NF7 rev2 VTT mod need it so bad...

I certainly like not having to cut traces, but what if you want to run a high Vdd??

You certainly cant run a 2v Vdd under these circumstances, or even 1.85v+......unless of course you are running a Vdimm of 3.7v.

Granted, most people are not going to be running a Vdimm of more than 3.4v for daily use....or even lower...but not being able to run a Vdd above 1.8v is quite a loss.

Myself, I am not entirely sure that such a high Vdd is needed...but I like to be able to go to 2v if I want.

Anyone have input on whether a Vdd of more than 1.7v is a necessity or not?

Regardless, thats one hell of a nice find tictac!! :toast:

i tested the mod already on my dead board

some how it manage to power on once again... here i measure with my DDM....

12V @ 13.5V
3.3v @ 3.9V
Vdd @ 1.63V
Vbt @ 1.63v
Vdimm @ 3.9v ( I had done vdimm = i/o mod)

well if u have very high vdd voltage(1.8 or > 2.0V )... there is still way to do it... cut the wire and put trimmer on the white line.....

start from o ohm... raise the resistance to lower the Vbt voltage... :D

Quote:

---

Originally posted by tictac
i tested the mod already on my dead board

some how it manage to power on once again... here i measure with my DDM....

12V @ 13.5V
3.3v @ 3.9V
Vdd @ 1.63V
Vbt @ 1.63v
Vdimm @ 3.9v ( I had done vdimm = i/o mod)

---

3.9V on your mem? Nice ...
Is this measured on the vdimm mosfet or is it the bios value?

DDM ... digital multimeter...

on mosfet

sorry but connecting the +3.3v line to the VDDR power directly power up my VTT to an half of the +3.3v voltage, so with this mod no VTT mod is required, or i'm wrong.... i have measured VTT under the dimm slot, not on the mosfet...

Dirty_Punk.... is it VDIMM mod to 3.3V power conector, power up VTT to an half of the +3.3v voltage ?
I do it to,.. but it isn't half of the +3.3v voltage... or maybe I measured it wrong ?

what about, if I use V I/O, add some fix resistor to reduce it, and add trimer (pot) to make it half of VDimm, than take it to VBT pin..?

Quote:

---

Originally posted by Dirty_Punk
sorry but connecting the +3.3v line to the VDDR power directly power up my VTT to an half of the +3.3v voltage, so with this mod no VTT mod is required, or i'm wrong.... i have measured VTT under the dimm slot, not on the mosfet...

---

try measure on the mosfet or the R750 leg facing the near lever... the vtt is not half of vdimm voltage after vdimm mod... :(

u can also measure vtt on ce 48 & ce53 Rubycon caps.... :)

Also, you can measure Vtt at the DIMM itself... use empty socket..pin 1 is Vref...aka..Vtt. Pin 1 end of Socket is at edge of PCboard end of DIMM. Pin 1 side is the side closest to the CPU socket. So pin 1 of DIMM1 is the pin closest to the Winbond volt/fan monitoring chip between DIMMs and CPU socket. Pin 7 is VDimm on same DIMM socket. Sharp multimeter probes are a plus here ;-)

That VDD "boost" mod in your picture TicTac shorts the 5Volt rail to something in the VDD circuit.... I would NOT recommend it.

I would also not recomend the other shortcut method... shorting 3.3V rail to Vdimm, or shorting VDD to Vtt... since those are driven by switching regulator circuits that are going to be in an unknown state.... the outputs driven higher than inputs. Could lead to damage longterm, and even component failure and unknown consequences.

Provided the 3.3V rail is held .3-.4V higher than the intended Vdimm voltage, the Vdimm regulator circuit works just fine and with complete stability. The current capacity is WELL beyond what the DDR's need...hence the Vdimm mosfet never really gets hot.

If you have voltage instability....wild fluctuations....in you Vdimm when set higher by the mod, it means your 3.3V rail is simply NOT high enuff. Go into your PSU and raise it, or use the 3.3Vsense PSU mod.

Also an issue, in raising 3.3V rail with PSU mods, most PSU's have overvoltage protection levels, and it's possible to reach high enough to start triggering this....leading to weird fluctuations in 3.3V rail. Im fortunate in that the Fortron530 has 3.3-5-12V adjustment pots AND pots on the overvoltage levels. Some dont have this and you may not be able to get your 3.3V rail to 3.6-3.7 with stability, especially at high operating temperatures after hours of gaming. You will NEED a serious power supply when pushing the NF7 this hard.... 530W Fortron or 550W SuperFlower are my recomendations (both $70-80 at Newegg, Directron, etc)

It's a tough balance, having VDD adjustable is nice, though I havent tested extensively on the impact it has on overclocking.... but I have the Swifty on the NB to handle up to 2.0Vdd. Question is, does the Vdd level impact the Northbridge's end of the memory transfer levels. If you goose Vdimm (and therefore VDDQ) then the rams are working at a voltage level above normal.

