Coldbug fix for MSI P55-GD80, Trinergy, and Fuzion

Garrett_Wu @ MSI

Quote:

---

Dear All

many overclockers were complaining coldbug boot issue during MSI Elite OC Summit. So after summit, our R&D guys work day and night, try to find out a solution.

And guess what, we found it!

here is official coldbug fix for MSI P55-GD80, Trinergy and Fuzion. Anyone happens to have these boards, please try out and feedback to us!

---

MSI P55-GD80 (back side)
http://farm3.static.flickr.com/2693/...4b639cfb_o.png


MSI Big Bang Trinergy and Fuzion (front side)
http://farm3.static.flickr.com/2758/...25378ca6_o.png

Thanks MSI for working this out for us :clap:

Lets hears some feedback on this mod

WOW i will test them asap :)

Damnit CHarlie, you just proved me wrong, I dont like that ;)

Charles, what are those resistants for? Can you give further details??

Looking forward to hearing if this really works

Looks like MSI is working hard on fixing their boards, good job! :up: :clap:

Nice to see this kind of co-operation between manufacturer and community...+1 :up: Might have to snag me an MSI board!!:D

Sweet! I'll test it on my GD80 for sure

Quote:

---

Originally Posted by Expat GriZ

Nice to see this kind of co-operation between manufacturer and community...+1

---

True :up:
Thanks for sharing FUGGER. :)

Im benching on Tuesday... if I have time + inclination, i'll give this a shot :D (Dice)

Thanks MSI!

Awesome! Big thanks to MSI for listening to us!

Quote:

---

Originally Posted by K404

Im benching on Tuesday... if I have time + inclination, i'll give this a shot :D (Dice)

Thanks MSI!

---

You don't need this mod for dice. The boards were having problems at about -90~-100. They should work fine on dice.

My 860 doesnt like dice at all.

Quote:

---

Originally Posted by K404

My 860 doesnt like dice at all.

---

test with dice please :D

thanks msi and fugher

I'll see if I can find the time :) DoD (Day of Dice) is usually a race against the clock and I have a list of scores I need to crank out :p:

Quote:

---

Originally Posted by M.Beier

Damnit CHarlie, you just proved me wrong, I dont like that ;)

---

You see?.......It wasn't "bios related" after all... :p: :D

Quote:

---

Originally Posted by hipro5

You see?.......It wasn't "bios related" after all... :p: :D

---

hehe, yeah, ok I admit ;)
But that is good news, good thing that I aint right always it seems? :yepp:

Quote:

---

Originally Posted by hipro5

You see?.......It wasn't "bios related" after all... :p: :D

---

:rofl:

I know you don't meant it as a joke, but I'm laughing so hard you can't imagine :up:

coldbug fix for P55-GD85, but we still recommand overclockers to use P55-GD80/Trinergy/Fuzion.

http://farm5.static.flickr.com/4017/...b1cc0ae3_o.png

Garrett_Wu

What does that resistor related? :) Any X58 mobo similarity curcutry to mod? :D

Garrett_Wu
coldbug fix for MSI P55-GD65?!?)))

Quote:

---

Originally Posted by TiN_EOF

Garrett_Wu

What does that resistor related? :) Any X58 mobo similarity curcutry to mod? :D

---

something with a PCI-E signal I read in hwbot forum. without the resistor it will be amplified.
source

I think just a kind of termination was removed...

:up: Good news
:yepp:

Thread necro-ish.... Are there any negative effects to shorting these points? Why did the companies choose to set the resistance the way they did?

Will it be revised for all future boards for better sub-zero behaviour?

:)

You increase voltage within CPU - potentially dangerous

Manufacturers didn't set the resistance, they just followed Intel guidelines.

I was under the impression it was signalling, not power? :)

Quote:

---

Originally Posted by K404

I was under the impression it was signalling, not power? :)

---

It's both actually Kenny if you think about it. Looking at the description, it sets the standing bias for the PCI/e bus reciever stages (presumably to facilitate an adequate level of voltage swing/ideal curve). IF it's increasing bias
to the receiver stages, yes it may be potentially harmful long term. I don't think a shorted value is strictly necessary actually. All vendors just copied what Sham did because they were under pressure to do so.


In this instance, a larger PEGR BIAS change is only required during post I think. Had companies of found this out earlier, I'm sure they'd of added an automated routine to change the value for a few seconds. Mind you, it's usually EVGA that do that kind of stuff. Most other boards are completely devoid of anything to do with active compensation or signal controlling circuitry, which is one of the reasons I sometimes struggle to see what they're asking you to pay for on their top-level boards other than the bling, a :banana::banana::banana::banana: load of chokes and a funky box.

As a precaution, what I would do, is only use this mod while benching sub-zero. If you return to air/water cooling - reverse the mod.

later
Raja

Thanks Raja :D

Have EVGA set it as a pulse?

Is shorting the resistor on MSI boards shorting out the entire resistor network, or just reducing the voltage/ has the signal amplitude become the supply voltage? Whats the signal strength like before and after?

Im... possibly over-simplifying the circuit here :p:

Quote:

---

Originally Posted by K404

Thanks Raja :D

Have EVGA set it as a pulse?

Is shorting the resistor on MSI boards shorting out the entire resistor network, or just reducing the voltage/ has the signal amplitude become the supply voltage? Whats the signal strength like before and after?

Im... possibly over-simplifying the circuit here :p:

---

I don't think it's setup in 'pulse' value as you put it on the EVGA board, I think the value at post remains static (don't quote me on that though, Sham's a bit of a whizz at fiddling with this :banana::banana::banana::banana:). I have not looked at EVGA's full implementation so can't comment on whether shorting the resistor results in open to ground - depends on what else they have in-series to provide the 63 step control.

On other boards, yes, it's fully open because the resistor is to ground. The drive voltage is likely VTT (it supplies just about everything on these boards). I can't comment further as I have no idea how things are setup internally, although I do know these are differential pairs (most have a reference voltage set around the mid point), so I'd think the change in bias just results in satisfying a required time oriented bus data transmit sense during post (change in slew behaviour).

The board engineers know a lot more than I do and have access to stuff that most of us don't. Unless there's soemthing radically different to how it appears this stuff works, the straight short to ground is a typical quick fix manoeuvre without any real finesse or consideration over long term implications.

later
Raja

Ok, had a bit more time to look at this now that I've got some work out of the way. The resistor sets BIAS comp, which is not quite the same thing as setting the standing quiescent BIAS for a FET -BIAS comp is used to attenuate (trim by freqeuncy) the differential signal bias across the transmission line to mitigate noise on an inactive/un-terminated line during transmit (so that noise does not cross-leak into an 'active' reciever thus corrupting a logic interpretation).

As such, the long term effects of full grounding may not be nasty (only Intel can answer that properly though). That all being said, there is likely to be a sweet spot for any given operating frequency and temperature to wring the most from the PEG bus and it will vary from board to board and CPU to CPU depending upon layout and trace capacitance (though more would need to be know to understand how it all works fully during a transmit).


later
Raja

my CB/CBB mod on fuzion
http://img39.imageshack.us/img39/4107/cbmodfuzion.jpg
will test it later with clarky
:p: