**Official DFI LanParty UT P35-T2R Review/Overclock/Guide Thread**

[Praz]

If all you used was scotch tape the thickness isn't going to compromise the mount. But unless the the base is shortened slightly you can't get a good mount.

[WoZZeR999]

When you say the base is shortened, do you mean the plate, or the plate holders. I did notice that I had to push the plate brackets in as far as possible to get the push-pins to line up. Is the thermal tape think enough to compensate for the raised SMD chips?

I had to play around with the thermal tape a bit to try to get it as flat as possible. I'm wondering if handling the tape (I fixed it after both side of the tape were off) is the culprit for the bad temps.

EDIT: I ran prime to check the PWM area, and the temps were still climbing when it had hit around 65-70. I shut it off because I didn't want anything to burn out. When I get home later today, I hope to have this PWM thing fixed.

[kberguy1]

If you used it as shipped, lack of cooling of the PWM slave chips is why it died. Doesn't take but an instant.

Actually, Praz, I used your suggestion and notched the tape around the raised metal caps. The thermal tape was covering the lower row of components that were closer to the rear of the motherboard. I used Arctic Cooling MX-2 thermal paste on the taller parts, as they had thermal paste originally anyway.

What I think happened is that I tightened the fasteners to a point where they made the layer of thermal tape too thin, giving a gap that the metal MOSFET cooler could short across. If anyone is going to mount these coolers with anything other than the push pins, I urge caution when tightening the fasteners. I know I learned an expensive lesson...

The good news is my DFI X38-T2R MB has arrived, although it might have to wait until next payday to get picked up. My future posts will mostly be on the DFI X38 thread. I hope to hear from a few of you there.

[runmc]

Ron,

I was experiencing the same problem yesterday with the 333/667 strap.
Timings were set at 5-5-5-5-12, the rest at AUTO.

From the moment I changed the Performance level to 7/8 or 9, I got the C1 error and it wouldn't boot at all. This was also the case with every other setting I changed manually (1 at a time)

I changed the strap to 333/800 and had no booting problems anymore. I can change every timing manually without getting that C1 error.

I'm using the Ballistix PC8000 on my UT (bios 0214).

I loaded optimized bios settings and then went back and changed all my settings except dram timing. I was then able to boot at 333/667. I had tried flashing the bios and that wasn't the problem.
As you mentioned, I also think it has something to do with performance level. I can run at 8 - 9 but not 7. I had timings set at 5-5-5-5-15 and the rest Auto. Maybe one of us will narrow it down to exactly which setting is the problem. I'm going to work with the performance level when I get time and see if that's the issue.

[Praz]

Actually, Praz, I used your suggestion and notched the tape around the raised metal caps. The thermal tape was covering the lower row of components that were closer to the rear of the motherboard. I used Arctic Cooling MX-2 thermal paste on the taller parts, as they had thermal paste originally anyway.

What I think happened is that I tightened the fasteners to a point where they made the layer of thermal tape too thin, giving a gap that the metal MOSFET cooler could short across. If anyone is going to mount these coolers with anything other than the push pins, I urge caution when tightening the fasteners. I know I learned an expensive lesson...

The good news is my DFI X38-T2R MB has arrived, although it might have to wait until next payday to get picked up. My future posts will mostly be on the DFI X38 thread. I hope to hear from a few of you there.

Although an excellent cooler, when used with these boards it is not a direct bolt-on application. Besides the component height issue and improper base length, extreme care needs to be taken when doing the actual mounting. As happened to you, a second chance will not likely be an option. When it's all said and done probably best to stick with the stock cooling solution.

[WoZZeR999]

I didn't re-mount anything last night because I didn't want to break the pushpins, but I did take some time to look if it was mounted incorrectly. Everything actually looks like it's sitting decently, and the cooler actually got hot (hotter than the stock cooler). I think the stock HS actually holds more heat, and since I am on water, I don't have direct air-flow directly on the HR-09. It seems that it sits in an area with somewhat stagnant air, so I'm thinking of either getting one of those "spot cooler" fans, or going back to the stock HS.

[GripS]

I didn't re-mount anything last night because I didn't want to break the pushpins, but I did take some time to look if it was mounted incorrectly. Everything actually looks like it's sitting decently, and the cooler actually got hot (hotter than the stock cooler). I think the stock HS actually holds more heat, and since I am on water, I don't have direct air-flow directly on the HR-09. It seems that it sits in an area with somewhat stagnant air, so I'm thinking of either getting one of those "spot cooler" fans, or going back to the stock HS.

The stock cooler is sufficient with adequate air flow. I mounted a 40mm fan on mine. The 'spot cool' cooler will probably work even better.

[mrcape]

My cooling is kicking butt. Replacing TIM, tightening the mounting of everything and adding a row of tiny mosfet sinks to the CPLs made a big difference.

Not needing tons of NB voltage helps too! The stock cooling is totally adequate if you have good airflow.

[saint-francis]

I agree. I have an 80 mm fan on my PWM sink and temps are LOW. Low like 40 under load. I took it off and replaced the thermal compound but I'm not sure how much of a difference that made because the stuff that was on there looked pretty nice (at least compared to the damn bubble gum that they usually put on).

