Printable View
hi
i' ve found the vdd mod for my is7-e
don't forgot, WOID WARRANTY
instruction:
you need to remove the R115 resistor (near the cpu fan connector, bottom and left)
use a 20k POT or 50k POT, set to 10k for start (it's very important )
increase resistance, increase voltage
measure the voltage from the mosfet near the agp slot, (below the PLL, right leg)
some measure:
without removed the resistor, r115 = 4.52 K ohms
unsoldered R115 = 10.0 K ohms
anyone can post a pic?
thx
Learn
Can anyone please confirm this?
Thanks
i happen to have a couple spare is7's layin around and plan on trying it to one of them this evening, ill post results..
hi,
i' ve finished the vagp mod, like the hipro5 way, it work OK, but the monitoring won't see the change.
Any results yet? :)Quote:
Originally posted by helsyeah
i happen to have a couple spare is7's layin around and plan on trying it to one of them this evening, ill post results..
sorry about the delay, senior projects are picking up and my last few nights have been shot to hell..
i performed the vnorth mod as described by learn, only i initally set my 20k pot to 4.35k which matched the resistance of R115. everything worked flawlessly. as directed, just increasing the resistance increases the vnorth voltage.
I will try to borrow one of my buddies digital cameras so i can post some picks, overall the mod is about the same level of difficulty as the vdimm mod. R115 is a bit of a pain to get to, but it is accessable.
As a side note, Learn, where is the trace you cut to do the vagp mod? i have hunted around quite a bit and have yet to find anyone else who has done the vagp mod on an Is7...
Awesome, how about OC results? I don't see why Vagp would need to be modded, since max voltage is already 1.65v, what is the point of increasing that unless you're OCing the agp? :confused:Quote:
Originally posted by helsyeah
sorry about the delay, senior projects are picking up and my last few nights have been shot to hell..
i performed the vnorth mod as described by learn, only i initally set my 20k pot to 4.35k which matched the resistance of R115. everything worked flawlessly. as directed, just increasing the resistance increases the vnorth voltage.
I will try to borrow one of my buddies digital cameras so i can post some picks, overall the mod is about the same level of difficulty as the vdimm mod. R115 is a bit of a pain to get to, but it is accessable.
As a side note, Learn, where is the trace you cut to do the vagp mod? i have hunted around quite a bit and have yet to find anyone else who has done the vagp mod on an Is7...
if you can post a pic on the back of the winbond w83303D, i can help you.Quote:
Originally posted by helsyeah
sorry about the delay, senior projects are picking up and my last few nights have been shot to hell..
i performed the vnorth mod as described by learn, only i initally set my 20k pot to 4.35k which matched the resistance of R115. everything worked flawlessly. as directed, just increasing the resistance increases the vnorth voltage.
I will try to borrow one of my buddies digital cameras so i can post some picks, overall the mod is about the same level of difficulty as the vdimm mod. R115 is a bit of a pain to get to, but it is accessable.
As a side note, Learn, where is the trace you cut to do the vagp mod? i have hunted around quite a bit and have yet to find anyone else who has done the vagp mod on an Is7...
this trace is between the winbond regulator and the southbridge,and at the bottom of the vddr mosfets.
edit: on the back of the motherboard
or follow the trace from this leg
http://www.thelab.gr/images/reviews/...gp/Vagp_12.jpg
http://www.thelab.gr/reviews/ic7mods_en5.php
it's a very small trace, it's not easy to do.
perhaps, it's possible to find an easy way.
vagp help for higher fsb (with the 1:1 mod) and also for higher fsb with the 5:4 and 3:2 ratio, and also for an highest 3D stability.
sorry for my bad english, it limit me... a lot :S
actually vagp usually helps a ton on memory stability and increases the maximum stable fsb oc. I have not run across any concrete examples, but i know people run 1.95 or more vagp without issue. I dont know about a "safe" voltage as far as going over 1.65 volts, but i imagine if we're volt-modding our mobos we're not necessarily doing the "safest" thing :D.
as for results, I havnt seen how far she goes with the 3:2 divider yet, but i will tonight or tomorrow.