The DDR signals all go to the NB chip, so does it also need to be goosed..or does the NB have seperate power pins ALSO powered by Vdimm (vddq) to set the reference level at which the DDR transceivers within operate? Without the NB spec sheet, this is unknown.

When I pull this sytem apart...with mods on everything, ill test extensively the impact of Vdd mod vs Vdimm/ram speed.

just to add up another thing:

The Dirty Punk vdimm mod is as same as normal vdimm mod... except
- the normal vdimm mod get their voltage from mosfet
- the dirty punk vdimm mod get their main voltage from i/o volt...

And there is no others circuit after the Vdimm mosfet to the dimm slot...

Measure the resistance between Dirty_Punk dimm solder point to the Vdimm Mosfet Leg.... The resistance is ZER0

Let me explain....

By shorting the Vdimm line to the 3.3V rail, you place the output mosfet of the Vdimm circuit in a battle with the output mosfet of the 3.3Vrail in your PSU. The Vdimm circuit mosfet now has it;s output driven to a higher/equal voltage than it's input...and they dont like that. It may never be an issue, but it may also end in damage to the transistor.... there really is no reason to bother sans for saving the $3 for a POT to properly mod the Vdimm.

Yes it is very stable, but so is Vdimm mod with a pot provided you raise your 3.3V rail to make room for the raising of Vdimm. You need to provide "some" headroom. Provided this is done, you'll see rock solid stable Vdimm from normal mod. if you see fluctuations it is because the 3.3V rail is too low for the regulator to work properly.

I simply recommend you stick with the POT method, no point in risking damage needlessly. Im not even worried about the motherboard, $60 refurb on Newegg is cheap... but $400 in BH5 ram, I dont think anyone should be risking with the simpler direct to 3.3V mod. We also dont know how the Vdimm(vddq) issue ties into the Northbridge chip, and what affect shorting to 3.3V rail might have longterm on that chip. Modder beware is all Im saying.

Quote:

---

Originally posted by tictac
Bachus_anonym

I just check my dead board and compare it ... there are missing mosfet in ur board...

i'm confused here... do u remove it .... what mod is that for?

here my dead board(NF7-S Rev2.0) mosfet near the sata connector :confused:

---

WOW! u got me on that one... i have no clue! ATM, i have three board on hand... a good one (it has mosfet), "dead" one (no mosfet) and really dead also with mosfet...

i wonder what is that mosfet for ???

EDIT: tracing it leads to battery socket... hmm

About that mosfet I have it too

About increasing Vdd = more OC on FSB... can´t reach more FSB with high Vdd´s. 243 MHz @ 1,6V or 2V...

About Vtt mod...
Can´t it be done by the 3,3V line + pot and tune it to the desired voltage? Or this have the same effect of dirty_punk Vdimm mod?
I mean this effect... :D

Quote:

---

By shorting the Vdimm line to the 3.3V rail, you place the output mosfet of the Vdimm circuit in a battle with the output mosfet of the 3.3Vrail in your PSU.

---

And thanks uwackme for the tips...

btw, I have solid Vdimm, from the typical Vdimm mod, and Vdimm mosfet doesn´t get really warm.

Keep the good work guys :)

Some NF7s don't like high vdd, some do. My first didn't, my second did.

Can someone post results before and after the mod? I'd like to know if this is really going to help.

Quote:

---

Originally posted by HaLDoL
Can someone post results before and after the mod? I'd like to know if this is really going to help.

---

http://www.xtremesystems.org/forums/...threadid=33240

Quote:

---

Originally posted by uwackme
Also, you can measure Vtt at the DIMM itself... use empty socket..pin 1 is Vref...aka..Vtt. Pin 1 end of Socket is at edge of PCboard end of DIMM. Pin 1 side is the side closest to the CPU socket. So pin 1 of DIMM1 is the pin closest to the Winbond volt/fan monitoring chip between DIMMs and CPU socket. Pin 7 is VDimm on same DIMM socket. Sharp multimeter probes are a plus here ;-)

---

i have measured on the DIMM socket the VTT, not on the mosfet... so on the dimm socket i measured over 1.7v with over 3.4v of VDDR, but this with power off motherboard (also VDDR can be measured with power off motherboard with this mod, with normal mod not because you measure always aroud 2.85v)

howether with +3.3v directly connected to dimm pin you get a much more stable VDIMM
with both at 3.2v i get about 3-5mhz more with this mod VS normal VDDR mod

When the Vdimm regulator is working correctly...ie: 3.3V rail is high enough over Vdimm desired...I get +/- .01V stability on the 4 boards I've fiddle with this week. All 4 systems have Fortron supplies with 3.3V adjustment inside though.

