Overclock.net banner
2M views 13K replies 1K participants last post by  austin86 
#1 ·

Hi everyone! I have been following a thread started a while ago about an awesome mod that is super simple and gives new life to the LGA775 platform. The original thread has become cluttered and the original poster hasn't updated his posts so I figured I would put together all of the findings into one thread to help any one interested. If you are interested in this mod, this is the place to find all the information you need about it!

Overview

This mod is for LGA775 motherboards only! It allows the use of LGA771 Xeon CPU's that are very affordable and powerful in many LGA775 motherboards with a few small modifications to the motherboard and CPU.

How does it work?

We are using a small sticker to flip two pins on the bottom of the CPU, then rotating the CPU to the correct position on in a LGA775 motherboard. If you want a more in-depth explanation, check this thread out: Explanation

Why would I want to do this?

Many people simply do not have the money to dump their old rig and buy a new one. The LGA775 platform has began to show its age in recent games and programs with most consumer processors. LGA771 Xeon processors are very affordable and blow most consumer level LGA775 cpu's out of the water. For example, a Xeon X5460 can be had for around $40. For $40 you can get a quad core CPU with 12mb of cache that runs at 3.16GHz stock. You also must remember that these CPU's were meant for servers, which means they are binned higher than the consumer level or even the extreme edition LGA775 CPU's. This usually equates to better overclocks and lower operating temperatures.

Compatibility

The following boards/chipsets have been proven compatible:

https://docs.google.com/spreadsheet/pub?key=0ArP20vNBl1vSdHUxTTFLMEdmejU3Vi1WaVVqLVF4Vmc&single=true&gid=0&output=html&widget=true

Choosing the right motherboard for this mod is extremely important. Some boards need a simple microcode update while others will not post with the Xeon CPU's at all. I will attempt to create the most complete compatibility chart I can, but also take a look at post #3 and see if your board is listed.
If you are running a motherboard with the X38, X48, Q35, or Q45 chipset there is a high chance it may not be compatible with this mod!

The Mod

Before you modify your motherboard, update your bios to the latest version if you haven't already.

Disclaimer: I am not responsible for any damage you may do to your motherboard or CPU while performing this mod!

First gather the necessary tools and components:

A sharp knife: Xacto, box cutter, or anything similar
A LGA775 motherboard with a compatible Chipset
A LGA771 Xeon CPU
A "Pin Mod" Sticker Link 1
Optional:
A pair of tweezers
A well lit room or a lamp to light up your work space

Lets start with the motherboard. First located the two CPU guides in the CPU socket:


Now use your knife and cut them both off



Once you do that you are done with half the mod!

Now get your CPU and "Pin Mod" stickers, tweezers come in handy when applying the stickers



And stick the sticker on with the little triangle facing the triangle on the bottom of the CPU



Now make sure when placing the CPU into the socket you line up the triangle on the CPU and the triangle on the socket!



You did it! Now put your system back together and get to overclocking!

Conclusion

If you made it this far, you either completed the mod or have read through the thread and are still unsure about it. If you have completed the mod, make a post sharing your success, it will help get others to try the mod for themselves. If you still aren't unsure about it, have a look at some of the others who have done completed the mod.
Quote:
Originally Posted by Servos View Post

Weirdest/best upgrade deal I've ever made. Sold my q6700 enough to buy a e5440, the adapter, some mayhems dye (for main computer). And hands down the e5440>q6700

Well the evga 780i seems to play nice with the E5440.
I oc'ed it to 4.0ghz, but my vdroop was pretty bad. Didn't crash, but ibt faild after 2-3 tests.
So stuck it at 3.2 for 1:1 fsb:ram.

The crasiest part is the temps. I let this go for about an hour and nothing hits over 55*c (even at 4.0ghz). I thought maybe it realtemp was misreading the sensor, but I felt the exhaust and it wasn't particularly over-wam. My q6700 was hotter under load stock.

Didn't realize I was using old bios too, off to flash em.



bumped it up to 3.4ghz, somehow I had the multi at 8 instead of 8.5.
Quote:
Originally Posted by imi2003 View Post

Hi all
I just successfully used a Xenon E5430 to work with a Gigabyte G41MT-S2P motherboard using a pin mod sticker and modding the socket
There is only one problem since my E5430 is a E0 stepping it has disabled sse 4.1 etc.
Does any one have a modded bios for this motherboard or can any one point me in the direction of some sort of tutorial so I can have a go myself
oh and this motherboard is pants for overclocking wont go over 335fsb without locking up on boot


Thanks imi2003
Quote:
Originally Posted by jrad View Post

Awesome news i fortunately got lucky and my X5460 posted right away with bios P32 on my Evga 680i
biggrin.gif
wheee.gif
, I'm going to install it on my 780i since it overclocks better. I wasnt expecting it to work though on the 680i since it didnt officially support 45nm Yorkfield quad cores.

LL


LL


LL


900x900px-LL-322f2b54_two.jpeg


900x900px-LL-e2e65b1d_five.jpeg


Going to start testing overclocks with it next to see how high it can go, i wonder if it will overclock as nice as the Q6600s.

