DELL XPS 420 VID-Vcore-Vdroop pencil volt mod Q6600 3.1 @ 1.25 v

[XeN0s]

Has anyone gotten their Dell XPS 420 Q6600 at beyond 3.1/3.2 GHz, and how did you do the volt mod?

I reach stability issues at beyond 3.1 GHz, when I approach 350 FSB speeds, x9 mult. I use RMClock to disable EIST and C1E and run Max performance.

SetFSB brings me from 2.99 to 3.11 GHz on 2-PAG maybe I should try the PAG.

Memset doesn't work because my MCHBAR is locked, but I'm at a 5:6 FSBRAM ratio.

4x1 GB 5-5-5-18 DDR2 RAM PC6400, running @ about 420 MHz, (5.9 WEI, noticed no difference in my overclockability with GSkill PC6400 5-5-5-15 3x2GB WEI 7.3)

PSU is Antec 750 watt modular.

Tape modded, wanting to increase volts for stability, temps aren't issue atm, just can't clock more.

Any tips appreciated.




I've hit an OC wall here it seems. I remember seeing a thread about the Q6600 volt mod but it seemed inconclusive. Talked about the G33 Intel motherboard I believe, and I have Dell stuff with not the same thing in picture to pencil - I can't find the thing in the photo mentioned to adjust vdroop with graphite from pencil.

Would I be OK to just increase volts without messing with vdroop if my VID is stable at 1.25 v via RMClock? I could try connecting some pins on the top left of my LGA 775 chip, should bring to 1.35 volts, but I can't find the vdroop on the Dell X38 motherboard.





Dell XPS 420 Q6600 @ 3.1 GHz 1333 FSB Tape [Max 46 C Air]

 

[XeN0s]

Apparently with the Core 2 series and later vdroop does not need to be disabled.

So the only variable left is my VID.

This was my guide, these mods apply to all LGA 775 chips:

http://www.overclock.net/intel-cpus/341123-intel-bsel-vid-mod-guide.html

I've done both of these mods, the FSB tape mod first and have run that stable to 3.2 GHZ stock voltage 1.25 for months now.



I just did the VID 1.45 mod with silver ink pen and more electrical tape, maybe I didn't put enough ink on, it is really thin even though it doesn't appear so.

should equal



RMClock, CPU-Z, Core Temp, and Everest all report 1.25v, that must be accurate, which means it did nothing for me.

I wonder if I should try doing the 1.4v mod instead, or thickening the ink/making it more precise by blocking with tape before inking.



Or maybe I should do this which someone did to their Q6600 so it should be a more reliable test if there is another factor. It's also a less complicated mod, should be 1.31 v, found here.

 

[TheOneTrueSplit]

XPS 420 Q6600 Overclock

I too have done the "tape mod" to get my Dell XPS 420 (UK spec) from 2.4GHz to 3.0GHz and then SetFSB to get it up to 3.2GHz, anything more than that and it gets unstable (as in I get BSDs).

I have got it up a bit more and run 3DMark etc and it was OK but when I run apps like FSX or Maxon Cinema 4D I get random BSDs.

I have not been as brave as yourself and done the voltage mods.

I think we have reached the limits for the XPS 420 which overall is not bad is it!

I have an XPS 8300 with an I7 Sandy Bridge on its way from Dell, somehow I managed to get the XPS 8300 with I7 2600, 12Gb of memory and a 1Tb hard disk for £760!!!! This was by using different "customise" links on Dell's UK XPS page, I should be paying more than £400 more for this rig!!!!! I know its not the K model Sandy Bridge chip but I hope to get some good results from both FSX and Cinema 4D (my main two apps).

 

[TheOneTrueSplit]

PS...

I would like to see some screen shots of your RMClock and SetFSB settings.

I am using the older "free" versions and would be happy to post images of my settings for comparison.

 

[XeN0s]

At some point I will redo the VID mod and try 1.4 or 1.35, I suspect I may have made the silver ink imprecise and maybe an extra pin is being connected, blocking off with tape should help.

I did just install 4x2GB DDR2 @ 800 MHz 5-5-5-15 G Skill RAM (noticed no OC increase with 1 of 4 sticks installed, or 2 of 4 in dual channel): http://www.newegg.com/Product/Product.aspx?Item=N82E16820231122

I use FPU Julia test from Everest as the most intensive CPU OC stability test even though it doesnt run longer than 10 seconds; it seems that if that test causes an Everest crash or a blue screen then the games will crash as well. 10 MHZ slower than a crashing FPU Julia test seems to always be 10 MHz slower than a BF3 crash though it might take 20 minutes in game, 10 MHz lower is perfectly stable, so that benchmark is perfect to tell me my stability.

It made my overclock a little more stable at slightly higher frequency (previous stability 3.096 GHz, now 3.142 GHz definite increase reflected by FPU benchmarks in Everest, although games like BF3 and SW:TOR which were not available when I was overclocking previously do not run stable at my previous 3.1 GHz overclock (w/ mixed RAM) which was stable for all other games.

I had to downclock to the default 3 GHz in order to run stable in BF3 and SW:TOR with old RAM (3 different timings, 2 different frequencies bad RAM mix).

Since I have solid G Skill RAM now I can run BF3 and SW:TOR (and FPU tests) stable at 3.14 GHz.

I am working on trying to downclock the RAM to be 1:1 with my FSB (FSB 333, DRAM 400 5:6), I want to try 333 with different timings and see if that can help me overclock more on default VID.

Been using SPDTool, MCHBAR is locked on XPS 420 X38 chipset so I wonder if that limits editing RAM timings like with SPDTool; I am able to change JEDEC profile #2 (I have a #1 and #2 in my G Skills, as CPU-Z shows) from 400 MHz to 333 MHz but CPU-Z still tells me the RAM is running at 400 MHz. The only value I am changing is "SDRAM Cycle time at Maximum Supported CAS Latency" from 2.5 ns 400 MHz to 3 ns 333 MHz.

I am going to try Thaiphoon next.

My only limits now seem to be RAM clock/timings (if my RAM is limiting the OC) or VID.



The 700 MHz DDR frequency is inaccurate; all other programs tell me at 350 FSB my RAM is running at 418 MHz, effectively 835 MHz, not 700.

I didn't adjust any other settings from default except on these pages in RMClock: