• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Blaylocks GA-MA790X-UD4P vCore Volt Mod

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Blaylock

"That Backfired" Senior Member
Joined
Jun 5, 2013
Location
Go Blue!
Thank you to macsbeach98 and Rasparthe for the guide and assistance in this. You guys are great!

***Warning*** - Moddifying you motherboard in this manner can damage your computer components.

***Caution*** - Do NOT leave overvolted computers unattended. They can burn-up and cause a fire, burning your house to ashes.

***Danger*** - Soldering irons are really, really hot. Picking one up like a pencil hurts really, really bad. Just trust me on this one.

Still reading? Good. As you may have guessed this is my Gigabyte GA-MA790X-UD4P vCore Volt Mod project. By following macsbeach98's guide located here I successfully hard modded my board. Why you ask. Because 1.5v isn't enough for my plans to LN2/DICE my AM2 CPU's in my sig. With this mod I'll be able to go well above the manufactured recommended settings and push these babies as far as the can(likely beyond what they can handle. Stay tuned for the fireworks folks).

Enough talky. Here's what I did. I purchased (1) 10k Ohm Cerment Potentiometer (Trimmer) from Radio Shack for about $3-$4. Stripped 1 wire from an old IDE ribbon cable. After practicing soldering on very, very small points on a dead MB I decided I was ready to give it a try. It's been a long time since I've done any precision soldering and it was definitely a challenge. I've include an image with a penny to show just how small these points are. 1 point is about half the size of the letters on the penny.

I stripped about 3mm off one end of the wire and tinned it. Clipped that down to about 1mm and soldered it to the correct point. Then measured the length needed to mount the trimmer where I wanted it. Macs used the back of the keyboard/mouse I/O panel and it seems like as good a place as any so I used the same. I then adjusted the trimmer to max resistance (full clockwise). I believe this is to minimize the risk of overheating it while soldering to it. Then I soldered the other end of the wire to the "bottom" lead on the trimmer. I then cut a short piece of wire to make the ground. Soldered that to the middle lead(the actual adjustment pole) and then to the keyboard/mouse I/O shield(ground). The trimmers top lead will not be used in this application so it's best to trim it off. Used a little double stick tape to secure it in place and you're done. Hot glue would have been better for this but my hot glue gun has grown legs and has apparently gone on vacation.

View attachment 161487

View attachment 161488

View attachment 161489

View attachment 161490

View attachment 161491

I did have some troubles getting it to post, but that's because I had the settings on the trimmer wrong. If you follow the steps I did you'll want the trimmer turned fully clockwise(Max resistance) to set the vCore close to the BIOS setting. Otherwise you're way overvolting it. I'm just glad this system survived my brain fart.

Here are the results.

At stock settings before the volt mod my Athlon 64 X2 6000+ had 1.40v in BIOS and displayed 1.392 in CPUz.
View attachment 161492


At stock settings after the mod the (1.40v in BIOS) CPUz displayed 1.456v. A bump of 0.064 with the trimmer at max resistance.
View attachment 161493


To offset this "free" bump in voltage I need to remember that what I set in the BIOS will be increased by approximately 0.064v. Here's a shot of CPUz with the Bios set to 1.35v. Actual vCore is now a hair over 1.4v for my "Moddified" stock setting.
View attachment 161494


This is all with the trimmer still set at max resistance(lowest vCore increase). How much will turning the screw get me? You'll have to stay tuned for that one. I'll need to get some cold!!!

Stay tuned.
 
@atm the one I got was a 15 turn.

@shrimp yes they do, but I need to find out where and "try" and hold just above that. I have a few ideas on how to accomplish this, but they're only theory at this point.
 
Nice mod Blaylock! :thup:

I hope it serves you well when you crank up the volts for some DICE fun!!
 
I can tell you a 6000+ will cold bug before a 6400+ having tried several. Pretty much luck of the draw. Most at 0f or 0c which ever it didn't like first, it was a show stopper. Chilled liquid and 1.7v to about 1.8v is pretty doable and usually yields great overclocking. edit additional comment: 6400+ is good for about -45c, the 6000+ may be a little under that. Give or take. It's not going to be fill the pot with LN2, but regulate how little you pour in and how often. It's tricky. Doing this during high clock freezing, no posts and blue screens. Oh the fun fun fun.