Thanks so much to all who contributed to this thread
cheers.gif


Edit: alright just got it up and running on my XFX 780i also, running bios P09.
LL


I think this engineering sample X5460 might have an unlocked multi because its letting me select any multiplier i want
ph34r-smiley.gif
Quote:
Originally Posted by TB13 View Post

So, I posted on bios-mods.com earlier today and never got a reply. Being incredibly impatient and bored I decided to try and mod the BIOS myself, I'll let the picture do the rest of the talking...
rolleyes.gif




So ladies and gents, if anyone here needs a BIOS modded to support Xeons or other 775 CPU's let me know and I can do it.

I'll be back later with some overclocking results
biggrin.gif
Quote:
Originally Posted by TB13 View Post

Bump for this thread!

These guys showed up this afternoon and I got to installing them




They are somewhat like the ones OP had, but they do not need to be cut and have adhesive pre applied. Peel the adhesive, mod the socket, and drop it in. Bam, quad core awesomeness!



Booted up no problem on my Foxconn G31 board.



Its running a bit warm, but I will pick up a new heatsink sometime soon and hopefully overclock it a bit.
 
See less See more
7
#2,152 ·
Quote:
Originally Posted by soco View Post

Which program are you using to monitor your temperature? Are you adjusting tjmax?

Credit goes to tomshardware -- http://www.tomshardware.co.uk/intel-dts-specs,news-29460.html

45nm Xeon Quad-Core CPUs
Model....................................................Tj Max (Stepping unspecified)
E7440, E7430, E7420.......................90°C
L7445...................................................80°C
X54xx series....................................... 85°C
E54xx series........................................85°C
L5408...................................................95°C
L5430, L5420, L5410........................70°C
X33xx....................................................95°C
L3360...................................................90°C
X33xx series........................................95°C
L3360...................................................90°C
I am just using the AIDA64 Extreme stress testing for the CPU, and noticed the built in monitoring shoot to the moon on temps. I will look at the TJ MAX setting and see if it helps. That was something else I thought about if the program was reading the temps correctly or not.
 
#2,153 ·
Quote:
Originally Posted by furion223 View Post

Quote:
Originally Posted by pioneerisloud View Post

Try a different sticker. I had that problem on my EP43-UD3L. Changed the sticker, and it works great now.
Got 2 xeons, both do the same, could be some faulty stickers?? got 2 stickers from some guy in israel, they got a "T" written on them or sth like that. When i mounted the second sticker on the second CPU it worked, as soon as i assembled my unit and tried to start it, it would go back to the same reboot loop. Anyway to rule this out, got any store that sells some stickers, maybe some you know work, or just random ebay ???

One more question, could someone also recommend me a good enough cooler for this CPU?? Been looking at something in the likes of "Cooler Master Hyper 212 EVO" would this work or is there anything better for that kind of money, around 50$
All I know is that I had that same problem with my P43, and a new sticker fixed it. I had two different kinds of stickers. The one you have to cut, and the one that you don't have to cut. The one that doesn't require cutting is what's been working best for me.

For a $50 cooler, I'd try to snag an H80 or H100. Even used. I'm sure you can find an H100 or H100i used for $50.
 
#2,154 ·
Quote:
Originally Posted by x11nt4 View Post

About the rewire job, I think I will do this. Probably find some thin wire of some sort to solder across the terminals along the back side. I would like to see if my P5E would work with this. I just don't want to get into cutting traces. I hope that wont be needed.
There's no cutting or soldering required. To prove my theory or disprove it only requires a conductive paint pen (easiest) or foil and adhesive. I really recommend against solder for this until it is at least proven. Soldering leaves no room for error and greatly increases the chances of something going wrong. I will show the first part of the mod (row AN) for starters. If it doesn't work the we can move to part two (row A).

Red lines in following image are just for reference.
As you can see the land in position AN 28 is only about 55% there. Hopefully this wont be a problem. There is always the possibility of filling in the groove with resin before attempting the mod, but I will leave that up to you. The resin would need to be completely flush to allow the CPU to seat properly.


A few multimeter readings must be done before the mod to ensure things for a smoothly as possible. Some of the lands are voltage others are return. If they are shorted together, your board will likely never boot again. The readings before the mod must match the readings after the mod or there is something wrong. Let me know which method you would like to try and we can begin. A windshield defroster repair kit may be the easiest and can be had at most auto parts stores. The foil method will be more tedious given how small the area is. Let me know what you would like to try and we can get started.
 
#2,155 ·
Quote:
Originally Posted by furion223 View Post

Hey there, can anyone recommend me a mobo that will surely work with this mod?? been trying to get it to work with my asus p5b-se, but had little success, my computer would restart when reaching the "Starting Windows" screen. Only got it to boot once but then when i restarted it came back to the same problem. Could the temperature be the problem, currently got a stock cooler that came with my core 2 duo e7200.

Anyhow, any mobo's that will surely work with this?? even better the ones that have DDR3 support.
Well my ga-EP45T-UD3LR worked out of the box with this. I did a microcode update to enable SSE4 and VT-x. Running PC2400 DDR3 @ 1666MHz 500 Bus right now.