The paticular is7-g that i performed the mod on has the issue where it cant boot when going past 280fsb with the 5:4 divide no matter what kind of voltage is thrown at the memory. Increasing the vnorth did not resolve that issue. I experenced the same thing with my IC7-Max3 as well. No amount of vnorth or vagp would let me use the 5:4 divider over 280 mhz.
I also have an is7-e that does not have any problem doing over 292fsb with the 5:4 divider with the memory that i have so its not my memory having problems with the raw speed. Actually, the memory is capable of 245mhz with 3.2v and 2-5-2-2 timings, at least at 1:1 mode. There is more stress on the memory with the 5:4 divider however. I currently am stuck at 235ish with the 5:4 divider, but im hoping that the vagp mod lets me get beyond that. Especially since i know my proc is capable of whats in my sig :D
learn, ill definately track down a digital camera then this weekend and get some pics posted... maybe if we can get some other mod-gurus in here we might find a better way to perform the vagp mod
ok, good idea :)
i have a problem over 280 5:4 with the pat enabled with my is7e, no problem without the pat.
bios problem ? (abit)
Thanks for the info guys. I tried pushing my Vagp to 1.65v (1.6v actual) and it helped prolong Prime's stability by a few minutes @ 300FSB on a 2.4C, 3:2 ratio, 2x 512MB Corsair PC3500 BH-5 chips, 3.01-3.09vdimm, 1.54VTT.
I believe my main problem is the PSU, the default rails are too low, 3.18-3.22v, 4.89v, and 11.68-11.80v.
Gonna do mod on all 3 rails when my Antec TP550 comes.
I'm concerned about my vcore, its 1.67v idle, but varies from 1.60-1.62v underload, is that normal? A full 0.07v sway from load to idle.
Vdimm idles at 3.09v, but sways from 3.01 to 3.04v underload... Is this all normal? Vcore is set to 1.7v in BIOS, but only reads max of 1.67v...
295FSB is completely stable so far, been doing 2 instances of prime95 for an hour or so.
Is there a "droop mod" for IS7? Seems like it helps the Asus boards a lot, the Vcore only varies by like 0.04v after the mod
i havnt run across a specific droop mod yet. there is a mod just posted today for vcore for the is7. However it is unclear to me why the is7 would need a vcore mod due to the insane vcores we have available... i would be interested in seeing if this specific mod did happen to deal with fairly drastic drop in vcore during load.
You get the same kind of variance in vcore and vdimm while underload as well?Quote:
Originally posted by helsyeah
i havnt run across a specific droop mod yet. there is a mod just posted today for vcore for the is7. However it is unclear to me why the is7 would need a vcore mod due to the insane vcores we have available... i would be interested in seeing if this specific mod did happen to deal with fairly drastic drop in vcore during load.
this is not normal, my vdimm maybe varies from 3.08 down to 3.07 at full load... so it would definately seem you have a weak PS... the true power should help a bitQuote:
Originally posted by HKPolice
Vdimm idles at 3.09v, but sways from 3.01 to 3.04v underload...
my vcore does drop at least .05 though with full load, and more when i crank the proc over 3.6ghz
I'm currently using an Antec TP430, but i'm selling it to my friend once my TP550 arrives. Problem is that the 3.3v rail varies from 3.18 to 3.22, what PSU do you use and how much does yours vary?Quote:
Originally posted by helsyeah
this is not normal, my vdimm maybe varies from 3.08 down to 3.07 at full load... so it would definately seem you have a weak PS... the true power should help a bit
my vcore does drop at least .05 though with full load, and more when i crank the proc over 3.6ghz
I did the Vdimm and VTT mods posted on AkibaPC's page here: http://www.akiba-pc.com/article.php?38.0 except that I used a 10K pot, and I think my resistance is higher than 1K.. but it works, when I set vdimm to 2.5v in bios, it drops down to 2.75v actual.
