it7 vmem mod

[phantompunisher]

i read the thread below but its still unclear as to witch resister to solder to. the picture seems to show one but it sounds like they are talking about a different one.

[hipro5]

Originally posted by phantompunisher
i read the thread below but its still unclear as to witch resister to solder to. the picture seems to show one but it sounds like they are talking about a different one.

Hi there.....It's the SECOND one from the left witch you have wroten(...or this one....).....

[phantompunisher]

this dosnt seem to be working. the board i have is actualy a it7 not a it7 max2. the layout is the same though but i have found the spec sheet for the mem voltage chip here http://www.fairchildsemi.com/ds/FA/FAN5063.pdf
yea resistor number two is the one to mod but there must be something else limiting the voltage

[hipro5]

What do you "exactly" get with the Mod?...

[phantompunisher]

i get 2.97 for a max witchis the same as when im using 3.2 in the modded bios. i belive i have found the problem: http://www.samhop.com.tw/pdf/SDP(B)55N03L.pdf

i have a drain source voltage of 2.98. i belive if this were 3.3 then the full 3.3 would flow through. the mosfet is not fed by the 5 volt line as i once thought. i have run tests from 2.5 -3.2 and taken measure ments at all points of interest. i will do some more investigation. its not for lack of a strong 3.3 line, if this is what the mosfet is run from, because my 3.3 is reading 3.29v

[hipro5]

Originally posted by phantompunisher
i get 2.97 for a max witchis the same as when im using 3.2 in the modded bios. i belive i have found the problem: http://www.samhop.com.tw/pdf/SDP(B)55N03L.pdf

i have a drain source voltage of 2.98. i belive if this were 3.3 then the full 3.3 would flow through. the mosfet is not fed by the 5 volt line as i once thought. i have run tests from 2.5 -3.2 and taken measure ments at all points of interest. i will do some more investigation. its not for lack of a strong 3.3 line, if this is what the mosfet is run from, because my 3.3 is reading 3.29v

You have to Mod your PSU to get at least 3,5Volts so you can get 3,2Vmem....coz the Mos-Fet "get's" Voltage from the 3,3Volts.....

[phantompunisher]

is that what you had to do? or did your board naturaly supply more voltage to the drain? i cant 3.5 my psu, that would run a lot of things out of spec

[phantompunisher]

welp i did my own mod after finding out why the it7 has only the 3 volts for memory. there is a diode in line with the 3.3 volt rail on its way to the drain. thats where you loose the .3 volts due to a voltage drop across the diode. its a very strange diode indeed. but anyway i bypassed it and ran the 3.3 to the drain and saw only a slight improvment in mem mhz. it was a fun prodject but ill just hold out for different memory thanks hipro5

[hipro5]

I have done that too but......you HAVE to Mod your PSU as well....

[phantompunisher]

nah i only wanted 3.3 volts to the memory. nothing higher.

[hipro5]

How's that???

[JCviggen]

I took my Vmem to 3.42 some time ago (Corsair 3200 C2 I think... maybe the 3500 already, not sure) by soldering the 3.3V line from the PSU straight to the mosfet, but I only ended up with worse overclocking... I think my mem would be happier at 3.1-3.2V

[hipro5]

Originally posted by JCviggen
I took my Vmem to 3.42 some time ago (Corsair 3200 C2 I think... maybe the 3500 already, not sure) by soldering the 3.3V line from the PSU straight to the mosfet, but I only ended up with worse overclocking... I think my mem would be happier at 3.1-3.2V

Hey, JC....I like your sign man.....

[phantompunisher]

i think i figured out what the diode is for. i belive its there to protect the ram from a static charge. so i guess bypassing it is fine. dont quote me though

[Tom Holck]

Originally posted by phantompunisher
i think i figured out what the diode is for. i belive its there to protect the ram from a static charge. so i guess bypassing it is fine. dont quote me though

Have you a picture, where the diode is placed?


Best OC regards


Tom Holck
Denmark
The Homeland of Prometeia

[phantompunisher]

on closer inspection you also bypass the load capaciter. anyone looking to do this as a perminate mod may want to look into adding that cap back in if the mem voltage drops on the start of heavy use. but here is is. jumpering the 3.3 to the bottom (as it is oriented in this pic) of the diode will give you 3.3 to the drain of the mosfet alowing you to select up to 3.2 int he bios and 3.3 with a mod to the second resister as mention at the top of this thread

[phantompunisher]

actualy just run a jumper from one end of the diode to the other that will give the full 3.3 to the mosfet and keep the capacitor in the circuit

[MrIcee]

Guys

I am very happy to see a success happening here with the IT7 mem volt mods. I would like to publish this as a guide for XtremeSystems with full credit to those successfully coming up with all aspects of the mod. I have an IT7 Max2 ver2.0 that I'll eventually be setting up as a 2nd test bench, and this mod is a major breakthru. Email me at MrIcee@xtremesystems.org with details
Thanks in advance !

Randi

[hipro5]

Originally posted by MrIcee
Guys

I am very happy to see a success happening here with the IT7 mem volt mods. I would like to publish this as a guide for XtremeSystems with full credit to those successfully coming up with all aspects of the mod. I have an IT7 Max2 ver2.0 that I'll eventually be setting up as a 2nd test bench, and this mod is a major breakthru. Email me at MrIcee@xtremesystems.org with details
Thanks in advance !

Randi

Thanks MrIcee for your good words and it'll be good to publish it so people can see it and do it to get some benefit as well....

