8RDA/+ memory mod?

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

I thought I remember seeing a thread on this topic already but I just searched the index and can't come up with anything. Is this mod. being worked on? If so, where is the thread(if there is one)?

[Fewture]

It's here: http://www.xtremesystems.org/forums/showthread.php?s=&threadid=6199

You need to cut the pin or solder it off the mobo like the BD7-II. Riscy stuff. MrIcee killed his poor 8RDA while attemting this mod..

Die-Hards only.... (so go for it! :D )

[Jeff]

Anybody here have a professional surface mount soldering station and enough experience to do this mod? If so, how much should I include with my board? :D

[muzz]

I wish there were ANOTHER way to do this mod , I'm pretty good with a soldering iron, but that is a small area/small ic .., it is pretty risky.
I only have an Ungar 45w iron ( could get new element/tip combo)with a flat tip right now, and maybe could do it, but it is a tough one with the above mentioned eqpt.

muzz

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

Can someone confirm that pin 22 of the W83301DR-0 is the pin that goes to the big fat trace with 'U29' stenciled on it? I just want to make sure I'm looking at the right pin/trace.

I'm going to desolder the capacitor CE38 and see if this may be an easier path to go.

Also, what about just breaking the trace and soldering wires there? Does anyone know if the trace for pin 22 also goes under the chip(to the back side of the board)?

Just thinking out loud guys... :confused:

And just for my info... we are trying to drop some voltage here so the chip will 'correct' itself by adding some Vdimm right?

Ok, just desoldered CE38 and verified that the positive side is on that same 'fat trace' that pin 22 is on. It also goes to Q32.

So either of those would be good solder points for one side of the variable resistor. I wish I knew for certain whether or not there was another trace going from pin 22 under the chip. If there is NOT, then it'd be reasonable to assume you could just carefully cut the 'fat trace' and then flake off enough of the protective PCB coating to solder a wire there. Then the other wire from the resistor could go to that capacitor or Q32.

Again, just thinking out loud.



Ok, so pin22 is just the sense line for monitoring Vdimm. Hmmm... If that's so, then I doubt there would be any trace under the chip for pin22. If you look at the PCB, it looks like that 'fat trace' is the voltage plain for Vdimm and the trace going to pin22 is just for sensing. Why would you need to send that voltage under the chip? I guess there could be a reason but I'm going to bet there isn't.

So a 2k variable resistor set at 0 would be an ok way to start this mod right? (0 resistance would equal a short which is what the existing trace is acting like right?)


The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

Ok, I just carefully scraped the protective coating off the fat trace and soldered a wire to it... no problem. I want to breaking the trace and soldering in a resistor except I want to get some feedback and make sure I'm on the right track.

From reading about Randi's experience, it seems the worst that could happen would be ~3.22v being fed to the memory modules... which I'm almost 100% sure my Corsair could handle for at least a short time.

I'm 90% sure I'm on the right track. I just want a little feedback before I continue.

That and I have to go do the Christmas thing this evening. If there's enough feedback when I get home, I'll give this mod a whirl late tonight.

[muzz]

Good Luck Bro !!!! I will be paying CLOSE attention to your findings. :banana:

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

Ok, seems like it works. Only problem is that the BIOS does NOT show the increase but if you measure your voltage at Q32 you see the increase.

Like I mentioned above, my first(and only so far) attempt was with a 2k variable resistor. I'm pretty sure this is way overkill since you only need ~200Ohms to get 3.2v(which I think is close to the maximum Vdimm according to Randi's findings). I'm going to pick up a 200-300Ohm variable resistor and do a little more testing as to what resistance give what voltage.

And I did it like I explained above. I never touched the actual chip... I'll try to take some pictures tomorrow but since it's Christmas... I'm not sure when I'll be able to.

BTW, I do NOT have this board running in a system yet so I have NOT tested this mod fully. I don't see there being any problems but you need to take all the information here on an "at your own risk" basis. ;)

[muzz]

Thanks alot Jeff and if it turns out correctly congrats bro!!!!!!!:toast: :toast:
You obviuosly measured b4 and after correct? And from the correct point?
Put it this way MBM or the USDM that comes with it show the voltage to vdimm.. have you checked those yet?

