Im game if someone will confirm it should work.
Ill pull them from a dead mobo.
Printable View
Im game if someone will confirm it should work.
Ill pull them from a dead mobo.
You can "double strap" them (put two in parallel). Should hold up fine as long as the traces dont catch fire due to current draw :D I'm going to google "double strapping mosfets" and read up on it some more. I suggest you do as well if you read this and go to it before I reply back with something important lol ;)
I got the shader up a bit farther on the Asus (2375+mhz) but i'm running into overcurrent protection at 1.95v loaded 904/2375/475.. I can run ATi Tool's artifact tester at 954/2480/475 2.15v lol.
Sparkle has arrived, should have it in my hands tomorrow night.
MOSFET STRAPPING INFO:
Scroll down to point #5
Add a ferrite bead to remove parasitic oscillations (pdf)
MOSFET/Driver IC 101
I'd try some of that if I had more time Stevil, the mosfets on this thing are about to burn thru the pcb even with cooling on them. Been trying to catch Andy...this card doesn't like more memory voltage so I'll have to play with it like this. Need to find 122 more points :D
'01 ORB Link
Not sure how much more benching time I'll get, posting an '03 backup ;)
'03 ORB link
http://img.photobucket.com/albums/v3...y/03_10321.jpg
Squeaked into #1 for 03 with that backup, wonder if SpikeP has a backup too? :toast:
Nice core clocks, my card needs 1.9v for 800+ and therefore is not a "golden" chip gpu wise, in fact it is more s**t brown. memory is good though thanks heavens.
Regards
Andy
I'll take your memory clocks over these core clocks :up:
In fact 5 more mhz memory clocks and I think I could beat you drago's based on opening fps...I can't complete any test in '01 @ 555 however. Based on the data sheet for these modules (re-badged FP-25 Hynix), vref & vddr are both @ 2.0v for the over 400mhz parts. These cards all use 400mhz parts and my card is already running 2.0v default. 2.05v and it freaks out. Only thing I can think of is vref may be on a separate feed.
Hehe. Been waiting for someone to compete on 03. better get the card back in teh rig and a few more runs.
Towards Zakelwe, do i have backup........erm, yeah!
on the same day i benched through with this score. Left it till someone broke the last one. not a drastic increase, but still gets me back up there :up:
http://img141.imageshack.us/img141/4194/10427en9.jpg
GHZ is quicker in GT1 and GT4. I am faster in GT2 and GT3, so interesting!
ORB Link: http://service.futuremark.com/compare?2k3=5736589
better find some ways to increase the vcore and vmem again now. memory is being fed 2.08v at the moment. So close to 600mhz now.
- Paul :p:
hot damn nearly 600MHz on the Memory VERY nice SpikeP
Yeah, people would kill for those memory clocks. Actually it's turning out quite nice because there is a good mix of people with one or the other so is balancing.
Very nice score, popped you back at the top :toast: I know G H Z is fairly competitive so his claim of "I cannot bench much more" is what I'd expect him to claim before he benches like a mofo :yepp: :D
I've got some pictures of my new heatsink modified to fit on the half height card.
http://www.zen141854.zen.co.uk/heatsink.jpg
Close observers will spot where I trimmed it on one side to fit the motherboard. For people who have not spotted it, it is the side closest to the camera.
Here it is on the motherboard
http://www.zen141854.zen.co.uk/heatsink2.jpg
Have to prop it up on the keyboard so it boots.
Here's the reverse of the card, lots of duct tape
http://www.zen141854.zen.co.uk/duct.jpg
check the pot, I like mine at least 1 inch above the PCB waving about with duct tape so it does not short. If only Kingpin or hipro new these sorts of tricks eh?
Early morning shed temps
http://www.zen141854.zen.co.uk/shed.jpg
The BIOS reads low but the 780i nvidia chipset still is pretty warm.
Finally, one of the advantages of benching in a shed
http://www.zen141854.zen.co.uk/shed2.jpg
After that I watch the moon and the stars come out and it is pretty magical some nights.
It's REAL nature at LOD 0 and looks great.
Regards
Andy
Nice Scythe Ninja Andy :p:
i have already seen your score on the ORB Zaklewe. just need to get some software issues sorted, and i shall try claiming that spot back again!
- Paul :p:
i'll do some benching this weekend and slap together a TTBT on my card and see how i go
looks like it's all about RAM >> nice runs there spike :up:
don't touch LOD it looks like it's real life nature :shocked: :D
that's one reason why i would wake up at 4am to go fishing mate
stunning sunrises i tell ya
maybe i'll grab a camera next time
T_M you should be looking to cool your RAM mate not the core
covering it like that is a bad idea matey.......
have a look at GHZ and spike
it's alllllllll in the RAM
Wonder how the ram would deal with phase change? Might fab a little plate so I can strap my evap to the ram instead of the GPU. Slotted 8 hours this weekend for various benching, time to make my bid for number 1.