On the other hand, I've seen some of these chips run 50-60c 3400mhz on air and was happy as can be. Chilling having no effect, this happened alot with Brisbanes if I recall Windsors seemed to clock well with the cold. but not so frozen.

Most semprons will like above freezing temps and low end athlons.

Anything Agena is just going to be what the cpu can do more so than the cooling. Early revision chips might have been on average 3ghz and later revision where around 3.2ghz IE: 9850BE was only 2.5ghz be damned the darn chip go over 3ghz unless you bought a few and got lucky.

Most fun you'll have is finding Phenom II chips and LN the f out of them. You'll get good LN2 response from those chips ;) On average 6ghz plus.

I did find a rarity in s754 class, was a paris core. It loved the cold. ALL other chips I owned in that socket wouldn't post even in most cases running a simple TEC. Just no post :( When this paris core posted right up running negative, I decided to try the LN2 on it. Worked out pretty good. Diamond in the ruffz. http://valid.canardpc.com/show_oc.php?id=2606835

Oh and TEC works great for testing cold on chips. lot less work then the LN2 pot ;) especially on the single core stuff from back in the day.

I should mention AM2 is one of my favorite all time sockets for benching :D
 
Last edited:
IE: 9850BE was only 2.5ghz be damned the darn chip go over 3ghz unless you bought a few and got lucky.

I actually got 3038 MHz from a 9850 (non BE) on air ! It didn't do much at that speed , but it ran long enough to post a CPUZ verification. LOL I don't have the link handy ATM , but it's some of my older posts. (I was overly proud and posted that puppy)
 
I actually got 3038 MHz from a 9850 (non BE) on air ! It didn't do much at that speed , but it ran long enough to post a CPUZ verification. LOL I don't have the link handy ATM , but it's some of my older posts. (I was overly proud and posted that puppy)

It's the older stuff that brings know how onto now a days PC overclocking. Only now the boards are factory tweaked, no need for hard mods any more. Even a Good AM2 board is gonna give you 1.9v cpu or more. In most cases with all the AMD cold bugged stuff, it's rare to use that much voltage any ways. Talking 1.7v, the chip better like chilly.
 
Awesome. What's the answer? Can you link it here for all posteriors? Lol
 
It's the older stuff that brings know how onto now a days PC overclocking. Only now the boards are factory tweaked, no need for hard mods any more. Even a Good AM2 board is gonna give you 1.9v cpu or more. In most cases with all the AMD cold bugged stuff, it's rare to use that much voltage any ways. Talking 1.7v, the chip better like chilly.

I had a trash can of RAM in it, too. I had no clue. :facepalm: Stock voltage. It wouldn't go 1 MHz faster no matter where I set the voltage. I now suspect the RAM may have been an issue, and it wasn't a BE. :rofl: Mobo was a M4A785-M. http://valid.canardpc.com/show_oc.php?id=2400472
 
I need to review this thread again, as I am now wondering if this is possible with the Proliant ML115 G5 board I have in the WhiteElephant (sig).
Mr. Scott, you know this board a bit...
It maxes out at 1.3v and I can reach around ~blk 230 for most benches, but CPUCool will let me go too 266.
Can I get more with a hard-mod on this mobo. Is it even possible?
I do have two of these boards...

feedback/advice

Cheers
 
You need to find the power control chip and then the data sheet if possible or look for other mods to the same chip. I recently did this to a 9800GTX+. Was having a hard time finding the same card/PCB. Used the name of the chip in a search and found a card with the same PCB, different model, the only difference was the resistance readings so I needed different pots than the online guide. It's definitely a labour of love. Hats off to those that take the plunge like you Blaylock.
Anytime Mike.
It is a very good board. Getting harder to find now.

There were a bunch on ebay just recently. Not cheap but cheaper than most. I didn't realize at the time it had this limitation with the AM2 chips
 
Back