Here are a couple videos I did for my youtube channel with the E5430 I was using at the time when I started this mod before i did the microcode update. Now using a E5450...



 
#2,156 ·
hi everyone!
i've been lurking this thread for a while and ordered e5430 and few dosen of adapters for my pq5..

but i have another, maybe interesting, question..
the other day i got this motherboard
http://www.asrock.com/mb/ATI/775Twins-HDTV/?cat=CPU
and Presler is best thing that this board can push so my question is.. could i put Dempsey Xeon (771 as well) on this board?
http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Dempsey.22_.2865_nm.29
what are the odds that this stone age chipset (ati xpress 200) will support it?

dempsey xeons are really dirty cheap on ebay, for 15 bucks i could get e5080 which is singing on 3.7ghz..
 
#2,157 ·
Quote:
Originally Posted by x11nt4 View Post

Ya I had this problem with my EP45T-UD3LR board. without knowing what memory you are running I can't tell you for sure, but mine has this issue with not enough voltage to the memory, and not enough to the MCH and MCH reference (tricky that reference voltage stuff! )

This post helped me get up to 500MHz bus speed. Hope this helps. - http://hardforum.com/showthread.php?t=1331808
For you, not having enough voltage to RAM caused the video problems too? I didn't overclock at all and left everything on auto so I figured I did something wrong but it's been "working" normally other than the video problems and other stuff I described. The video problem was definitely not happening before I put my E5450 in and was still using my Q6600 so I thought I did something wrong.

I'm not sure, but maybe my video problem is completely separate.
 
#2,158 ·
Quote:
Originally Posted by nixass View Post

hi everyone!
i've been lurking this thread for a while and ordered e5430 and few dosen of adapters for my pq5..

but i have another, maybe interesting, question..
the other day i got this motherboard
http://www.asrock.com/mb/ATI/775Twins-HDTV/?cat=CPU
and Presler is best thing that this board can push so my question is.. could i put Dempsey Xeon (771 as well) on this board?
http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Dempsey.22_.2865_nm.29
what are the odds that this stone age chipset (ati xpress 200) will support it?

dempsey xeons are really dirty cheap on ebay, for 15 bucks i could get e5080 which is singing on 3.7ghz..
It wouldn't be worth the effort. 95W and yet they're only a dual core. You're better off staying with the e5430 and your pq5.
 
#2,160 ·
Quote:
Originally Posted by TerminalVoltage View Post

There's no cutting or soldering required. To prove my theory or disprove it only requires a conductive paint pen (easiest) or foil and adhesive. I really recommend against solder for this until it is at least proven. Soldering leaves no room for error and greatly increases the chances of something going wrong. I will show the first part of the mod (row AN) for starters. If it doesn't work the we can move to part two (row A).

Red lines in following image are just for reference.
As you can see the land in position AN 28 is only about 55% there. Hopefully this wont be a problem. There is always the possibility of filling in the groove with resin before attempting the mod, but I will leave that up to you. The resin would need to be completely flush to allow the CPU to seat properly.


A few multimeter readings must be done before the mod to ensure things for a smoothly as possible. Some of the lands are voltage others are return. If they are shorted together, your board will likely never boot again. The readings before the mod must match the readings after the mod or there is something wrong. Let me know which method you would like to try and we can begin. A windshield defroster repair kit may be the easiest and can be had at most auto parts stores. The foil method will be more tedious given how small the area is. Let me know what you would like to try and we can get started.
For some reason I was thinking the pins on the socket came through to the back of the MB, which they dont. so no I would not solder. If I am understanding your theory here, are you wanting to create a link with the conductive pen to reconnect the signal from the cpu lands that are currently not connected, back to the correct pins on the socket? I have an LGA771 MB I can check voltages from these pins too.

I think your theory is a good one because I felt when I tried this, the bios was looking for a condition from the CPU that it was not seeing, thus it just said NO to a startup. But I am wondering, if this is an ECC related thing, could the bios be modded to not check for ECC capabilities from the CPU? I thought too, but it has been more than a month since I started this modding, that I had read the gigabyte X38 boards did work and it was only the ASUS that had this problem. Is that not the case ?
 
#2,161 ·
Quote:
Originally Posted by x11nt4 View Post

For some reason I was thinking the pins on the socket came through to the back of the MB, which they dont. so no I would not solder. If I am understanding your theory here, are you wanting to create a link with the conductive pen to reconnect the signal from the cpu lands that are currently not connected, back to the correct pins on the socket? I have an LGA771 MB I can check voltages from these pins too.

