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

AN7 Vdimm mod gone Bad, need help!

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

Darien

Member
Joined
Jul 1, 2005
Location
Hawaii
Hi guys,

I perfromed the AN7 Vdimm guide posted here:
http://www.vr-zone.com/?i=759&s=1

and it works fantastic Thx!! actually I was advised to just solder a wire to the r316 resistor solder the other end of the wire to a VR - trimmer and then ground the 2nd leg of the Trimmer.

This is a picture of my motherboard with the modifications
http://www.ocforums.com/attachment.php?attachmentid=43445&stc=1

and it worked great for over 1 week doing lots of testing and tweaking, no problems! but then my stupidity set in... I did a big NO NO and adjusted the voltage with the computer running and I fried the resistor because of a volt fluxuation... I think... the resistor is grounded now? and now the board still works, but the Vdimm is stuck to MAX... the same as the 3.3v rail... and I can't turn it down in the BIOS or with the Trimmer...

after trouble shooting I realized that the right side of the R316 reistor has 3.5 volts(the side closest to the RAM slots), while the left side has almost no voltatge at all.... AND the left side of the resistor has zero Ohms resistance to any ground point on the board...

Unfortuneately I think that maybe I didn't fry the resistor, but rather the circuit that leaves the left side of the resistor...

can anyone here give me any insight on how to fix this? I would unsolder the resistor and replace it, but I'm not sure thats the problem...

This isn't my first volt mod and any help you guys can give would be very much appreciated.

TIA

P.S. I think I know 1 way to fix/work around this, but I'm hoping to find out exactly what happened and maybe fix it the correct way.
 
Last edited:
SolidxSnake said:
de-solder the VR you put there, see if it works.

yeah, sorry I wasn't clear on that... I did completely take the mod out, didn't matter one way of the other...

see the thing was, when the mod was hooked up no matter how I set the VR, there was 0 Ohms across my VR... so I thought my VR was bad... when I disconnected the ground leg of the VR completely it had over 10 Ohms across the VR... the reason why... the left side of the r316 resistor was grounded and had zero 0 ohms with any other ground point...

so with the VR completly 100% disconnected and out of the case, or with it connected, it still acts the same...

putting a grounded VR to the r316 won't matter if r316 is grounded to begin with... even if the grounded VR has set to 0 Ohms to ground the r316.... r316 is grounded already... which is why it's stuck on Full voltage

right side of the resistor has 3.5 volts the left side of it is grounded...
 
hmm. Try soldering the VR in place of the old resistor, see if you can control volts with those two spots. if you can, then the resistor's junk, find the resistance of it and replace it
 
SolidxSnake said:
hmm. Try soldering the VR in place of the old resistor, see if you can control volts with those two spots. if you can, then the resistor's junk, find the resistance of it and replace it

Yeah, when I was saying I thought I knew the fix before, this was it

1alt_method.jpg

just remove the resistor and like the pic shows, don't reattach the left side of the old resistor point, (which is a dead ground now) and instead just route it straight to the Vdimm Mosfet... then attach the VR to that... I'm pretty sure that will work... but I don't know 100%, and if that does work I don't know if my Vtt will still track and If I can still turn down the volt in my Bios.... but at least is will let me turn down the Vdimm....
 
Darien said:
Yeah, when I was saying I thought I knew the fix before, this was it

1alt_method.jpg

just remove the resistor and like the pic shows, don't reattach the left side of the old resistor point, (which is a dead ground now) and instead just route it straight to the Vdimm Mosfet... then attach the VR to that... I'm pretty sure that will work... but I don't know 100%, and if that does work I don't know if my Vtt will still track and If I can still turn down the volt in my Bios.... but at least is will let me turn down the Vdimm....


hmm. I was just thinkin of soldering a VR to the points the old SMD resistor was at. Nothing more.

But your plan looks good too :D
 
SolidxSnake said:
hmm. I was just thinkin of soldering a VR to the points the old SMD resistor was at. Nothing more.

But your plan looks good too :D

Yeah, just basically replace the R316 resistor... well... I thougt of that... but I'm not sure the r316 resistor is the problem...

is the left solder pad where the r316 resistor is, normally a ground point? if it's normally a ground point then I'd definately replace it... but I was thinking the line to the left solder pad was ruined, is now grounded and therefore needs to be bypassed with the mod I posted above....

Again, I'm happy to try replacing the R316 resistor, but do you know is the left solder pad where the r316 resistor is, normally a ground point?

Thx Again
 
Very tough and trying, but victory at last Thank God

everything got far far more complicated than it should have, but by some miracle, and alot of work, I got it working and I know what happeded....

but basically like the guide from VR-zone says, 0 ohms across the r316 solder pads is a zero increase, it's actually a decrease... if you remove the r316 resistor and short the solder pads (0 Ohms / no resistance) you will far below the voltage set in your BIOS, often too low to even POST into the BIOS... If you put maximum/too much resistance, remove the resistor altogether, or completly sever the line between the solder pads you will be stuck at maximum resistance...

the only plausable solution was to connect the two r316 solder Pads, with a VR Trimmer... I used a 1k 15turn trimmer.... this was insanely difficult... I won't even go into it... but unless you got mad crazy soldering skills, or have NO other choice (like me) don't do it this way...

