X800-X850 AGP/PCIx POSSIBLE EXTRA PIPE UNLOCKING

[klootzakje]

agp and pcie arent the same card though they do have the same core :O

 

[F!reF0X]

Not a chance in Hell but nice thought.

Why not, it did worked on removing ram modules, Why should'nt it work on a gpu. With a good hairdryer with less air input can easy reach high tempertures. (try it with your hand on the air input) It much safer then an oven with 250 degress. I think we will see a Molex conector on the pcb melting away, if somebody is going to try it in the oven

 

[SPHERE]

http://froogle.google.com/froogle?s...&tab=if&q=x800+se+express&btnG=Search+Froogle

i cant find em on froogle lol

 

[klootzakje]

x800gt doesnt have a molex connector anyway

 

[SPHERE]

caps and other plastic parts don't like the oven lol

dude if you realy want to take your core off go ahead but i gurantee you you wont get it back on because it doesnt use pins it uses balls of solder that will smear all over the bottom of the core

 

[F!reF0X]

:| no gpu pins?

 

[SPHERE]

:| no gpu pins?

nope no pins just balls of solder...

 

[dolf]

Sean, I will prepare a file with what I need to be measured and will send it to you as soon as possible.

 

[dolf]

Guys, the idea to remove the GPU to fix and to return it on the PCB sounds nice but please
cool down. If you want to see what is on the back side of the GPU just ask your friends, fellow-citizen about burned X800, X850 with R480 (PCIx). Do the same like Sean did and the things will become much more clear and much more securely.
Otherwise probably some of you will successfully remove the GPU but to restore it on the proper position and with all connections in order - I don't believe unless some of you haven't a special equipment .

My suggestion is before to destroy a working card first to see what is on the back side of the core but from already dead card .

Even after the discovery for the back side cuts I will continue to work on R420 because I still have hope .

Successfully searching.

 

[fresko995]

Results...

hi dolf,
finally, I found some time to perform the measurements on my Non-VIVO-X800Pro in accordance with your instructions (I didn't resolder the resistor you initially told me to move to position R576!).
my results are in the attached document.
I some times had difficulties to get values while measuring the "PD"-point; it's the one from the "lasercut-bridge". that "dot" is damn small to put the multimeter-probe on.
also keep in mind that I'm not a professional in electronics as you are. performing some basic multimeter-tasks is all I'm familiar with.
Although Sean has found that unlocking the last pipelines might by impossible at all (resistors on the backside of the GPU), I still have hope. thanks to your everlasting optimism in this issue....
could you please send me your own measurement results so I can compare them as well?
later,
fresko

 

[dolf]

Hi Fresko,
Please check my file and try to answer to my new questions. Later on I will put my comments over the the results but first I want to obtain the all possible data.

I hope that after having the results I won't become pessimist.

Sean's results will be also very important but they will come later on.

For the final of my post I will use my favorite sentence:
What is pessimist? The answer: pessimist it is well informed optimist.

 

[Nobru_rv]

U need a hot air workstation mostly used for GSM phones repairing. Its easy to remove the core. It is VERY HARD to solder it back.

 

[dolf]

It is just talking. We haven't intention to unsolder the GPU. It is just wish.

Thanks for the reply.

 

[klootzakje]

hey dolf... will you need me to do some measurements? if so... does it matter much if i do some voltmods?

 

[dolf]

Hi Klootzakje,
I was busy this week and no time left for core investigation. Tomorrow if possible I will prepare the material for Sean and probably when I receive the feedback (from Fresko also) and summarise it I will be ready to give some difinite answer about locking .

If you want to voltmod the card do it. The mod is not connected with my investigation.

Please, if you have time measure the resistance between the pads on the core where you haven't resistors (empty resistore places).

Successfull modding

 

[SPHERE]

Hi Klootzakje,
I was busy this week and no time left for core investigation. Tomorrow if possible I will prepare the material for Sean and probably when I receive the feedback (from Fresko also) and summarise it I will be ready to give some difinite answer about locking .

If you want to voltmod the card do it. The mod is not connected with my investigation.

Please, if you have time measure the resistance between the pads on the core where you haven't resistors (empty resistore places).

