Official DFI LanParty UT X38-T2R(and LT) Discussion/Review/Overclock/Guide Thread

[Aerou]

There are always an exception to everything but for most users changing the TIM used for the PWM MOSFETs is not going to result in a decrease of temperatures. DFI uses Shinetsu X23-7783D on these components. It is probably the best commonly available TIM for this use.

and what paste does DFI use on the SB HS?

[Praz]

and what paste does DFI use on the SB HS?

I've never asked. But because of the cost of X23-7783D I doubt if it's used where not needed.

[eva2000]

so why is tape being advised for mosfets i used mx-2 and they dropped to +3 ambient idle from +8 and i have more volts in them now, so why use tape with the semi hard mount the dfi dosnt have a gap from the push pins

same here used MX-2 on my DFI X38 LT on SB/NB and PWM area and really nice low temps under load

this is the stock stuff (Shinetsu X23-7783D ?) which i removed - pretty well applied on PWM area



if it's Shinetsu X23-7783D, then its consistency is pretty much same as MX-2 TIM.

[Origin_Unknown]

if only one of the thermalrite heatsinks would fit there

[zanzabar]

Ok 2 questions... Why is it reading almost 1.7v to the northbridge when i set the voltage at like 1.444v?

u need to update the bios its a glitch that they fixed in the voltage tables



and is that paste that came with the beard x23 i just left it in the box since i thought that it was random stuff that would be worse than mx2

[t_ski]

This looks like there is very little pressure on the mosfets. It looks like only the two on the edge are barley making contact. I wonder of some nylon screws and nuts might allow you to tighten it down a little better. Also, if the heatsink is aluminum, perhaps it's bent? Maybe not flat, but higher up in the middle...

By the way, what kind of watercooling can fit the NB? I have a Danger Den MPC-975X chipset block sitting next to me - can someone check the center-to-center distance of the two NB mounting push-pins? I can check the DD block and see if it would work on this board. I'd rather use a chipset wb on it than the huge stock heatsink...

[Praz]

The MCW30 fits if that's any help.

[t_ski]

Have one of those, too

[Praz]

Forgot I had one sitting here with no heatsink on it. 60mm center to center

[zanzabar]

what about those chips with cpl on them they get kinda hot if some1 was to make something shouldent that be covered aswell

[jakk]

Hi. I have oc'd before,but I have never had this good of a board to oc on. I am currently at 3.4 stable with my g0 q6600. I'm having a problem getting it to 3.6, and I was wondering if I could get some input as to what my problem might be. My first question is how everyone posts those nice bios templates,I'm not sure how to do that.
I currently have my cpu vtt set at 1.227v,I'm not sure what a safe limit is for it, and I sure don't want to fry my q6600! I'm running 2x1gb of patriot ddr800 at 2.23v. When my fsb is at 400,I can run ddr800 so I don't think ram is my problem,but I'm not sure. If I try to boot at 400,I bsod at windows loading. I'm guessing it's a voltage issue but,like I said,I don't want to fry anything. I'm on a d-tek fuzion and a pa 120.2 for temp control. Thanks for your help.

[Praz]

My first question is how everyone posts those nice bios templates,I'm not sure how to do that.

Just substitute your values.

[jakk]

Ok,here is what I am at. Everything marked 'stock' is board default.

CPU Feature Page
Thermal Management Control................Enabled
PPM (EIST) Mode...........................Disabled
Limit CPUID MaxVal........................Disabled
CIE Function..............................Auto
Execute Disable Bit.......................Enabled
Virtualization Technology.................Enabled
Core Multi-Processing.....................Enabled

Main BIOS Page
Exist Setup Shutdown......................Mode 2
Shutdown After AC Loss....................Enabled
O. C. Fail Retry Counter..................1
CLOCK VC0 Divider.........................Auto
CPU Clock Ratio...........................9x
CPU Clock.................................380 MHz
Boot Up Clock.............................265
DRAM Speed................................333/800
PCIE Clock................................100 MHz
PCIE Slot Config..........................1X 1X

CPU Spread Spectrum.......................Disabled
PCIE Spread Spectrum......................Disabled
SATA Spread Spectrum......................Disabled