Do it the way I did it the first time, but use the correct VR trimmer!!

just solder a wire to the r316 resistor, solder that wire to a 1K Ohm 15 turn (or more) VR trimmer, solder the second leg of the 1K Ohm 15 turn (or more) VR trimmer to a ground point.

MY MISTAKE: I believe what happened to the first mod I did... I used a 20k Ohm (instead of 1k Ohm) VR that probably had only 15 turns... not very sensitive (20 times less sensitive) and VERY easy to make a rapid voltage change... I believe this is what happened, I changed the voltage too quickly with the computer running and got a voltage fluxuation and it blew the r316 resistor... when this happened it somehow radically increased the resistance in the r316 resistor thus forcing my Vdimm to full throtle maximum... I've heard the R316 resistor is a 0 Ohm resistor, I don't think that's the case probably about a 300 Ohm resistor (guessing, it said 87A), but after it blew I removed it from the board and measured the resistance of the r316 resistor and it measured about 2k Ohms!!! I can get full throtle Vdimm with less that 1k Ohm resistance....

the trimmer you want to use for this mod is a 1k Ohm VR Trimmer with atleast 15 turns preferably 25 turns

NOW:

In trying to fix this, I screwed up so bad I truly thought I had ruined the board... this is for anyone who screws this up royally... there is hope...

after I removed the r316 resistor, and I was trying to solder connections to the solder pads, I tried a few different things... It was insanely hard to make clean solder connections to those tiny pads... one thing lead to another I touched the solder pad one to many times and the solder pad was removed from the PCB of the motherboard!!! at this point no solder would stick to the board at all!!! I tried for 4 hours to make a clean solder connection where I had burnt out that pad and finally I did it..... I burned out the 2nd metal solder pad of the r316 resistor :temper: ... I almost cried!!! :cry: but I found success...:cool:

I took a sharp tweaser and scrapped away some of the PCB surrounding the holes in the mobo where the solder pads used to be, so I could expose some of the copper lines that the solder pads connected to...

then I got a conductive Ink Pen from Radio Shack.... (got much more complicated, but I'll just tell you the easy way to fix it)... filled in the two holes where the solder pads were with conductive ink... DO NOT CONNECT THE TWO HOLES WITH THE CONNECTIVE INK!! at this point we again have two conductive points... the problem is you can't solder onto them... what I did to the left one was super glue.... yes super glue the wire to it, I got the super glue on and around the surronding PCB, but that was OK, it's not conductive.... for the right Pad I drew a line with the conductive ink out of the right ink hole to the leg of a super tiny miny mosfet 2mm directly to the right of the hole.... then I soldered the 2nd wire to the that mini mosfet leg..... which was also incredibly hard... then ofcourse I connected the two wires to the 1K Ohm 15turn VR trimmer.

the point... DO IT RIGHT THE FIRST TIME!!!! but even you're as dumb as me, and screw it up this bad, it can still be fixed...

my board is working great, with Vdimm adjustable from the BIOS and from the trimmer... :beer:

A NOTE about adjusting the Vdimm... don't make huge volt changes in the BIOS... for instance don't take it from 3.2 to 2.8 all in one shot... change it one setting at a time, go back into the BIOS and recheck the voltage... if your actual voltage is too far out of line with your voltage setting the AN7 will try to make a correction push your Vdimm full open again, and then you have to trim it back down... actually this seems to happen VERY often, when saving a new BIOS profile and sometimes with a FSB adjustment... basically anytime you change anything go back into the BIOS to check your real Vdimm.

One last thing... Only do this mod if you need more that 3.2v... don't do it for the heck of it... don't do it and then run TCCD at 2.9... you can... but it becomes a pain because adjusting the volts in the BIOS becomes very finiky... everytime you change something in the BIOS you need to recheck you actual voltage because it may get set to full Vdimm by default, especially when saving a new bios profile or changing FSB...
 
Last edited:
Wait a minute... you're using the middle leg of an FET? I'm not positive, but I always thought the middle leg of a MOSFET was the - lead, where voltage was sent out

either way, thats some pretty crazy stuff.
 
It only takes a little voltage build up on the soldering iron or some static electricity to pop a chip including MosFets. Toss the board. Even if you get it to produce power, it's going to frag on you.
I once screwdrivered a board, just 1 trace, just 1. Stupid 370/SKT-A POS clamps :mad:
So i soldered the smallest bead I could get on a piece of 30AG wire I could and fired it up. Everytime you boot the thing, you have to reboot and tweek bios to get it to pick up the frequency correctly. The solder create a "signal speed bump" an causes the processor to mis-read the clock freq.
 
AlabamaCajun said:
It only takes a little voltage build up on the soldering iron or some static electricity to pop a chip including MosFets. Toss the board. Even if you get it to produce power, it's going to frag on you.
I once screwdrivered a board, just 1 trace, just 1. Stupid 370/SKT-A POS clamps :mad:
So i soldered the smallest bead I could get on a piece of 30AG wire I could and fired it up. Everytime you boot the thing, you have to reboot and tweek bios to get it to pick up the frequency correctly. The solder create a "signal speed bump" an causes the processor to mis-read the clock freq.

Thx for the input, seems to be working fine, benches good, plenty stable, no guarantee if won't fail, never is... in the immortal words of Arnold Swarzenneger "...if he dies, he dies...: :shrug: if the board blows I'll just get another one... but the fix I did is better than just throwing it away, especially since it's working fine.
 
quick note over 6 months later its still working great. overclocks like a monster =)
 
Back