I thought the heatsink was the only think that would get lapped, but when I saw the GPU as well I'm like
Did you guys lap both or did you just lap the heatsink? I really dont' see the point in risking a card for a few degrees in temperature. So I'm thinking I'm probably just going to remove the shim and lap the heatsink. I'm not looking for mega gains just better than what I had (atleast 370/335).
The kit came with 600, 800, 1000, 1500, 2000 grit sand paper and polish compound. I've read that polishing will doesn't help but actually worse if you use it. ANyone know?
Don't lap the R300 die. There's no way you can improve the surface unless you have $millions worth of equipment and you'll almost certainly damage it anyway.
I'd give it a good clean and then lap the heatsink to a shine - you can use polish if you want, just make sure you remove all the residue afterwards.
Yes, don't go anywhere near the die with sandpaper... the only chips you should lap are the ones with the black cover over them... heatspreader (eg Northbridge, GeForce graphics cards, chips on soundcards).
The R300 just looks fragile (like the Athlon/Duron processor dies') and I would think that too much pressure, or scraping away the surface, would most definately damage it.
Desktop - XP2100 @ 2.4GHz, Abit NF7-S, 1gb DDR500 Crucial Ballistix, 120 + 160gb HDD, Windows XP
Laptop - Athlon 64 3000, 1gb RAM, 40gb HDD, Radeon Mobility 200, Windows Vista
Server - P4 1.8GHz, 768mb RAM, 2x 40gb HDDs, Win2k3 Server
Cool. Cause i wasn't going to do it anyways. Hmm Do I need to re attach the shim if I remvoe it? I plan on using the stock sink, so I'de imagien retaching it won't be a problem if I apply preasure flat on the die.
Ok it's done, but I'm seriously nerveous that I damamged something. When prying off the shim I heavily sracthed underneath where the shim used to rest. After I lapped the sink and reattached it I also noticed that with the heavy fan mounted that the pins struggle to keep the sink mounted tight to the die. I'm not sure if the vibration of the fan will loosen it.
But I'm running it so far, and everything seems ok. I just ran 1 cycle of 3DMark2003 just to see if I had indeed damaged something. It went fine. I guess the next step is an endurance test. Really upse that ATI put this shim in the first place. You'de think they would have some high tech peice of equipment to attach this accurately.
Originally posted by Tuxy the pins struggle to keep the sink mounted tight to the die
I think the pins must be longer than needed (of course, with the shim fitted they are the correct length) as my card has done exactly the same. If I pull the heatsink down slightly, there is a small gap between the die and the heatsink... it's left me feeling quite nervous, and at the first sight of artifacting, i'm checking my card
Desktop - XP2100 @ 2.4GHz, Abit NF7-S, 1gb DDR500 Crucial Ballistix, 120 + 160gb HDD, Windows XP
Laptop - Athlon 64 3000, 1gb RAM, 40gb HDD, Radeon Mobility 200, Windows Vista
Server - P4 1.8GHz, 768mb RAM, 2x 40gb HDDs, Win2k3 Server
ATi don't make the boards themselves. The shim is supposed to be level with or only fractionally proud of the die. The reason it's higher is because of bad QC (and a "better be safe that sorry" approach!). You can get lucky and get a board with a well-installed shim (like ol' man) but the majority of AIBM boards I've used have had a little too much clearance for most TIMs to span effectively.
Pretty easy to increase the HS mounting pressure - just use washers.
Originally posted by MuFu ATi don't make the boards themselves. The shim is supposed to be level with or only fractionally proud of the die. The reason it's higher is because of bad QC (and a "better be safe that sorry" approach!). You can get lucky and get a board with a well-installed shim (like ol' man) but the majority of AIBM boards I've used have had a little too much clearance for most TIMs to span effectively.
Pretty easy to increase the HS mounting pressure - just use washers.
MuFu.
No my shim was high also. i simply put a piece of copper between the core and the HS to raise the HS off the shim. That was the whole point I was making. The piece of copper is just small enough that is does not touch the shim and raises the HS off the core. You need to take a closer look at my pics. The shim has nothing to do with the HS install. I didn't want to remove my shim. It really is not necessary unless you want to do a hurry up job and risk wrecking the GPU PCB. Tuxy sounded like he got prety close to doing it. I have seen alot of people cut pcb traces trying things like that.
If you use a piece of copper as thick as mine you can use any size HS imaginable pretty much as it raises the whole HS off the caps etc.... My HS is a coolermaster PIV HS.
Ah, ok. Thought you had a large copper block covering the whole package. What's the interface between the copper and the Al?
You'd have to be a total klutz to damage the GPU when removing the shim, IMHO. You only need to stick a knife about 0.5cm under one corner to be able to lever it off. Not everybody wants to mess around with copper blocks, although I agree it seems like a very effective solution. Hell, I have a big copper block on mine.
I ended of mouting the fan on a bracket away from the sink so it will sit better on the core. Dropped 1 degree so now I'm registering under 30C at idle.
Man - I scraped right down to the green on mine (no reason really - just looks neater with no glue on it, lol). As long as you don't knock off any of those SMT components it'll be fine.
So far I've managed 365/330 anything over that I start to artifact. I realize that I didn't do a good job lapping the stock sink when I went ahead and hacked away at a PII heatsink and lapped it.
Seeing how I'm not impressed with the stock heatsink, I'm going to go ahead and mount a larger aluminum one. Only this time much more time and effort to maknig the mounting holes. I'm gonna go to Home Deopt and look for a drill bit that creates a screw impression so I adjust and secure the sink properly to the card.
Reply With Quote
Bookmarks