I think your theory is a good one because I felt when I tried this, the bios was looking for a condition from the CPU that it was not seeing, thus it just said NO to a startup. But I am wondering, if this is an ECC related thing, could the bios be modded to not check for ECC capabilities from the CPU? I thought too, but it has been more than a month since I started this modding, that I had read the gigabyte X38 boards did work and it was only the ASUS that had this problem. Is that not the case ?
Voltage readings in this case won't do any good there is too much circuitry involved. I do think we're on the same page here. We will basically be creating a new set of lands with the pain pen using the existing lands I have mapped out. When done the previously unused pins will make a complete circuit and the motherboard will hopefully boot. Conductive paint pin may not be the most rigorous media to use, but as long as you're not continuously removing the CPU it should be fine. It should be good enough to prove whether on not this will work. You'll need decent masking or electrical tape too. It can't allow the paint to get under it while drying. If you have the board out, look at the socket with the CPU in it and see if it looks like the pins in rows A 28 and AN 28 are making contact with the CPU. You'll probably need a magnifying glass or take a macro shot with a camera to be certain.
I know I'm not sharing all my info for this mod yet, but I want to be certain that all the steps are understood and done in order before beginning. The last person I spoke to about doing the mod up and fried their CPU before I even said what to do lol. Patients is the key here. I work on state of the art avionics so I'm confident in my methods. I have a two part goal here:
1. Get the mod to work in X series chipsets.
2. Don't fry the motherboard...

The only board I know of that worked was a gigabyte X38 series with a 33XX series Xeon CPU. All others I know of have failed no matter the brand. The issue is the POST so I doubt modding the bios will do any good as this is before the boot up sequence. Modding the BIOS in this fashion will likely create a board that never boots...
 
#2,162 ·
This is going to be a dumb question.

My x5470 showed up today. I'm going to install it later, but presuming that I can boot, what is the best program to be using to monitor temps and is there anything I need to modify to get accurate temps? The board is a Gigabyte P35-DS3L Rev 2 if that makes a difference. I modded the BIOS myself and posted it earlier in this thread, but give me a shout if you need it. I can repost.
 
#2,163 ·
Quote:
Originally Posted by sjwaste View Post

This is going to be a dumb question.

My x5470 showed up today. I'm going to install it later, but presuming that I can boot, what is the best program to be using to monitor temps and is there anything I need to modify to get accurate temps? The board is a Gigabyte P35-DS3L Rev 2 if that makes a difference. I modded the BIOS myself and posted it earlier in this thread, but give me a shout if you need it. I can repost.
Use CoreTemp, and adjust the temperatures to -15 (all cores). That'll change TJMax from 100 to 85, which the real TJMax is 85.
 
#2,164 ·
Quote:
Originally Posted by kataomoi View Post

For you, not having enough voltage to RAM caused the video problems too? I didn't overclock at all and left everything on auto so I figured I did something wrong but it's been "working" normally other than the video problems and other stuff I described. The video problem was definitely not happening before I put my E5450 in and was still using my Q6600 so I thought I did something wrong.

I'm not sure, but maybe my video problem is completely separate.
You went from a 1066 bus processor to a 1333 bus processor which would be like an overclock. So you might do a reset to optimal because I have noticed some "stickage" with bios settings.

Well, I didn't have lines like you describe, but the PCI-E goes through the MCH. Mine was freezes, or graphics wise, was the driver would not load or come up with garbled video then a reboot. But I am running an Radeon 5770. First thing I think of when I see lines on the screen is video card memory gone bad. Could have just been a coincidence. The suggested voltages on that page helped me a bunch even without overclocking. What I kept seeing over and over with these gigabyte boards is the importance to A reset bios to optimal B hit ctrl+F1 at the main bios screen to enable extra settings, and third was to manually set everything to what is stated in white to the left of the [AUTO] settings. people have been having issues with this board for years so there is a pretty good amount of stuff I found when I comes to voltages even if it is just stock settings. Manual set the PCI-E to 100MHz and "FAST" also seems to be a common response.
 
#2,165 ·
Quote:
Originally Posted by kataomoi View Post

For you, not having enough voltage to RAM caused the video problems too? I didn't overclock at all and left everything on auto so I figured I did something wrong but it's been "working" normally other than the video problems and other stuff I described. The video problem was definitely not happening before I put my E5450 in and was still using my Q6600 so I thought I did something wrong.

I'm not sure, but maybe my video problem is completely separate.
Here is a guide I also used. Very good. This was written by ProfJim at tweaktown forums...

Sometimes, a new mobo will have "issues" that include rebooting, instability, or boot loops.

The following is primarily for EP4x motherboards but might help other Gigabyte models.

The following settings should be set after you have loaded Optimized settings:

PCI Express Frequency (Mhz)

: [Auto] <--- set to 100

Performance Enhance

: [Turbo] <--- set to Standard when overclocking
Extreme Memory Profile ---- (X.M.P.) ---- : [Disabled] <--- Leave this Disabled

Motherboard Voltage Control
CPU
CPU Vcore

1.?????