I'm having a few technical problems at the moment. such as I have blown my card to kingdom come :rofl:
So you have all the chance in the world of claiming that spot back !
Well I did last for 3 weeks :D
In was running 2.25v and it kept hanging and so I decided to run rivatuner to see temps and was getting 106C and then it would not boot up.
Either I put too much aitric epoxy on or it had not cured or some other slip up due to lack of practice. I could try and knock it off with a tap at 90 degrees to the sink and start praying but I think it is an ex parrot.
Regards
Andy
hehe, did you get an '01 done before it fried? :D
My only hope is ram cooling also, we'll see.
Hmm, I think I :banana::banana::banana::banana:ed up I made this mod on my motherboard, and it worked perfectly until now.
Earlier this evening I decided to game a little on my rig, I pulled out the 8400GS and replaced it with a 8600GT. I also pulled out the connector to the mod.
The mobo was stuck in POST. I tried to replace it with the 8400GS, and the same problem occured. (the 8400GS got insanly hot btw)
Then I took a look under the mobo and this is what I saw:
http://img98.imageshack.us/img98/8794/picture010gb5.jpg
http://img98.imageshack.us/img98/174/picture012cr3.jpg
Well i will not be doing any more runs till sunday night i think. possibly tomorrow night. card is still with me.I am running 2.2v through the ram now. it freezes at 610mhz, but seems good for 605mhz. I need to get myself a old hdd to put xp on so i can hopefully gain a bit more through that as well. Also i am finding the shaders are meaning more than the core on 3dmark03.
Finally. Realised people are running 03 on either 3.0 LOD or 3.6 LOD. i`m running 3.4 :up:
- Paul :p:
I use LOD 3.0 because I have good memory speed and therefore extra LOD is probably not that good an idea compared to guys short on memory bandwidth. You should notice the first game going down in fps with increased LOD for game 1, so it's a compromise. The only way to judge is to run loads of runs and see what's best for you. To be honest I have not done this much this time around.
Well my card is not dead :eek: I stripped the volt mod off and put the card in the freezer for 20 mins and then gently knocked the heatsink off with a hammer at 90 degrees to the chip tap tap tap and it came off without removing the gpu and I booted it up with no heatsink on into BIOS and it was OK. phew. It might be a bit tired now though of course, ie might not do 3d at all or such.
I think the problem was I put too much artic silver epoxy on, I am not expert in using it as it has been several years. It acted as a blanket. So now I have gone for a smaller heatsink and a lot thinner layer of artic silver epoxy.
http://www.zen141854.zen.co.uk/zalman.jpg
It looks a lot neater but I have to say I am not overly confident about the results. Even if the gpu is fine my memory might have decided to go back to 550 again !
Who's idea was it to have this competition again ? :rolleyes:
Regards
Andy
Good to hear it still boots. Mine has been through hell. Bad solder mod, smoking cables. 3v through ram for a split second. hehe. still with me and kicking. I am gona have to have 3 or 4 wires to the GPU next though. Going on the core from 1.7v to 1.8v, the actual voltage went from 1.69v to 1.74v. wires get warm as well.
Heres a few pics of the wire that melted. I stil duno what i did wrong, but it must of been drastic!
http://img185.imageshack.us/img185/6683/p1000343ek5.jpg
http://img153.imageshack.us/img153/6978/p1000344ks4.jpg
http://img185.imageshack.us/img185/4820/p1000345dh4.jpg
The LOD, i have to agree with ya about the first test there, but i still found i was gaining on GT2, so the score ended higher.
Just like to point out as well that when i started this contest, i expected to be way down the list, so even if i end up anywhere in the top 5 of 3dmark03, i will be happy :)
Don't let that kind of talk make you think i am finished yet. ooooo no! :D
- Paul :up:
lol these poms are such showoffs :p: :D
I love the smell of burnt plastic in the morning, or was it napalm?
These little chips are pretty tough, the Dutch guys have been giving them hell as well of course.
Good old TSMC 80nm process :up:
Regards
Andy
You need to clamp the heatsink to the core to make proper contact while the glue cures lol...
Why do low end cards always seem to survive such horribile adventures?
Try this with a $600,- Ultra and it will be dead as a doornail. :smoke:
If only i had the nerves to put 1.8V through my G92 core, lol, i'm getting nerves feeding it over 1.5V
I take it shaders are pretty sensitive to raised voltage?
Good going guys! :up: