[First Look] DFI 680i LT NF-680i LT SLI (56K Warning~~~)

[MikeMK]

LOL... yeah, its a pain that Intel cant make their IHS's anywhere near flat! LOL!

As for voltage stability - ive used the galaxy on a fair few setups over the last few months - mostly SLI 8800 setups, and its been rock solid. However, I have noticed the DFI is reading the 12v rail as 11.71v at stock idle, which is a bit odd, because on all other setups its been around 12.20v and not dropped a touch when you add load. I'll reserve judgement until I do a bit more testing.

Do you think the galaxy actually prevented you i getting an overclock?

[Borgod]

For me it's a combination of poor voltage stability, high CPU temps because of insufficient radiator heat dissipation and outrageous NB temps that I can't control no matter which orientation, fans and remounts I do with the NB cooler.

I'm ordering a mips wc kit next week, so NB will be watercooled in the not too distant future hopefully.

I'm also going to replace my Galaxy with an Ultra X3 1000w PSU... and it's going to cost me an arm and a leg to import one

[Spazilton]

LOL... yeah, its a pain that Intel cant make their IHS's anywhere near flat! LOL!

As for voltage stability - ive used the galaxy on a fair few setups over the last few months - mostly SLI 8800 setups, and its been rock solid. However, I have noticed the DFI is reading the 12v rail as 11.71v at stock idle, which is a bit odd, because on all other setups its been around 12.20v and not dropped a touch when you add load. I'll reserve judgement until I do a bit more testing.

Do you think the galaxy actually prevented you i getting an overclock?

The DFI is misreporting all PS voltages low.

Pretty much all the PS voltages are reading .3v low, Ive verified this with a DMM.


I finally found a work around for my throttling issue, and its not just related to Quads. My E6600 at 3.7 1.5975v bios would start to throttle after a few minutes of Small FFT even though core temp was 51c.

RM clock autoloading with TM1 and 2 disabled has solved my problem. Not the soulution I want, but it will work.

@MikeMK, you might want to reconsider your WB for a Q6600, the Storm doesn't work all that well with Quads, since the coolest part of the block is inbetween the 2 dies.

[Grinch]

The DFI is misreporting all PS voltages low.

Pretty much all the PS voltages are reading .3v low, Ive verified this with a DMM.


I finally found a work around for my throttling issue, and its not just related to Quads. My E6600 at 3.7 1.5975v bios would start to throttle after a few minutes of Small FFT even though core temp was 51c.

RM clock autoloading with TM1 and 2 disabled has solved my problem. Not the soulution I want, but it will work.

@MikeMK, you might want to reconsider your WB for a Q6600, the Storm doesn't work all that well with Quads, since the coolest part of the block is inbetween the 2 dies.

I would suggest a fuzion or an apogee GTX

[MikeMK]

I would suggest a fuzion or an apogee GTX

Funny you should say that, ordered one a few days ago (a fusion that is)... arrived this morning. Gonna sort out my watercooling tomorrow and stick that on and see what it can do. However, looking at the IHS it may be better to lap it aswell.

[Jobeo]

For me it's a combination of poor voltage stability, high CPU temps because of insufficient radiator heat dissipation and outrageous NB temps that I can't control no matter which orientation, fans and remounts I do with the NB cooler.

whats your NB volts and what kinda temps are you seeing ?


regarding PSU and the +12vl i had a PC P&C 750w and guardian reported the 12v rail around 11.90v but i upgraded to the 1kw and now my 12v is rock solid at 12.03... it doesn't waiver so i'm very pleased with this new PC P&C PSU...

[Grinch]

Funny you should say that, ordered one a few days ago (a fusion that is)... arrived this morning. Gonna sort out my watercooling tomorrow and stick that on and see what it can do. However, looking at the IHS it may be better to lap it aswell.

i lapped mine a while back and it knocked 6C off of my load temps

[Borgod]

whats your NB volts and what kinda temps are you seeing ?

regarding PSU and the +12vl i had a PC P&C 750w and guardian reported the 12v rail around 11.90v but i upgraded to the 1kw and now my 12v is rock solid at 12.03... it doesn't waiver so i'm very pleased with this new PC P&C PSU...

My NB is at stock voltages right now and it's idling at 68 degrees.
When I bump it up to handle 3600Mhz it is just shy of 90 degrees and I can't get it stable.

I've seen my 12v rails go as low as 11.7v under load.

i lapped mine a while back and it knocked 6C off of my load temps

Now thats interesting!
Had you lapped your CPU aswell?

Perhaps the benefits of a convex base only come from a hard mount? Since I refuse to do that and will only mount my block by hand (until they bleed however) just maybe a lapping of the block may benefit aswell.

6 degrees!!

[Grinch]

My NB is at stock voltages right now and it's idling at 68 degrees.
When I bump it up to handle 3600Mhz it is just shy of 90 degrees and I can't get it stable.

I've seen my 12v rails go as low as 11.7v under load.


Now thats interesting!
Had you lapped your CPU aswell?

Perhaps the benefits of a convex base only come from a hard mount? Since I refuse to do that and will only mount my block by hand (until they bleed however) just maybe a lapping of the block may benefit aswell.

6 degrees!!

cpu was the only thing lapped..

[Borgod]

So with both CPU and fUZION block lapped temps dropped 6 degrees?

[MikeMK]

I think he's saying that only his CPU was lapped, and not the fusion. Nice temp drop. Def gonna do mine.