: [Auto] <--- set to the 1.????? value displayed in your bios
CPU Termination

1.200V*

: [Auto] <--- set to the value to the left (Auto over-volts this too much when overclocking) Auto is OK if not overclocking
CPU PLL

1.500V*

: [Auto] <--- set to the value to the left (Auto over-volts this too much when overclocking) Auto is OK if not overclocking

MCH/ICH
MCH Core

1.100V

: [Auto] <--- set to 1.20 for 2x2GB ram (1.24 - 1.26, or 1.05 might be needed)

ICH I/O

1.500V

: [Auto] <--- set to the value displayed in your bios
ICH Core

1.100V

: [Auto] <--- set to the value displayed in your bios

DRAM
DRAM Voltage

1.800V

: [Auto] <--- set to the max rated ram voltage (for most memory)

Advanced Settings

The following settings will improve system stability for overclocked systems.
Limit CPUID Max. to 3

: [Disabled]
No-Execute Memory Protect

: [Enabled]
CPU Enhanced Halt (C1E)

: [Disabled]
C2/C2E State Support

: [Disabled]
x C4/C4E State Support

: [Disabled]
CPU Thermal Monitor 2 (TM2)

: [Enabled]
CPU EIST Function

: [Disabled]
Virtualization Technology

: [Disabled] <--- Enabled if you use Vmware/Virtual PC

Integrated Peripherals
Legacy USB Storage Detect

: [Enabled] <--- set to Disabled when overclocking (this must be enabled to flash or boot from a USB drive)

These settings will eliminate many mobo problems unless there are hardware and/or cable connection problems.

===============
 
#2,166 ·
@ x11nt4
Thanks I'll keep that information in mind.

More on the video problem, which I do believe is completely separate from the 771 mod itself, is that I could easily reproduce the problem if I plugged my monitor into my dvi only slot (I'm using a asus gtx 670, one slot is DVI/VGA and the other is simply DVI only). It would continuously turn on and off. During a restart I left it on my login screen and it would flicker on and off and eventually the screen got garbled but the login screen was still readable. Though, it never forced a restart on me or anything. Something like this
The problem would not occur at all if I used a VGA>DVI adapter for my monitor or using my old CRT. Another thing, when I plugged my CRT into one slot, and my DVI-monitor into the following slot, whenever I would move my mouse, right click my desktop, or move a window the DVI-monitor would turn off/on without fail, just like in the youtube link but more frequently depending on what I was doing to force it. Without nvidia drivers it never occurred and when I tested by using my old ATi 4870 it never occurred. I wish I had another computer I could test my video card in just to be sure but I don't have such resources at hand.

I put my Q6600 back in just to see and it looks like those horizontal lines still show up randomly and pretty much everything above happens still.

I'm sure the 771 mod worked out fine for me albeit the original few problems of reboot cycles during a restart, but in return I found another separate problem to deal with. At least I still have warranty on my video card.
 
#2,167 ·
Quote:
Originally Posted by TerminalVoltage View Post

Voltage readings in this case won't do any good there is too much circuitry involved. I do think we're on the same page here. We will basically be creating a new set of lands with the pain pen using the existing lands I have mapped out. When done the previously unused pins will make a complete circuit and the motherboard will hopefully boot. Conductive paint pin may not be the most rigorous media to use, but as long as you're not continuously removing the CPU it should be fine. It should be good enough to prove whether on not this will work. You'll need decent masking or electrical tape too. It can't allow the paint to get under it while drying. If you have the board out, look at the socket with the CPU in it and see if it looks like the pins in rows A 28 and AN 28 are making contact with the CPU. You'll probably need a magnifying glass or take a macro shot with a camera to be certain.
I know I'm not sharing all my info for this mod yet, but I want to be certain that all the steps are understood and done in order before beginning. The last person I spoke to about doing the mod up and fried their CPU before I even said what to do lol. Patients is the key here. I work on state of the art avionics so I'm confident in my methods. I have a two part goal here:
1. Get the mod to work in X series chipsets.
2. Don't fry the motherboard...

The only board I know of that worked was a gigabyte X38 series with a 33XX series Xeon CPU. All others I know of have failed no matter the brand. The issue is the POST so I doubt modding the bios will do any good as this is before the boot up sequence. Modding the BIOS in this fashion will likely create a board that never boots...
Of course the X33xx series is a 775 chip so I would expect that. Mine runs fine with the X3220 I've used since day 1 with the P5E. Patients is definitely the hard part of any mod
smile.gif
I have a couple LGA 775 boards, although they are G31 and a 965, but I can use those to at least test to see if after this mod if it still comes alive. I would think that should be a good test at least electrically speaking before moving it into the X38. They are a couple boards I don't have any real use for anyway so if one has it's gene blown out of the lamp, then oh well! lol I have 3 LGA771 chips to play with but I will take a look at this and compare these sockets.

Just to be sure here, your saying you believe intel took the 4 lands on the cpu ( 8 counting both sides ) that are missing from the 771 package, and relocated them to the 2 extra lands ( 4 counting both sides ) where they are missing on the LGA775 where the notches are in the cpu correct? From your telling me you believe that the VTT isn't probably the issue, but rather the VSS/VCC
 
#2,168 ·
Quote:
Originally Posted by kataomoi View Post

@ x11nt4
Thanks I'll keep that information in mind.