Voltage Setting Page 
CPU VID Control...........................stock
CPU VID Special Add.......................Auto
DRAM Voltage Control......................2.230V
SB Core/CPU PLL Voltage...................1.510V
NB Core Voltage...........................1.265V
CPU VTT Voltage...........................1.222V
VCore Droop Control.......................Enabled
Clockgen Voltage Control..................3.45V
GTL+ Buffers Strength.....................Strong
Host Slew Rate............................Weak
GTL REF Voltage Control...................Disabled
CPU GTL 1/3 REF Volt......................N/A
CPU GTL 0/2 REF Volt......................N/A
North Bridge GTL REF Volt ................N/A

DRAM Timing Page
Enhance Data Transmitting.................Fast
Enhance Addressing........................Fast
T2 Dispatch...............................Enabled
Clock Setting Fine Delay..................Listed Below
CAS Latency Time (tCL)....................5
RAS# to CAS# Delay (tRCD).................5
RAS# Precharge (tRP)......................5
Precharge Delay (tRAS)....................12
All Precharge to Act......................auto
REF to ACT Delay (tRFC)...................auto
Performance Level.........................auto
Read Delay Phase Adjust...................Listed Below
MCH ODT Latency...........................auto
Write to PRE Delay (tWR)..................auto
Rank Write to Read (tWTR).................auto
ACT to ACT Delay (tRRD)...................auto
Read to Write Delay (tRDWR)...............auto
Ranks Write to Write (tWRWR)..............auto
Ranks Read to Read (tRDRD)................auto
Ranks Write to Read (tWRRD)...............auto
Read CAS# Precharge (tRTP)................auto
ALL PRE to Refresh........................auto

Read Delay Phase Adjust Page
Channel 1 Phase 0 Pull-In.................Enabled
Channel 1 Phase 1 Pull-In.................Enabled
Channel 1 Phase 2 Pull-In.................Enabled
Channel 1 Phase 3 Pull-In.................Enabled
Channel 1 Phase 4 Pull-In.................Enabled

Channel 2 Phase 0 Pull-In.................Enabled
Channel 2 Phase 1 Pull-In.................Auto
Channel 2 Phase 2 Pull-In.................Enabled
Channel 2 Phase 3 Pull-In.................Enabled
Channel 2 Phase 4 Pull-In.................Enabled

Clock Setting Fine Delay Page
Ch1 Clock Crossing Setting................stock
DIMM 1 Clock fine delay...................stock
DIMM 2 Clock fine delay...................stock
DIMM 1 Control fine delay.................stock
DIMM 2 Control fine delay.................stock
Ch 1 Command fine delay...................stock

Ch2 Clock Crossing Setting................stock
DIMM 3 Clock fine delay...................stock
DIMM 4 Clock fine delay...................stock
DIMM 3 Control fine delay.................stock
DIMM 4 Control fine delay.................stock
Ch 2 Command fine delay...................stock

Ch1Ch2 CommonClock Setting................stock

Ch1 RDCAS GNT-Chip Delay..................stock
Ch1 WRCAS GNT-Chip Delay..................stock
Ch1 Command to CS Delay...................stock

Ch2 RDCAS GNT-Chip Delay..................stock
Ch2 WRCAS GNT-Chip Delay..................stock
Ch2 Command to CS Delay...................stock

[crazyea]

STABLE Q6600 G0 @3.825 - 1.5v on my DFI LanParty LT X38-T2R.

Love this new set up!

59c Loaded on all 4 cores w/ my new watercooling set up! 65 in SG is due to my setting cpu offset too high in bios, everest is showing the cores properly. Idle temps are very good as well! Click images for lager pic.


Idle


Load


My New WC Set up!

[crazyea]

The MCW30 fits if that's any help.

yes, I am using it. Highly recommended!

[linflas]

Nice clocking, and nice clean rig, well done.

[t_ski]

Forgot I had one sitting here with no heatsink on it. 60mm center to center

Ah, then - the DD 975X block is too far apart. I'll have to see if they have another top that might work with it, or else I'll use the Swifetch one. The base is larger on the DD block, so I was thinking it would sit better. Add in the pins on the DD block and I think it may cool even better. I have yet to test it though.

[crazyea]

Nice clocking, and nice clean rig, well done.

Thanks!