Just puzzling over my tubing for the NB. Ive always had it my loops like this:

Res--->Pump--->rad1--->CPU--->NB--->rad 2--->GPU1--->GPU2

However, after looking at it, its pretty tight to bend the tubing from the fusion to the MCW30 - it doesnt kink the tubing, so its not THAT bad, but i can still see it having a problem with the tubing causing the MCW30 to tilt and lift one part of it away from the NB. If the mobo used screw mountings it wouldnt be an issue!

Any ideas?

[Grinch]

I think he's saying that only his CPU was lapped, and not the fusion. Nice temp drop. Def gonna do mine.

Just puzzling over my tubing for the NB. Ive always had it my loops like this:

Res--->Pump--->rad1--->CPU--->NB--->rad 2--->GPU1--->GPU2

However, after looking at it, its pretty tight to bend the tubing from the fusion to the MCW30 - it doesnt kink the tubing, so its not THAT bad, but i can still see it having a problem with the tubing causing the MCW30 to tilt and lift one part of it away from the NB. If the mobo used screw mountings it wouldnt be an issue!

Any ideas?

correct only lapped cpu..

does this give you any ideas?




being that I only have 1 video card..makes for easier tubing..

[MikeMK]

Thanks for the pic m8. Yes it definately makes it easier having the one card. U can see the prob i would have going from the NB to a GPU in the top PCI-E slot. Pretty much a no-goer.
CPU to NB works, but because the clips run accross the centre of the MCW30it allows it to move with pressure for the tube. if the clips ran accross either end of the block u wouldnt have the same problem.

Anyway, ive taken some of the pressure out of the tubes now, by rerouting them a bit. Better contact, but still not brilliant. Idling at 50degrees ish. Still better than around 75 degrees it was b4!

Talking of which, i have a feeling that the triangle on the bottom of the NB foam shim is causing some of the contact problem. Has anyone tried taking it off? Did it help?

[Grinch]

I made the MISTAKE of taking off the foam padding around the NB...DO NOT TAKE IT OFF! What happens is if your mcw30 is not PERFECTLY flat it will cause your mobo to short out and give you all sorts of stability errors...I think what happens is that the copper actually touches some of the metal on NB chip...(looks like sauter points..really dont know word)

[Raja@ASUS]

the contact with the nb itself will not be the issue, however the smt resistors and capacitors will short if the block skews under pressure. best to leave the foam in place...

regards
Raja

[MikeMK]

Cool. Thanks for the tip guys. Really not having much luck tbh. After various different mounts/tubing arrangements etc, the best temps ive had are 48-52 degrees idle (worse than stock) and worst temps around 80 degrees.

I have an EVGA sat in a box right next to me, and I know that uses mounting holes - soooo tempted to dig it out! LOL!

Such a shame, im really looking fwd to playing with the DFI, but im guessing that unless u can cool the NB sufficiently then you arent gonna get the best results, which defies the point of keeping over the EVGA.

[Grinch]

my nb is at 40C right now...

[Grinch]

After some tweaking and alot of hours of benckmarking I have found my new sweet spot for me..

[century child]

Cool. Thanks for the tip guys. Really not having much luck tbh. After various different mounts/tubing arrangements etc, the best temps ive had are 48-52 degrees idle (worse than stock) and worst temps around 80 degrees.

I have an EVGA sat in a box right next to me, and I know that uses mounting holes - soooo tempted to dig it out! LOL!

Such a shame, im really looking fwd to playing with the DFI, but im guessing that unless u can cool the NB sufficiently then you arent gonna get the best results, which defies the point of keeping over the EVGA.

You should consider one of these NB blocks. This would most likely eliminate your contact problem:
http://www.mips-computer.de/sg_WATER...PSETCOOLER.htm

[Jobeo]

ya if i had an evga laying next to this DFI and was about the build a system i wouldn't hesitate in my decision...

i'm once again frustrated with the board. i thought i had a stable 400 FSB clock as it prime blended for many hours and seemed rock solid stable after extended use.

that is of course until the SUPCOM test where you're at max units and i get reboot crashes. not even a BSOD the thing just reboots. its very easy to want to blame supcom but i can play at stock clocking for hours on end, no matter how many units are in action.

i tried setting the memory back to 2T and auto timings, that seemed to delay the reboot crash a bit but it still rebooted. all the voltages are high enough and temps are fine so the only thing i'm left with is the ******* VTT and GTL's... argghhhhhhh... =/ =/ =/

[MikeMK]

ya if i had an evga laying next to this DFI and was about the build a system i wouldn't hesitate in my decision...

Well, ive made my decision... ive opened up the EVGA - the mcw30 fits like a glove, and all is up and running installing windows now

I'll have a think over the DFI, and debate whether to grab a new NB block or not and give it another shot. I guess it depends on how happy i am with how the EVGA performs. Will see over the next few days

[Grinch]

I know my evga would not post @ 500fsb...don't get me wrong I much prefer the layout for the NB & SB on the evga..the hook thing really bothers me on the dfi..but now that I have it all under control I am taking the dfi over the evga..dfi board seems ALOT more stable for me..

[owcraftsman]

does anyone know if MIPS will ship to to USA?

Also I'm wondering how well they perform?


Source: http://www.mips-computer.de/index.htm

Dam these look sweet

[Grinch]

send an email and ask?

[BAR*B*Q]

hmmm... got my first freeze on post. how can i clear cmos without risking to kill memory ?