Running 295FSB @ 1.55vcore load, stable so far :)
One more thing, I see you're running an IC7, with 3:2 divider with over 300FSB, do you have FULL PAT enabled?? I am unable to select full PAT with 3:2 divider, no matter what FSB. Can't get full PAT beyond around 260FSB with 5:4 divider... and I have the same RAM as you
im actually using an antec smartpower 400 and it holds up pretty well... i did do the mod to get my 3.3 rail up to 3.5v. The only time i see my rails start to drop at all is when i go over 3.65ish ghz on the proc, then my 3.3 rail drops from 3.5 down to 3.4ish other than that, the 12v and 5 v rails stay within .02v of their idle values of 11.98v and 5.02v
Hmm, either my winbond monitor chip is old or inaccurate, or something.... I did monitor my 12v line with a multimeter before and it didn't vary more than 0.03v when booting up and prime. It was 12.14 - 12.17v, even tho winbond hardware doctor says 11.74v idle.Quote:
Originally posted by helsyeah
im actually using an antec smartpower 400 and it holds up pretty well... i did do the mod to get my 3.3 rail up to 3.5v. The only time i see my rails start to drop at all is when i go over 3.65ish ghz on the proc, then my 3.3 rail drops from 3.5 down to 3.4ish other than that, the 12v and 5 v rails stay within .02v of their idle values of 11.98v and 5.02v
actually im running my is7-e atm, the ic7 was an experiment to see how far i could push my procQuote:
Originally posted by HKPolice
One more thing, I see you're running an IC7, with 3:2 divider with over 300FSB, do you have FULL PAT enabled?? I am unable to select full PAT with 3:2 divider, no matter what FSB. Can't get full PAT beyond around 260FSB with 5:4 divider... and I have the same RAM as you
as to PAT, it was fully enabled when i was running @ 308fsb and i had pat all the way up to 275 using the 5:4 divider, but at 280fsb i hit the invisible wall where i could not post with the 5:4 divider.
on my is7 I actually have my GAT settings set to turbo, normal, auto, disabled, disabled. memtest and cpuz do not see pat as enabled, but i get over 150 mbps more unbuffered bandwidth on my memory by just having turbo enabled in GAT vs. having it set to auto.
So on your IS7-E PAT was fully enabled upto 308FSB with 3:2 divider? But that's not possible, cuz if you set the NB strap to 800PSB, it only does partial PAT unless you select F1 setting.Quote:
Originally posted by helsyeah
actually im running my is7-e atm, the ic7 was an experiment to see how far i could push my proc
as to PAT, it was fully enabled when i was running @ 308fsb and i had pat all the way up to 275 using the 5:4 divider, but at 280fsb i hit the invisible wall where i could not post with the 5:4 divider.
on my is7 I actually have my GAT settings set to turbo, normal, auto, disabled, disabled. memtest and cpuz do not see pat as enabled, but i get over 150 mbps more unbuffered bandwidth on my memory by just having turbo enabled in GAT vs. having it set to auto.
If you select PSB667, FULL PAT is enabled all the time, even if you select auto for GAT, but it only allows 1:1 and 5:4 dividers.
CPU-Z shows PAT as disabled if it's partially enabled. You need CTIAW to truly see the state of your PAT.
oops, sorry about the confusion, i meant that on my ic7 pat was fully enabled @ 308fsb...
308FSB with 3:2 divider? :oQuote:
Originally posted by helsyeah
oops, sorry about the confusion, i meant that on my ic7 pat was fully enabled @ 308fsb...
lol, yes with the 3:2 divider
Wow... thats what I'm aiming for, 310+FSB with 3:2 divider and FULL pat on 2-2-2-5 BH5s, gimmie your mobo!!!! :p
lol yeah the ic7-max3 did pretty damn well except for the 280fsb wall with the 5:4 dividerQuote:
Originally posted by HKPolice
Wow... thats what I'm aiming for, 310+FSB with 3:2 divider and FULL pat on 2-2-2-5 BH5s, gimmie your mobo!!!! :p
actually im hoping to get my is7-e up to 308 with the 5:4 divider... might take 3.5v on the memory, but ehh :D
then maybe ill be able to break 23k on 3dmark01
ok learn i got an image of the back of the is7. its of the area directly behind the south bridge and winbond chip... let me know if it isnt the exact area you need
http://helsyeah.shackspace.com/back%20of%20is7.jpg
Ok, heres the pics i promised for the vdd mod if anyone else wants to check it out
excuse the really crappy quality picture, tis what happens when a scanner is used instead of a digital camera :D
heres the location of the R115.
http://helsyeah.shackspace.com/vnorthmod.jpg
As described by Learn, remove R115 and solder two legs of a 20k pot into its place.