More on the video problem, which I do believe is completely separate from the 771 mod itself, is that I could easily reproduce the problem if I plugged my monitor into my dvi only slot (I'm using a asus gtx 670, one slot is DVI/VGA and the other is simply DVI only). It would continuously turn on and off. During a restart I left it on my login screen and it would flicker on and off and eventually the screen got garbled but the login screen was still readable. Though, it never forced a restart on me or anything. Something like this
The problem would not occur at all if I used a VGA>DVI adapter for my monitor or using my old CRT. Another thing, when I plugged my CRT into one slot, and my DVI-monitor into the following slot, whenever I would move my mouse, right click my desktop, or move a window the DVI-monitor would turn off/on without fail, just like in the youtube link but more frequently depending on what I was doing to force it. Without nvidia drivers it never occurred and when I tested by using my old ATi 4870 it never occurred. I wish I had another computer I could test my video card in just to be sure but I don't have such resources at hand.

I put my Q6600 back in just to see and it looks like those horizontal lines still show up randomly and pretty much everything above happens still.

I'm sure the 771 mod worked out fine for me albeit the original few problems of reboot cycles during a restart, but in return I found another separate problem to deal with. At least I still have warranty on my video card.
His video sure looks to me like a video card that is having crashes and the drivers are trying to reset the card....

I would try setting the core and memory speed as low as the drives will allow for that card and see if that makes a difference. Could be a power supply issue. When the drivers are installed, it turns things on in the video card that are not enabled when running with default svga drivers. Might try a different driver. Possibly over heating problem but it sounds like power supply or drivers. I had a pair of 3870's back in the day with a PowerPC & Cooling 750W psu. It worked fine for about 6mo, but it gradually started crashing and strange video problems like what is shown in the video. It was the PSU couldn't supply enough to the cards anymore. That's when I said hell with it and went to a 1200W. That was 4 or 5 years ago now, never had one issue since with any of the upgrades I have done since.
 
#2,169 ·
Trying to see how high I can go here. It seems pretty stable so far as I continue to go through some benchmarks.
I tried 4500mhz (1800 x 10 ) at 1.45v -- can get into windows but can't do anything else.

One thing I have noticed is that with vdroop I couldn't even come close. Using the pencil mod helped out a ton for my 780i.


 
#2,170 ·
Quote:
Originally Posted by TerminalVoltage View Post

There's no cutting or soldering required. To prove my theory or disprove it only requires a conductive paint pen (easiest) or foil and adhesive. I really recommend against solder for this until it is at least proven. Soldering leaves no room for error and greatly increases the chances of something going wrong. I will show the first part of the mod (row AN) for starters. If it doesn't work the we can move to part two (row A).

Red lines in following image are just for reference.
As you can see the land in position AN 28 is only about 55% there. Hopefully this wont be a problem. There is always the possibility of filling in the groove with resin before attempting the mod, but I will leave that up to you. The resin would need to be completely flush to allow the CPU to seat properly.


A few multimeter readings must be done before the mod to ensure things for a smoothly as possible. Some of the lands are voltage others are return. If they are shorted together, your board will likely never boot again. The readings before the mod must match the readings after the mod or there is something wrong. Let me know which method you would like to try and we can begin. A windshield defroster repair kit may be the easiest and can be had at most auto parts stores. The foil method will be more tedious given how small the area is. Let me know what you would like to try and we can get started.
Hard to say how much a continuity test means on a cpu, however I show 0 ( full continuity ) between lands A25-26, also between E30 & F30, and between A25/26 to E/F30. Lands AN25/26 show full continuity. E1 to F1 shows full continuity. From AN25/26 to E/F1 I show a 2 for continuity I don't think they are direct linked. On the 775 cpu I see full continuity ( 0 ) between lands and AN25/26. I get full cont. between AN25 and AN29/AN30. AN28/27 Show full continuity. A25-A30 show full continuity

It looks to me like the 771 is getting a vtt signal. Only the VSS i'm guessing is missing. Looks to me AN27/28 need to go to E1 and F1. If I am understanding what the continuity check is telling me here
 
#2,171 ·
Quote:
Originally Posted by x11nt4 View Post

Of course the X33xx series is a 775 chip so I would expect that. Mine runs fine with the X3220 I've used since day 1 with the P5E. Patients is definitely the hard part of any mod
smile.gif
I have a couple LGA 775 boards, although they are G31 and a 965, but I can use those to at least test to see if after this mod if it still comes alive. I would think that should be a good test at least electrically speaking before moving it into the X38. They are a couple boards I don't have any real use for anyway so if one has it's gene blown out of the lamp, then oh well! lol I have 3 LGA771 chips to play with but I will take a look at this and compare these sockets.

Just to be sure here, your saying you believe intel took the 4 lands on the cpu ( 8 counting both sides ) that are missing from the 771 package, and relocated them to the 2 extra lands ( 4 counting both sides ) where they are missing on the LGA775 where the notches are in the cpu correct? From your telling me you believe that the VTT isn't probably the issue, but rather the VSS/VCC
Maybe I'm mistaken on the CPU series for the successful X38 mod, its been a while.

I'm not sure if it's one, the other or both as far as Vcc/Vss and Vtt. My gut says Vtt, but lets be methodical here. I figured try row AN first as it's pretty straight forward as to what needs to be done there. Row A on the other hand has a couple of options and may cause some issues if incorrect. If we do one at a time, we can rule out any errors. Plus if it works on the first try (fingers crossed) it will be less work. I have attached an images of what I'm proposing for clarification for row AN and instructions.