Successfull modding

sry dude i kinda disected the core (a peice of the die fell off and i was kinda like meh! f*** it! so i broke it in half lol )

 

[dolf]

Sean I am very sorry to hear that. But even now the core can help (hope that you have the correct part of the dissection - that with laser cuts).
If possible try to obtain the following information:

1. (On the back side 5/4 bridges) they should have joined end. Is this end is general for both 4 and 5 bridge groups?
2. If it is not general for 4 and 5 bridge groups try to discover to what is/are connected. (All resistors from the front side are connected to Ground. But it is possible some of them to use connection through the PCB - core ball).
The first possibility is connection of the bridges to GND, the second one is to have direct connection to +3.3V (top of the front-side bridges). The third possibility is connection to ball (pin) from the core.
3. If you discover connections with these points, please cut the remaining bridges one after other and see whether the zero ohm connection is changed and to what (which ball, point from the front side)?
4. Can you find direct connection between not joined bridge part for every single end (after cutting) to ball of the core (or resistor on the front side)?

For me is interesting only the core-side with bridges (I hope). This information could be crucial .

Thanks in advance for the support and time spent but all that is done just for the community not personally for me .

 

[eeirertomb]

ya!

dolf your the best i just unlock the pipes!

 

[SPHERE]

Sean I am very sorry to hear that. But even now the core can help (hope that you have the correct part of the dissection - that with laser cuts).
If possible try to obtain the following information:

1. (On the back side 5/4 bridges) they should have joined end. Is this end is general for both 4 and 5 bridge groups?
2. If it is not general for 4 and 5 bridge groups try to discover to what is/are connected. (All resistors from the front side are connected to Ground. But it is possible some of them to use connection through the PCB - core ball).
The first possibility is connection of the bridges to GND, the second one is to have direct connection to +3.3V (top of the front-side bridges). The third possibility is connection to ball (pin) from the core.
3. If you discover connections with these points, please cut the remaining bridges one after other and see whether the zero ohm connection is changed and to what (which ball, point from the front side)?
4. Can you find direct connection between not joined bridge part for every single end (after cutting) to ball of the core (or resistor on the front side)?

For me is interesting only the core-side with bridges (I hope). This information could be crucial .

Thanks in advance for the support and time spend but all that is done just for the community not personally for me .

i only broke it in half cause i didnt think you where gonna get back to me

btw i kinda lied when i said broke in "half"


see by that i meant shreaded and peeled the layers apart

if its any concolation at all i am verry shure that ati does these cuts before the core is even on and before its even marked for pro or xt

(im quite certan that all of these perminatly locked pipes are bad even if you where able to unlock them )

realy sorry dude

if i get another ill be shure to look you up

i still have the lasercuts i believe but they arent attached to much else and im not at home atm

 

[SPHERE]

dolf your the best i just unlock the pipes!

kool

 

[Pavlin]

dolf your the best i just unlock the pipes!

Great !!
What's your card and what did you do exactly ?

PS: I'm getting soon Sapphire X800Pro VIVO PCI-E with 2ns Samsung memory and I'm gonna make everything to unlock it !

 

[ChAce]

If you get 2ns your chances of XT PE speeds are virtually zero. You need 1.6ns ram And if you're going for PCI-E, get the GTO2.

 

[Pavlin]

If you get 2ns your chances of XT PE speeds are virtually zero. You need 1.6ns ram And if you're going for PCI-E, get the GTO2.

Well i see no problem speeding the card at least to X800XT PE speeds . Also 1180Mhz aren't impossible for 2.0ns samsung. More important is though will the card can be unlocked. I've already order it, also GTO2 at the moment is 299$ in newegg which isn't best buy offer .

 

[4g0ny]

Hi again guys, i am now playing with my ASUS x800 PRO R420 too, but mine is one of the first cards.. ive read that if you have a lasercut on your core, you can connect the missing pipes with conductive pen? well i have a lasercut, but after removing the cover layer and connecting with a (graphite) pencil, i still have 12 pipes? did i screw up? or do i need conductive pen.. or is this regular bs and do i have the same prob as you all..
btw i got 2ns mem and the 500/500 bios, it says XT but 12 pipes..
any suggestions?

well anyway, Klootzakje has to RMA his x800RX cause we blew it up, hope he wont stay out of commission for too long

 

[klootzakje]

got caught in the "a little more" syndrome