Hey Guys,
is here anybody wich knows a HD4890 VMem mod?
Thanks a lot.
Best regards
Dave
Printable View
Hey Guys,
is here anybody wich knows a HD4890 VMem mod?
Thanks a lot.
Best regards
Dave
bios mod Guide http://www.xtremesystems.org/forums/...d.php?t=230111
^That has nothing to do with vMEM.
Hey Kabauterman :D Post up some pics of the card and we'll get you sorted out. Id be surprised if there arent already guides for it though.... its quite a popular card? :D
front http://www.techpowerup.com/reviews/P...front_full.jpg
back http://www.techpowerup.com/reviews/P.../back_full.jpg
I will take a look on this too :)
Maybe these pics will be ok, to find the vreg provide the Mem voltage....
Thanks!
VT242WF is the chip of interest, haven't seen it on any other card.
Now, the question is, where do I get a spare VT242WF to replace the one I busted on my 4890...
:D
Ouch... :shrug:
Plz can you mark this chip in the picture?
How did you blow this chip up... did you try a VMem Mod?
The front/back pics from TPU you posted are of a Powercolor non-reference card which has different vMEM chips than VT242WF on AMD reference card. On the linked Powercolor card memory voltages (two of them) are controlled by those 8pin Upi-Semi chips, whose model I can't tell by those pics. One is on the backside below CF connectors and one on frontside just below the small fake copper (= aluminum) heatsink near the 6pin plugs.
Month or two ago, I managed to chip away literally a dust-spec size a chunk of a corner of one of the VT242WFs, while fumbling with a custom VRM heatsink. VT242WF is a square shaped CSP part around 3mm each side and the chipped corner is so small it's just barely and barely visible to the naked eye. One bloody small bit missing (match for size reference) and no-worky... :rolleyes: Result was, the second memory phase broke = no image. I had to jury-rig both MvDD and MvDDQ to run from one single phase as both are 1.5v stock. But apparently there's a reason why the phases were non-unified to begin with... ŠP Anyways, the card works and is artifact free depending on game with underclocked memory (700-750MHz), but stock memory freq is totally unstable and artifacts heavily. Once I get blessed with enough inspiration, and repair my hot air-soldering iron I'll try to seperate the phases by hack-sawing an off board power phase to replace the broken one.
That card isn't a reference one
I'm guessing:
http://a.imagehost.org/0060/men.jpg
Thank you very much!Quote:
Originally Posted by LardArse
What do you think how "big" the VR must be? 100k? 50k?
Likely a very low value (50-100Ω). You can find out by measuring the resistance between ground and the marked pads and multiplying the result with 20.
those are on the front side of the board,can someone post reading points and place for the res on the back side,i dont want to run card with cooler os or hanghin things inside the cooler.
thanks
I know that is very old thread..but can anyone re-upload pics?
post 11..thx
vmem and vgpu