This mod as I'm sure you know, comes with risk. You are after all the first one I know of to try it. Proceed at your own risk.
cheers.gif




To begin you need a multimeter:
1. Measure and RECORD the resistance of one of the lands located in the green square labeled "A" with one of the lands in the square labeled "B". You will also need to pay attention to the placement of your leads, you are measuring through a circuit. Measuring with the red lead on "A" with black on "B" gave me over twice the resistance as the other way around. Record both values to be safe.

2. Measure and RECORD resistance on one of the lands in "A" with the land labeled "C". you should have a short no matter which way you place the meter leads.

3. Measure and RECORD resistance on one of the lands in "B" with the land labeled "D". You should have a short no matter which way you place the meter leads.

Now its time to paint. You will be creating the two squares labeled "A" and "B", one at a time. The first square will need to be fully dry or you will make a mess.

4.Clean area with alcohol and let completely dry.

5. Using a good quality tape (not Scotch) mask around the lands labeled "A" leaving only the area inside the square bare. Press down firmly on the tape to make sure it's fully stuck as to not allow any paint under it. It's also important that you keep the gap between the two squares in place as previously shown. One square is power the other is return, short them together and most likely you'll have a dead board. See image below, the yellow represents tape.


6. Paint the square labeled "A" and allow it to completely dry before moving on to the next step. You will need to paint all the way to the edge where the groove is cut into the PCB. Be extremely careful not to let the paint run over the edge of the PCB, it may bridge the gap between the two squares you are painting and cause a short. Also, don't paint past the red line, we want the CPU to sit flush. You want the paint thick enough so that it isn't easily scratched through, but not so thick that it chips off or interferes with proper CPU mounting.

7. Remove the tape from around square "A" when paint is dry and mask the CPU to paint square "B". procedures are the same as when you painted "A".

When done CPU should look like below with paint only in the blue area.


Now you need to take meter readings to verify proper modification.
8. Do the same measurements as were completed in steps 1, 2 & 3. The only difference being you will now be reading form "square A" & "square B" rather than the lands that were painted over. The readings should be the same as before. In no case should the squares be shorted together. If they are, something is painted wrong. Also note that the readings between lands "C" and "D" should be the same as the earlier measurements at "A" and "B" so you can measure from there to avoid damage to the paint from the meter leads.

cheers.gif
GOOD LUCK!
thumbsupsmiley.png
 
#2,172 ·
Quote:
Originally Posted by x11nt4 View Post

It looks to me like the 771 is getting a vtt signal. Only the VSS i'm guessing is missing. Looks to me AN27/28 need to go to E1 and F1. If I am understanding what the continuity check is telling me here
I should have clarified my intent a little better. The purpose of this mod is to give the socket 771 CPU the same voltage out/returns as its socket 775 CPU counterpart. This should get it working unless there is another swapped pin combo like in the original sticker mod that only affects certain chipsets.
confused.gif
I find this unlikely though. I also don't believe we need to mess with the lands on E1, F1, E30 and F30. One is a test land and I believe the other three are remapped in the socket to the positions that have missing lands on the CPU. Hence the mod.
thumb.gif
 
#2,173 ·
I have finally modded my nephew's Asus P5K Pro and successfully seated E5440 C0 in the first try. Everything runs fine and stable but there is one thing that baffles me. At stock frequency first two cores runs at 20 degrees C hotter than second two. 45C on the first two vs 25C on second two. This difference reduces as load increases. When running Linx difference is about 10 degrees C, 57C and 47C respectively. Cooler I used is CM Hyper 212 with Noctua NT-H1. I double checked if heat sink is seated well, and it was. This stinks like stuck/bugged on die temperature sensor, but I honestly I don't know who to trust, since those 25C at idle is unrealistically good temperatures. At mild OC 3.4GHz @ 1.216 VCore (stock voltage) there is almost no difference in idle and load temperatures.

I also noticed faint whining and hissing from the motherboard coils (I think it is from motherboard) while CPU is loaded, I mean there is clear difference in sound when Linx is running and when I hit stop. Can this be caused by weak PSU?
 
#2,174 ·
Quote:
Originally Posted by TerminalVoltage View Post

Maybe I'm mistaken on the CPU series for the successful X38 mod, its been a while.

I'm not sure if it's one, the other or both as far as Vcc/Vss and Vtt. My gut says Vtt, but lets be methodical here. I figured try row AN first as it's pretty straight forward as to what needs to be done there. Row A on the other hand has a couple of options and may cause some issues if incorrect. If we do one at a time, we can rule out any errors. Plus if it works on the first try (fingers crossed) it will be less work. I have attached an images of what I'm proposing for clarification for row AN and instructions.

This mod as I'm sure you know, comes with risk. You are after all the first one I know of to try it. Proceed at your own risk.
cheers.gif




To begin you need a multimeter:
1. Measure and RECORD the resistance of one of the lands located in the green square labeled "A" with one of the lands in the square labeled "B". You will also need to pay attention to the placement of your leads, you are measuring through a circuit. Measuring with the red lead on "A" with black on "B" gave me over twice the resistance as the other way around. Record both values to be safe.