I set the 20k pot to 4.35k which was the resistance i read from R115. To raise the vdd voltage increase the resistance on the pot.
as with the IC7 the vdd is measured from the mosfet above the agp slot towards the socket.
http://helsyeah.shackspace.com/readvnorth.jpg
Sweeeeet, thanks man!!!!! :)
Still waiting for my Antec TP550 to come and do volt mods on that first... argh.... can't wait... fjla;sjklsfdakl;fsda :(
The only problem is, no1 knows if this actually improves stability, yet.... argh
thx for all the pics my friend.Quote:
Originally posted by helsyeah
ok learn i got an image of the back of the is7. its of the area directly behind the south bridge and winbond chip... let me know if it isnt the exact area you need
http://myweb.cableone.net/jeremybtmn...20of%20is7.jpg
it's the exact area.
i'll "photoshop" this
Learn, excellent, i look forward to your "chopping" :D
Ok, i tested to see how high i could go with the 3:2 divider and the vdd set to 1.65... I was able to reach 308fsb again as i had with my ic7-max3. It was stable through memtest86+ but i didnt load it into windows yet. Once Learn gets a diagram of what is needed for the agp mod then ill perform it and see how much further i can go...Quote:
Originally posted by HKPolice
The only problem is, no1 knows if this actually improves stability, yet.... argh
this is the small trace.
http://learn.chez.tiscali.fr/vagp-mod-is7.jpg
http://learn.chez.tiscali.fr/readvnorth-readvagp.jpg
sweet, thanks Learn, i'll be performing the mod tonight
I did follow the trace back to the front of the mobo and it went to one of the capacitors on the front of the mobo right below the south bridge... from there it seemed to go to an internal trace in one of the other layers... i would be curious where it resurfaced and the possibility for finding a better location to do the agp mod.... i wonder if theres a circuit diagram of the IS7 floating around anywhere...
hi
yes, a circuit diagram should be very helpful.
i've forgot,
increase resistance, increase voltage.
another thing, the bios monitoring seem won't see the voltage change.
Hey do you guys know where to measure Vdimm voltage? Mine is flucuating too much, I would like to solder a thick wire from the 3.3v rail directly to the MOSFET.
Plz help
Thanks
Also, where do I measure VTT voltage? I think I've got the right MOSFET for Vdimm.... but I'm gettin 3.34v when Winbond HWMonitor says it is 3.23v.... :(
i have no fluctuating problem with my is7e 1.2,
over 3.6v of vddr, i've a problem.below that, no
the bios monitoring seem won't report properly the real voltage.
you can measure the vddr from the two big mosfets below the dimms.(as usual it's the right leg)
take care, it's dangerous (do not short circuit the vddr and the 3.3v rail...)
i've soldered a wires from the vddr,vtt,vdd and vagp mosfets, it's a safe way to mesure the real voltage
edit:
for the vtt it's the smal mosfet (Q67) below the atx connector
So did the Vagp mod help?
I seem to be stuck.... I can pass memtest86 loops @ 295FSB 5:4 ratio, but prime95 causes the system to freeze/reboot after a while... Sometimes even when the CPU was 55C.
I have a fan blowing over the ram, and i even epoxied some heatsinks on the agp and vdimm mosfet, as well as the southbridge, with a fan blowing all of em.
Also, is there a capacitor mod for the IS7? Like the one here:
http://www.thelab.gr/reviews/ic7mods_en1.php
Thanks
chipset maxed? or your cpu?Quote:
Originally posted by HKPolice
So did the Vagp mod help?
I seem to be stuck.... I can pass memtest86 loops @ 295FSB 5:4 ratio, but prime95 causes the system to freeze/reboot after a while... Sometimes even when the CPU was 55C.
I have a fan blowing over the ram, and i even epoxied some heatsinks on the agp and vdimm mosfet, as well as the southbridge, with a fan blowing all of em.
it help a little.
jump from 285 5:4 to 290 5:4 (pat fully enabled, with super pi),
over 290 with the pat fully enabled, it crash quickly for me.
and jump from 255/256 1:1 (pat fully enabled) to 266 1:1 (memory limited )
it's possible, i thinkQuote:
Originally posted by HKPolice
Also, is there a capacitor mod for the IS7? Like the one here:
http://www.thelab.gr/reviews/ic7mods_en1.php
Thanks
but it help?