[C-N]

Hi all, I don't actually have this x38 board but I do have the p35 with a similar layout so the issue & remedy should transpose.

I had to tighten my PWM heatsink & in-fact my whole heatpipe up twice on my p35 board probably due to the thermal compound sinking or squishing out with heat expansion etc.

One thing I noticed which doesn't help much is when mounting my TRUE 120 even though the thing is still not mounted solid (you can still slide it around) the board warps/buckles probably due to the tension on the mount & the weight of the TRUE 120. I am now on water using a D-Tek block & it is still easily possible to warp/buckle the board before the mounting screws bottom on there spacers.

The thing I have noted from these 2 mounts air/water is that if you over tighten the CPU mount & warp/buckle the board the PWM temps also increase. If you get bad warping enough to separate the PWM from the heat sink along its very long 8 phases this is enough to cause a PWM cascade failure resulting in a reboot.

I now have a very light water block mount, my board is straight & my PWM runs in the low 40's Orthosing with my Q6600 @ 480 X 8.

So to sum it up all you having PWM issues/high temps check your board is not warping around the CPU/PWM area.

CN

[crazyea]

so if the board warps a bit, do you just back it of until its not warped anymore or are you using a backplate with your block?

[nzbleach]

that would explain it C-N. Good, logical thinking there.

[C-N]

so if the board warps a bit, do you just back it of until its not warped anymore or are you using a backplate with your block?

When I noticed the PWM temps rise & the board started rebooting with the TRUE I had just put it in the case after running Orthos overnight @ 475 x 8 on my Tek station. I then took the board back out & looked all over it & found the TRUE was buckling the board. I guess when the TRUE was pushing down on the board it was fine but once I fitted it in the case & the weight of the true levered down on the board the warping must have been too much for the PWM to handle hence the temp & reboot problem.

I messed about for hours re-mounting the TRUE in the case but I could not get a decent mount where all my cores were running even so I switched over to water.

Initially I tightened the D-Tek block up to the stops & it too warped the board so I added a backplate from my old Tunic tower 120 & screwed the D-Tek screws into it, I measures & marked the screws 8mm from the end in an attempt to get all 4 screw at the same pressure. Basically now I only have the 4 D-Tek mounting nipped up literally just taking up the slack on the springs + 2 turns & all my cores run within 2oC of each other under load.

that would explain it C-N. Good, logical thinking there.

Thanks nzbleach but it was just what I found maybe it can help others maybe not.

IMO we have all probably had many boards warp around the CPU area & it hasn't mattered with 4,5 maybe even 6 Phase PWM's because they were short enough to not be affected by board warp, but these new 8 phase really long PWM's cant take the warping.

CN

[luihed]

Very good point C-N, had the same problem with my P35T2R especially with my QX... My mounting is very diff than norm and this was my solution which helped stability on the QX....



It solved 90% of the issues I was having with the QX but unfortunately not all of them... DFI taiwan was kind enough to upgrade me to an X38 ( thanks Jimmy )...

The X38 however doesnt seem to be too affected when the board gets a lil warped, dont know why cos its pretty much identical to the P35 ...

I hate to say this but Im not too impressed with this board in terms of fsb clocking... My P35 got my E6600 to 600fsb, from the short test that I did, I couldnt get past 570 with this board, I really hope its just a setting thing.... This board however runs my QX so smooth and I hit wayyy higher fsb on my QX vs the P35, hopefully I get some decent fsb when I get a wolfie....

As for the TIM discussion, Thermal conductivity in (w/(m-k))... copper=401 aluminum=237 lead=36 water=.6 Thermal grease .7-3 ( silver compound 2-3 ) .... does it really matter which tim you use? my $.02

[BulldogPO]

Any news about DDR3 version?

[t_ski]

Initially I tightened the D-Tek block up to the stops & it too warped the board so I added a backplate from my old Tunic tower 120 & screwed the D-Tek screws into it, I measures & marked the screws 8mm from the end in an attempt to get all 4 screw at the same pressure. Basically now I only have the 4 D-Tek mounting nipped up literally just taking up the slack on the springs + 2 turns & all my cores run within 2oC of each other under load.

So can we assume you are using the default mounting for the D-tek, and not the optional Pro mounting kit that includes a backplate? Do you think the backplate would make a difference if you are not using one right now?