On the before/after results.....D23 3Dfire bios, Fortron 350W supply, 3.3V rail @ 3.6V, normal Vdimm mod, cut/pot Vtt mod, NO Vdd mod on this board yet, so stock 1.7V setting.

PC2100 valueram: 3.0Vdimm 1.5Vtt

Before 2x 256M 200Mhz, 11,3,3,2.5 cpuint enb, total stability

After 2x 256M 210Mhz 11,3,3,2.5 cpuint enb, total stability.

HyperX PC3500 CH-5 chips: 3.3Vdimm 1.65Vtt

Before: 1Gig 220Mhz 11,3,3,2.5 cpuint enb, total stability

After: 1Gig 230Mhz 11,3,3,2.5 cpuint enb, total stability.

So it's definitely worth doing in my book. not the panacea I was hoping for with CH-5 chips, but still better.

Ill have D10 bios results on a fully mod'd board with 530W supply later today. This board already does 220Mhz 11,3,2,cas2 cpuint enb with 1Gig....3.4Vdimm, 1.85Vdd, and no Vtt mod.

Total stability meaning memtest83 v3.0 and M$oft memory test running errors free for 20 passes on extended tests.

Quote:

---

Originally posted by uwackme
Let me explain....

By shorting the Vdimm line to the 3.3V rail, you place the output mosfet of the Vdimm circuit in a battle with the output mosfet of the 3.3Vrail in your PSU. The Vdimm circuit mosfet now has it;s output driven to a higher/equal voltage than it's input...and they dont like that. It may never be an issue, but it may also end in damage to the transistor.... there really is no reason to bother sans for saving the $3 for a POT to properly mod the Vdimm.

Yes it is very stable, but so is Vdimm mod with a pot provided you raise your 3.3V rail to make room for the raising of Vdimm. You need to provide "some" headroom. Provided this is done, you'll see rock solid stable Vdimm from normal mod. if you see fluctuations it is because the 3.3V rail is too low for the regulator to work properly.

---

i have tryed both and with very different results..

with +3.3v at 3.5v and VDDR at 3.2v with normal VDDR mod i can get max to 250mhz only sandra benchable 2 stick of V-Data PC3200

with the +3.3v at 3.2v but directly connected to vdimm i have a rock stable vdimm (and this isn't so easy with NF7-S and dual channel) and can bench sandra at 255mhz (probably a bit over) and 250-252 s.pi with the same sticks of ram

i think that the power circuit of the VDDR on the NF7 when DC is enable has a great hit on stability... so bypassing it give much more stable VDDR...

howether i'm can't recommend this mod for 24/7h use because as you say we can't know exactly if the mosfet like or not this situation...

when u gona do Vbt mod Dirty_Punk?

it worth a try... :cool:

:toast:

My frens manage to Hit 260MHz FSB with your Vdimm mod
http://downloads.modlabs.net/bios/pic/asus.jpg

right, my try is without VTT mod... i must try with that, but now i haven't yet the V-Data memory so i must try with other ram...

if you set the VTT more than 1/2 or the VDDR, like 1.7v instead of 1.65v, can you have better results on overclock or not???

howether the biggest problem of the NF7 in overclock IMHO is the chipset... isn't strange that DFI with more VDD get more mhz and NF7 with only 1.5v get the max and more VDD doesn't give anything???

and another thing.... if i'm right MSI has done a new nforce mobo with great voltage options in bios, like 2.3v vcore, 3.3v vmem ecc ecc... and it have a voice that seems to be a "secondary VDD voltage" that can be overvolted (i'm not sure), can be this the problem of the NF7?!? that NF7 doen't overvolt this "secondary VDD voltage" together with normal VDD and DFI do this????
it's only an idea and doesn't know if it can right or only a stupid idea :D

ya the is other mosfet beside the vdd mosfet . What mosfet is that? is that the secondry vdd voltage....

which MSI board is that?

emmm... my fren with a screen shot 260MHz 2-2-2-11 timing

now his board is dead after trying to do the Vbt mod....

Vbt=Vdd mod

:( :( :(

anyone have same issue here?

hoether only fot the overclock test the easyest mod is connecting +3.3v to vdimm, no difficult soldering, no warranty void..

better dont do my vdd=vbt mod

i will remove it then.... :(

another dead story here:(mistamasta)
http://forums.amdmb.com/showthread.p...5&pagenumber=4