[PiLsY]

GJ phantom!

I should state that on my v2.0 board Hipros mod infact fried the vmem controller. However, bypassing the diode looks like it may meet with more success. Im picking up a new board in the morning, so i'll be able to report then what success (if any) I have. I think it may actually proove easier to use conductive paint to simply join the 2 ends of the diode. The jumpered method would give the ability to undo the mod, but if Ive figured this correctly 2.5v in bios should give around 2.85v which would be fine for day to day use. We shall find out tomorrow .

Again, nice work phantom, and to you as well Hipro5 for sussing out the starting point .

PiLsY.

[PiLsY]

Well, i couldnt wait for amtrak to deliver so I drove down to their depot for 6am and picked up by hand . Its now 7am here and ive been up all night so bear with me a bit on spelling and grammar.....

Right this mod does NOT work yet on the Max2 v2.0. Its very close, but it needs some tweaking. Ive performed the mod as listed above and seen some very odd results. However I am confident of finalising it and getting more predicatble voltage outputs. At this point in time Ive done the following :

Silver paint across diode, no other mod in place. Zero effect on voltages.

15k fixed resistor across R110 - Very odd results leading to lower than default voltage.

15k fixed resistor across R110 and silver paint across the diode. Nonesensical results listed below :

2.5v = 2.59v
2.6v = 2.59v
2.7v = 2.59v
2.8v = 2.59v
2.9v = 2.64v
3.0v = 2.72v
3.1v = 2.76v
3.2v = 2.85v

This is not a linear climb in voltage, its moving in fits and starts. Here are my results with no mods and a modded bios in place :

2.5v = 2.51v
2.6v = 2.59v
2.7v = 2.68v
2.8v = 2.76v
2.9v = 2.51v
3.0v = 2.59v
3.1v = 2.68v
3.2v = 2.76v

This shows the modded bios HAs worked but that it doesnt actually program the hardware correctly for extra voltage. With mods in place the 2.9v to 3.2v settings do offer a climb in voltage, without they simply mirror 2.5v to 2.8v. All in all very odd.

I hope to have all problems ironed out and some pictures up by tonight if I can find enough free time.

PiLsY.

[phantompunisher]

son of a gun it looks like they modded the bios in the v2. the sneaky bastages. what vmem chip is on the max2 v2? is it the fan5063m? what do you have for a mosfet drain voltage? wait... actualy the diode jumper wont do anything provided the bios and resister trick arnt working. no use having a higher drain voltage when there will be no signal to release it. this is a bummer your going to have to hunt down the resistors as the bios will be no help to you.
if it is the fan5063 look at pin 12. follow the trace to both of the resistors that you will be looking for. get a reading of the resistance of both. r2 will be connected to grond on one end.

post some picks when you can. this is odd that its diferent

you may have found r1 in witch case if you resistored across it it will have lower voltages

[PiLsY]

Same vreg, slightly ddifferent layout of resistor. Resistor 4 for example in the bank of 6 SMDs is blank. Nothing connected at all. Pin 7 connects to the resistor to the left of R110, then doubles back and goes through R110.

I should say this is now the second board ive had where the modded bios has done nothing. However, at this moment in time I am alone in this as nobody else on these forums has posted problems with the modded bios. It seems tow ork for everyone else. I am waiting on seeing pics of the vmem reg area of a board that the modded bios works on. Im convinced there are 2 revisions of this board, whether I have an earlier or later one I dont know. I should point out that I am buying form the UK, not the US. It wouldnt be the first time different continents had different revisions of boards. We never saw the early rev KX7-333 whereas the US seemed to get mainly those for the first 2 months of production.

Im trying 7.5k resistance atm, just waiting for the prometia to chill down to boot point .

PiLsY.

[phantompunisher]

pin 7? have you found the resisters connected to pin 12? if so what are there values? if there 5k like on my board then anything under 10kish will result in full mosfet drain voltage flowing through to the memory no mater what the bios setting. good luck with it. im going to try to sleep soon

[PiLsY]

Right, more results...

Re-pasted my other results to compare.

No mods apart from bios :

2.5v = 2.51v
2.6v = 2.59v
2.7v = 2.68v
2.8v = 2.76v
2.9v = 2.51v
3.0v = 2.59v
3.1v = 2.68v
3.2v = 2.76v

Shorted diode, modded bios :

2.5v = 2.51v
2.6v = 2.59v
2.7v = 2.68v
2.8v = 2.76v
2.9v = 2.51v
3.0v = 2.59v
3.1v = 2.68v
3.2v = 2.76v

Shorted diode, 15k resistor across R110, modded bios :

2.5v = 2.59v
2.6v = 2.59v
2.7v = 2.59v
2.8v = 2.59v
2.9v = 2.64v
3.0v = 2.72v
3.1v = 2.76v
3.2v = 2.85v

Shorted diode, 7.5k resistance across R110, modded bios :

2.5v = 2.67v
2.6v = 2.75v
2.7v = 2.84v
2.8v = 2.93v
2.9v = 2.67v
3.0v = 2.75v
3.1v = 2.84v
3.2v = 2.92v

Back to square one on voltage adjustment via bios, but getting increased values. The way its going I think a 5k resistor should see about 3.05 to 3.1v and a 2k resistor should see the full 3.3v max.

Interestingly enough measuring vmem off the mosfet gives 3.33v during pre-chill (halted boot), but as soon as it releases and the bios kicks in it drops down to pre-defined bios values listed above. Very very odd...

PiLsY.