EDIT the mbm check for stupidity...........

muzz

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

The only real measurements I've made have been on the mother board itself. I don't have this thing up and running yet and actually... tonight is the first time I've fired it up. Pretty funny that this board has NEVER run stock. :D

I'll try to load Windows and use MBM or USDM to verify the voltage but seeing as how the BIOS isn't reading the correct voltage... I'm not 100% sure these programs will either.

Just off the top of my head... maybe there is another trace to pin22 under the chip... the trace that is responsible for measuring Vdimm. ;) If that's the case, I can live with having to set my Vdimm using a DMM as opposed to a software measurement.

p.s. I've tested(in the BIOS only) up to 3.12v. I'm pretty sure the mod. is working just using the amount of heat my DIMM is producing. At 2.5v they don't too warm but at 3.12v they get a little warmer... a little quicker. ;)

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

[muzz]

Sweeeeeeeeeeeeeeeeet bro!!!! Yeeeeehawwwwwwwwww I just looked at that trace on my board, and that is a nice ,big, fat sucker with plenty of space!!!

Congrats Jeff :toast:
I started a thread at AOA and AMDMB linking this thread.

Happy Holidays bro.

EDIT: Try this at YOUR OWN RISK!!!!!!!

muzz

[muzz]

I see the same folks following this thread... how many of ya have opened your cases and looked at it???
heheheh:toast:

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

Here's a fuzzy picture of what I did...

Notice how I had to cut the 'fat trace' that was connected to pin 22. Just use a good deal of caution when doing something like this because there are multiple layers of PCB and you don't want to cut down to the next layer. :(

[muzz]

Well personally I know exactly what you did bro...... I thought about this, but said to self " self ya know that there has to be a place on the board that has a trace"... then responded to self." self there could be something attached on a lower layer - not ready for that chance stupid....." lmfao....... Well bro you seem to have found out that there was nothing( I couldn't afford to make that mistake, cuz this is really MY only machine.... there is another, but it's not mine ), and it is a really ez mod, the amount of space that is there is unreal... alot easier than the lift the leg thing. heheh

Already going through my variable resistor jar to find a <1k.. hopefully less than 500 ohm, in a flat plastic style....:toast:

Good Job Jeff

muzz

[Jeff]

Just be sure you can handle blowing up your board. :( I really have not tested this fully and while I don't believe anything bad will happen to my(and your) board, I've been known to make mistakes... BIG mistakes. :D

I'm sure if I'm barking up the wrong tree then either my system will fry... or someone else a hell-of-a lot smarter than me will correct me... or both. :D

[muzz]

Well I don't have the style vr I wanted, but I do have a 20 turn 1k.... which is good, I'll use the bottom 2 legs and glue the sucker to the board with some crazy glue( standing up )........ heheh. I will not be doing this tonight, as it's Midnight, and it's Xmas eve.. going to bed by 12:30 tonight.

[muzz]

I know me, if I start this now I will NOT go to bed at all..... I will be working on this, and then testing it all morning......... heheheh

[muzz]

Where is the vdimm measuring point?

[DaGooch]

Great work Jeff! Trying to think of a way of doing this mod by not cutting the trace here. Maybe desoldering CE38 and putting in a variable inline before or after CE38.

[muzz]

Originally posted by Jeff
Just be sure you can handle blowing up your board. :( I really have not tested this fully and while I don't believe anything bad will happen to my(and your) board, I've been known to make mistakes... BIG mistakes. :D

I'm sure if I'm barking up the wrong tree then either my system will fry... or someone else a hell-of-a lot smarter than me will correct me... or both. :D

Bro I would NEVER blame someone for showing me showing..... if I pick up the iron and start working. It becomes MY problem not yours.........:toast:

[muzz]

I see that this mod is slightly different than the bd7 mod...according to what Randi said( correct me if I'm wrong Randi. or anyone ) you were cutting pin 22 off the board ( or disconnecting it..... however you got pin 22 disconnected from the circuit), and installing a vr inline. IE In series to connect the original path up THROUGH the resistor( in series)...... so where as they wanted max resistance( they had a shunt- in parallel with the existing circuit to ground), we want least resistance to start with so the circuit "LOOKS" like it was never touched....... correct? Then increase resistance slowly........

I just wanna make sure everyone is on the same page here, and that this is the way it should be done.