2. Measure and RECORD resistance on one of the lands in "A" with the land labeled "C". you should have a short no matter which way you place the meter leads.

3. Measure and RECORD resistance on one of the lands in "B" with the land labeled "D". You should have a short no matter which way you place the meter leads.

Now its time to paint. You will be creating the two squares labeled "A" and "B", one at a time. The first square will need to be fully dry or you will make a mess.

4.Clean area with alcohol and let completely dry.

5. Using a good quality tape (not Scotch) mask around the lands labeled "A" leaving only the area inside the square bare. Press down firmly on the tape to make sure it's fully stuck as to not allow any paint under it. It's also important that you keep the gap between the two squares in place as previously shown. One square is power the other is return, short them together and most likely you'll have a dead board. See image below, the yellow represents tape.


6. Paint the square labeled "A" and allow it to completely dry before moving on to the next step. You will need to paint all the way to the edge where the groove is cut into the PCB. Be extremely careful not to let the paint run over the edge of the PCB, it may bridge the gap between the two squares you are painting and cause a short. Also, don't paint past the red line, we want the CPU to sit flush. You want the paint thick enough so that it isn't easily scratched through, but not so thick that it chips off or interferes with proper CPU mounting.

7. Remove the tape from around square "A" when paint is dry and mask the CPU to paint square "B". procedures are the same as when you painted "A".

When done CPU should look like below with paint only in the blue area.


Now you need to take meter readings to verify proper modification.
8. Do the same measurements as were completed in steps 1, 2 & 3. The only difference being you will now be reading form "square A" & "square B" rather than the lands that were painted over. The readings should be the same as before. In no case should the squares be shorted together. If they are, something is painted wrong. Also note that the readings between lands "C" and "D" should be the same as the earlier measurements at "A" and "B" so you can measure from there to avoid damage to the paint from the meter leads.

cheers.gif
GOOD LUCK!
thumbsupsmiley.png
Looking at your squared off areas to paint, it looks like this will connect AN27+AN28 to AM27+AM28 and AN29+AN30 to AM29+AM30 when it is put back into the socket correct?
 
#2,175 ·
Quote:
Originally Posted by NSTime View Post

I have finally modded my nephew's Asus P5K Pro and successfully seated E5440 C0 in the first try. Everything runs fine and stable but there is one thing that baffles me. At stock frequency first two cores runs at 20 degrees C hotter than second two. 45C on the first two vs 25C on second two. This difference reduces as load increases. When running Linx difference is about 10 degrees C, 57C and 47C respectively. Cooler I used is CM Hyper 212 with Noctua NT-H1. I double checked if heat sink is seated well, and it was. This stinks like stuck/bugged on die temperature sensor, but I honestly I don't know who to trust, since those 25C at idle is unrealistically good temperatures. At mild OC 3.4GHz @ 1.216 VCore (stock voltage) there is almost no difference in idle and load temperatures.
I don`t recall where I have read about temp sensors are not giving true readings below 45C (if I remember right). It must be on Intel`s white paper I think.
But yes, if temp readings at idle are too low and there is a diff between cores then that is normal.
If only at 60C -or more- readings between cores are almost the same with a 3-4C diff at most assuming all cores run at 100% continuously.
Quote:
Originally Posted by NSTime View Post

I also noticed faint whining and hissing from the motherboard coils (I think it is from motherboard) while CPU is loaded, I mean there is clear difference in sound when Linx is running and when I hit stop. Can this be caused by weak PSU?
Most probably the capacitors have gone bad. This makes the power stage whistling. If you encounter random BSODs while it seems stable and tests get it stable then sure a capacitor is bad.
Bad capacitors may damage the power stage and/or cpu. So better check.

It could also be PSU`s capacitors.
 
#2,176 ·
Well this is interesting,

I'm doing my second Xeon mod to a ASUS P5QL-E P43.

I had previously run a Q6600, and a X3370 for a while after a microcode update.

Now I got a X5470, and after some minor fiddling- notched the chip instead of the socket, and applied the sticker.

It POSTS fine, and hangs after the Windows 7 logo animation.

The weird part is, if I lower the multiplier to 8 or 8.5 it works ok. Any other ratio hangs. 9 worked for a while. Then hung on a shutdown.

I've runs Prime95, 3Dmark 6 at 8x/8.5x perfectly..... SSE4.1 and VT-x are showing...

As a side note, CPUz sees the CPU 'Specification' as a Xeon X5470 @ 3.33GHz., etc.. But changing the multiplier alters the Processor 'Name' to E5430/40/50 (for 8x/8.5x/9x ratios....) odd eh?!

I bought the 5470 for the golden x10 multiplier, but Windows fails at the logo with default bios settings, I have read from PioneerDude that he reckons the sticker was bad in his situation.... but this inability to run a high ratio is odd..

I 've had the board to over 400fsb on the x3370 (9x) without trouble. PSU is excellent 600w.

The cpu is running nicely as I write this post...at lower speed.

Does it want reseating again.. I'm kinda stumped.

The X5460 in my HTPC G35 worked out well and is operating now.

ideas?

Ralcool
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top