How did you enable full PAT? NB Strap 667? I can't even boot beyond like 260FSB 5:4 with full PAT :(Quote:
Originally posted by Learn
chipset maxed? or your cpu?
it help a little.
jump from 285 5:4 to 290 5:4 (pat fully enabled, with super pi),
over 290 with the pat fully enabled, it crash quickly for me.
and jump from 255/256 1:1 (pat fully enabled) to 266 1:1 (memory limited )
it's possible, i think
but it help?
Are your corsair sticks 256mb or 512, mine are 512
Also, how did you test for stability? Just superPi? I run 2 instances of Prime95 torture tests for at least an hour.
yes, NB strap 667.Quote:
Originally posted by HKPolice
How did you enable full PAT? NB Strap 667? I can't even boot beyond like 260FSB 5:4 with full PAT :(
Are your corsair sticks 256mb or 512, mine are 512
it's 2x 256mb sticks.
it's a bench report, not stability test.Quote:
Originally posted by HKPolice
Also, how did you test for stability? Just superPi? I run 2 instances of Prime95 torture tests for at least an hour.
i'm using prime95, seti@home, game...
like this?:D
http://learn.chez.tiscali.fr/3.0C/prime-4010-load.png
What motherboard revision is your board?Quote:
Originally posted by Learn
yes, NB strap 667.
it's 2x 256mb sticks.
it's a bench report, not stability test.
i'm using prime95, seti@home, game...
like this?:D
http://learn.chez.tiscali.fr/3.0C/prime-4010-load.png
IS7-E revision 1.2
I just did the Vagp mod, and it doesn't work. I tried varying the resistance from 10k upto 100k, but the Vagp voltage doesn't change!! :(
I know I did the mod right, because when I unplug the VR, it doesn't boot up.
Oh ya, my board is revision 1.0. Learn, where did you get the Vagp mod?
hi,
sound strange, it work for me.
you have a good ground? and VR?
where have you check the voltage? from the mosfet? bios?
myself, it's the apadated hipro5 ic7 mod, it must working.
Yep, I have a wire soldered onto the the back of the ATX connector's ground. I check the voltage from the mosfet, I have wires soldered onto VDD, Vagp, Vdimm, and VTT points.Quote:
Originally posted by Learn
hi,
sound strange, it work for me.
you have a good ground? and VR?
where have you check the voltage? from the mosfet? bios?
myself, it's the apadated hipro5 ic7 mod, it must working.
Are you sure that is the right trace to cut??
yes i' m sure.
(computer turned off)
can you check the resistance between the vagp mosfet (right leg) and the ground( VR set to 0 ohms), and if the value change when you increase the resistance.
Nope, the red wire is connected to ground it seems.Quote:
Originally posted by Learn
yes i' m sure.
(computer turned off)
can you check the resistance between the vagp mosfet (right leg) and the ground( VR set to 0 ohms), and if the value change when you increase the resistance.
I just checked everything again, and recut the trace. Seems to work now, but it takes a lot of tweaking to get it right. I think the difference between 1.75v and 1.91v is less than 1k ohm? Anything over ~1.95v, and the board won't boot.
Seems to help stability a bit, I can do 310FSB 3:2 ratio memtest86 without problems now, gonna do more testing later on.
cool, it work
i've never measured the resistance/voltage. yes, it's possible, it's less than 1k ohm. if you want you can use a lower VR.
it's the same here, i can't use more than 1.9v.
I did the VDD mod to my IS7-G motherboard but I think that I broke it :(
I was looking the hipro5 mods and this thread and I havent any problem until VDD mod. I hope the problem is a wrong potentiometer (5Kohm) Im not sure because Im at work now and I cant see which pot I used.
I placed a 3300uF capacitor in the corner of the northbridge like hipro5. I measured the default voltage and Its 1.45v ¿to much low for default?
I was reading all this thread and Learn and HKPolice have (for me) impressive OC. For me 270MHz FSB 1:1 with full PAT is simply impossible. When i read that somebody can get 290 or over at full PAT I cant believe ¿why can't I ?
