsoft Windows XP Professional x64 Edition

[hipro5]

It's simple......Think about "CPU".....When you feed your VCore with more Voltage, it stops "handling" it when you go too high....IF you put it under DI or LN2 for example, it can handle this high VCore and it handles even more than that......This way is how GPU handles the VGPU.....It's the same thing...

[T_M]

At only 1.51V you are not going to have to worry about OVP.
OVP generally kicks in around 1.65V+.

On my card my current max is 1.662/2.25V when running 938/1188MHz before i hit OVP and the system reboots then registeres only 0.100V vGPU. Thats when i know i failed due to OVP.

George, not sure if you missed my question earlier in this thread:
About how many turns of the 100K VR are required to get the desired 5.1K OVP resistance?
Because i cant measure resistance properly (as discussed before) im just running it blindly.

[Alexxx650]

Thank for answers. It turns out to me it is necessary to make better cooling, instead of OVP?

[T_M]

Yes, better cooling means you can put more volts thru.
Once you are using Dry Ice or better, then you should worry about OVP

[oohms]

hipro5 what about putting a resistor in series with VSENP and/or VSENN?

http://208.109.51.75/PrimarionMain/D...s/PDB-3540.pdf

[hipro5]

hipro5 what about putting a resistor in series with VSENP and/or VSENN?

http://208.109.51.75/PrimarionMain/D...s/PDB-3540.pdf

I think that we already have a BIG Vdoop on this design.....IF we put a resistor as you say, it'll make things even worse as for Vdroop.....
We HAVE to find out hte full datasheet of the IC so to search the design and eliminate the Vdroop of it if possible....

At least in my way of doing it, I have NEVER encoded shut down problems due to OVP.....I have frezzings if the core goes too high and that's it....I meassure the VGPU and it's there.....

[wittekakker]

I've added 3x 1800uF caps to my GT, and changed from Thermalright HR-03 to chilled water, no way my card would go any higher then before, removed all my mods now because there is no way I can get this card clocking any further. Do the inductors really boost the Vdroop that much?

[hvtecgr]

Hi hipro!!!my dfi p965 has the primarion 3539!!!I've searched a little and found another datasheet from other company which makes the excact same chip and i have the full datasheet from that!3539 has a pin named vr_ready and i came in conclusion that this must controll the vdroop!!But i'm not sure how it changes from 0 to 1!!!I don't know if that helps cause i don't see any pin named vr_ready in px3535!!!

[T_M]

Interestingly, my friend pro has measured his OVP resistor to find an inconsistent value too, so its not just me.

[T_M]

Yesterday i ran into trouble with my Ultra.
My vGPU mod was going weird, so i removed it to reapply the connection to resistor.
Unfortunately it appeared that the resistor was damaged (or at least the pad).
After much messing around and variouys attempts to solder it, i finally got what i thought would work.

Then i took the opportunity to check my OVP mod again, and while i was still getting different readings, occasionally i would get ballpark correct ones, which allowed me to adjust the VR to give me the required 5.1k.

Plugged the card back in to check my repair work, and holy 2.52V!
Looks like the OVP mod works
Pity about my vGPU mod.... so i swapped the 500 Ohm for a 10K, reran and the card is back working again, presumably with no OVP

[dinos22]

2.52

we need the ovp for 8800GT

people are reporting OVP kicking in at 1.3v which is only good for 1GHz on core with dice (lol ONLY haha)