[mrcape]

True, the stuff underneath was very evenly spread. It looked like AS5, grey, thick and good. I think just a tighten and a fan can do the trick.

[MikeMK]

Hey guys... after some of your thoughts. Ive hit some kind of wall with my CPU. I have it 100% (12hrs prime) stable at 4ghz 1.32v vcore, and it works great... but anything over that and it just wont play ball.
Even 4.1ghz wont last longer than 5 mins in prime no matter what I do. Ive tried increasing nb voltage, lossening mem timings, playing with GTL's and obviously increased vcore.

My vcore has been up to 1.42v on load so far, and even 4.1ghz still isnt stable - to me thats a hell of a jump for 100mhz extra. Ive had voltage walls before, but that is something.

VID on this chip is 1.1375v according to coretemp, which is pretty reasonable is it not? Running the latest 0214 bios. Just wondered if anyone had any thoughts on what i could try. Im pretty sure the chip should go further.

[WoZZeR999]

Have you tried throwing some higher PLL/VMCH/VTT at it?

[MikeMK]

Well, VTT is at 1.31v. I have gone higher with my other CPU's - like my G0 Q6600 - but i read somewhere that its not a good idea to put too much VTT through these chips, so i was a bit hesitant.

I have been working with GTL's around 85/85/85 which is 66% of VTT. NB chipset has been up to 1.55v, i have a fan on it, so temps are ok but imo it shouldnt really need higher than that for just 200mhz more. (at 4ghz i use 1.47v nb)

I'll play around with the PLL voltage and see if that makes any difference as i havnt tried that yet.

[cantankerous]

I would be weary of using excess PLL voltage as well. Too much of that helps, but is only short lived. After that the cpu is permanently dwarfed on how high the fsb can go. Too high a PLL voltage will surely have the FSB limit much lower than it was to begin with. I understand we are all greedy for every little bit of mhz we can get but I would much rather sacrifice 200mhz now then 400mhz a month from now due to a weakened chip. 4ghz is complete overkill for anything needed these days.

[MikeMK]

Very wise words there! However... theres always that nagging in your mind to go higher Especially seeing as my vcore is relatively low at 4ghz! Im happy with 4ghz for 24/7 but its always nice to push things further.

[Brother Esau]

Hello One and All I'm finally back online Did everybody that ordered a Bios Chip from me before my move get their Bios Chip alright?

Campbell

[ether.real]

Yup, got mine

[ababich]

Hello One and All I'm finally back online Did everybody that ordered a Bios Chip from me before my move get their Bios Chip alright?

Campbell

I sure did, thanks Cambell

[Brother Esau]

Well just got online after a month last night and other then some thank you PM in regards to the Bios chips I guess no news is good news But I just want to check in and make sure everybody is a Happy Camper

[Jupiler]

Well, VTT is at 1.31v. I have gone higher with my other CPU's - like my G0 Q6600 - but i read somewhere that its not a good idea to put too much VTT through these chips, so i was a bit hesitant.

I have been working with GTL's around 85/85/85 which is 66% of VTT. NB chipset has been up to 1.55v, i have a fan on it, so temps are ok but imo it shouldnt really need higher than that for just 200mhz more. (at 4ghz i use 1.47v nb)

I'll play around with the PLL voltage and see if that makes any difference as i havnt tried that yet.

Here's a nice article regarding the GTL settings with 45nm cpu's.
The guide is written by Pete_X (also member here at XS) and in German.
Here's the google translation to English :

http://translate.google.com/translat...hl=nl&ie=UTF-8

[Praz]

Here's a nice article regarding the GTL settings with 45nm cpu's.
The guide is written by Pete_X (also member here at XS) and in German.
Here's the google translation to English :

http://translate.google.com/translat...hl=nl&ie=UTF-8

The part about VTT is pretty much dead on. However the method of adjusting the GTL values is hit-n-miss. In each example where the GTLREF was changed, all three were adjusted (CPU 0/2, CPU1/3 and NB). Using that method if one or two happen to be correct when there is a failure and all three are adjusted a person is really not getting anywhere.

Campbell, nice to see you back.

[Brother Esau]

Hey, John whats up and hows the family?

[MikeMK]

Here's a nice article regarding the GTL settings with 45nm cpu's.
The guide is written by Pete_X (also member here at XS) and in German.
Here's the google translation to English :

http://translate.google.com/translat...hl=nl&ie=UTF-8

Thanks for posting that... very interesting indeed. Looks like its all in my GTL's i guess. I take it that explains why no matter how high the vcore is it wont go above 4ghz?

Lots more tweaking to do!

[Eldonko]

3110 580 FSB benchable on T2RS, better than a kick in the sack I guess. All about the GTL settings.

[cantankerous]

Just wondering what you guys are running your PCIE bus at for 24/7 use? I am at 105mhz very comfortably and have been like this since day one. I have tried going to 110mhz but things just seemed to get flakey. Games would crash to desktop etc until I lowered it back down to 105mhz. This sounds a bit low to me, so would like to know what others are achieving for 24/7 use with 100% stability.