I have P4 2.4C SL6WF Week 21 Like HKPolice and AData Vitesta 2x512 DDR500. I know it have worse timmings but higher FSB.
Please if you can, post some screenshots with sandra's Bandwidth test ( It must be aronund 6.5 - 7 GB/s with full PAT at 290MHz)
I found the problem, It was the vtt mod that I adjusted bad the potentiometer to the max resistance when this must begin from 0 ohms... Now the problem is resolved and the motherboard is working without problems.
My vnorth (VDD) now is 1.75v with a 3300uF cap to stabilize the voltage. 3.3V rail of my Antec Truecontrol is 3.55V, VDDR is 3.10v. I was testing VTT but I cant see any variation of voltage.
Now I have better stability and OC, but far from the people here.
I can boot at 290 1:1 but no windows stability. 285 1:1 looks stable but I didn't burn test like prime95. I tryed 280 5:4 with PAT but it don't works :(
Maybe I need higher VDD ?.
Im curious, what memory are you using, what is your vdimm set to and what timings are you using? tweaking those could solve the 280 5:4 problem. The type of chips used on the memory my also not like the 5:4 divider.Quote:
Originally posted by MDLSoft
I can boot at 290 1:1 but no windows stability. 285 1:1 looks stable but I didn't burn test like prime95. I tryed 280 5:4 with PAT but it don't works :(
Maybe I need higher VDD ?.
There is the chance that your motherboard is not capable of running over 280 with the 5:4 divider. I have one Is7 and an IC7 that cannot run over 280ish with the 5:4 divider. This seems to be a limitation of the northbridge and perhaps increasing vdd will help. I never tried over 1.8 vdd and it did not seem to help very much with getting over 280 with the 5:4 divider.
Quote:
Originally posted by MDLSoft
I did the VDD mod to my IS7-G motherboard but I think that I broke it :(
I was looking the hipro5 mods and this thread and I havent any problem until VDD mod. I hope the problem is a wrong potentiometer (5Kohm) Im not sure because Im at work now and I cant see which pot I used.
I placed a 3300uF capacitor in the corner of the northbridge like hipro5. I measured the default voltage and Its 1.45v ¿to much low for default?
I was reading all this thread and Learn and HKPolice have (for me) impressive OC. For me 270MHz FSB 1:1 with full PAT is simply impossible. When i read that somebody can get 290 or over at full PAT I cant believe ¿why can't I ?
I have P4 2.4C SL6WF Week 21 Like HKPolice and AData Vitesta 2x512 DDR500. I know it have worse timmings but higher FSB.
Please if you can, post some screenshots with sandra's Bandwidth test ( It must be aronund 6.5 - 7 GB/s with full PAT at 290MHz)
i've also 1.45v from default vdd voltage.
vtt mod help in 1:1 mod.
PAT fully enabled, n/b strap 667, 5:4, 2225, street racer, enhanced, 5 clocks, D, D
http://learn.chez.tiscali.fr/3.0C/290-5-4-PATE.png
PAT disabled, n/b strap 800, 5:4, 2225, auto, enhanced, 5 clocks, D, D
http://learn.chez.tiscali.fr/3.0C/290-5-4-PATD.png
PAT fully enabled, n/b strap 667, 1:1, 2225, F1, enhanced, 5 clocks, D, D
http://learn.chez.tiscali.fr/3.0C/mem-261.png
I created a small webpage for all the is7 mods...hope I got all the info right. Enjoy.
http://www.1337info.com/stevexx/voltmods/is7/is7.html
It takes a lot more than 10ohms for the VDD, I used a 50K VR and had to turn it up about half way to get 1.9v
As for the vdimm, 1K is also not enough if you wanna push more than ~3.2v or so, use a 10K VR
There is also a VTT mod you have to do
updated.
Hahaha I never claimed to be the "king of the world" :p: just that 1K usually isn't enough. If you're extreme enough to do these kind of mods, you're most likely running BH5 or CH5 ram which require ~3.3v to run @ max safe speeds.Quote:
Originally posted by bh2k
updated.
would it be possible to pay someone to do this to my board? :D
my hands are FAR from steady, and time is an issue as well :(
pm me :D
Pm hell-fire, he can do it for a price...