EDIT: I just thought about that mod from the bd7... basically what was done is they just added a "Safety resistor- (for def load) and went in parallel with that, but it was not necessary to have that there.. I remember( i think ) they just added that for saftey.. basically the vr they used would have gone str8 to ground if that resistor wasnt put there....... I noticed they have a bypass cap going to ground there.
So basically we are doing the same thing, but no extra resistor there. So my question is why are we starting at MAX resistance... when if the resistor wasn't in series and nothing was touched there would be no resistance?

Am I missing something here?

[Bushboy]

Originally posted by DaGoochMeister
Great work Jeff! Trying to think of a way of doing this mod by not cutting the trace here. Maybe desoldering CE38 and putting in a variable inline before or after CE38.

Just how hard is desoldering a pin? Is it a risky thing to do, I mean I know anything like that is but how likely is it to go wrong? Obviously I'd prefer that than cutting a trace because of obvious reasons, and also because I didnt buy my board from a store so no RMA lol.

[Jeff]

Originally posted by DaGoochMeister
Great work Jeff! Trying to think of a way of doing this mod by not cutting the trace here. Maybe desoldering CE38 and putting in a variable inline before or after CE38.

That was my first thought and why I desoldered it first. Only problem is pin 22 shares the positive side of the cap and basically that whole big fat trace. Taking the capacitor out doesn't break the trace/circuit.

[Jeff]

The memory mod described in this thread will VOID YOUR WARRANTEE!!! Only attempt the mod if you understand this and are willing to risk damaging your board.

@muzz:

You do NOT want to start out at maximum resistance since this is being added in line(series). If you start out at maximum, then you will be stealing the maximum amount of voltage away from the chip and it will compensate by adding the maximum amount of voltage needed for correction.

Basically you want to start out at 0 Ohms and then add a little at a time and see how much of a gain you get at Q32(the leg that's connected to the 'fat trace').

If measure there without doing the mod, you'll see your 2.5v(or whatever you choose in the BIOS). After the mod, if you set to 0 Ohms, you will stilll see this default voltage(nothing added). Now when you increase the resistance a little you will see it go up by .1, .2, etc. CAUTION: I HAVE NOT TRIED CHANGING THE RESISTANCE WHILE THE BOARD HAS POWER! I don't think anything bad will happen but I remember reading thread where Randi said some board didn't like having something changed so I've played the safe road here.

Oh, and as for adding a safety resistor... I don't think you'd want too since the safest resistance we want is 0 Ohms. If you add a resistor in parallel then you will just be increasing voltage.

[muzz]

Well Jeff thats what I was saying bro..... if you look at the bd7 discussion, it is stated to start at the maximum... at least thats what I thought it said. Look at the thread for the bd7 mod, I'm pretty sure it said MAX. which to me makes no sense....UNLESS you are purposely trying to get the error amp to correct itself.

BTW I setup my vr to 0 ohms last night, then was reading the bd7 discussion. thats why I asked. :)

Merry Xmas bro.. I'll be on in awhile.:toast:

[Jeff]

So how'd you make out muzz?

I'm loading WinXP right now and I will hopefully do some testing either tonight or tomorrow.

[muzz]

Been fuggin around the whole day today. been on/off. getting alot of phone calls and chit.
I'll be doing it tommorrow.
Plus I was doing a little orbage too.. got a little higher score tonight.... snowing like a bastardd here, ever see a garbage bucket packed with snow water......... lol
Still can't keep it cold enough because I have the 9700 and the cpu in the same loop both oc'd, running hv ..... don't take long to get the temps up to 21-25. which blows!!!!

[DDTUNG]

I've just read the entire thread and am confident that you are on the right track. The STR1sense pin is originally connected directly to the Vmem output and what you have done is introduced a voltage drop across the output and STR1sense which will cause the IC to increase its output. The only thing that remains is to dial in the resistance to get your desired voltage.:) And you are absolutely correct in starting off with 0 ohms and going up gradually.

DDTUNG:cool:

[Jeff]

Cool! DD's stamp of approval! :toast:

After fooling around a bit with Vdimm, I am pretty confident that what the BIOS and MBM is reporting as Vdimm is wrong. I noticed people said that the 2.9v setting in BIOS only gives 2.880v... well that's correct if you look with a software program but measuring it with a DMM shows it's really 2.906v. Another strange thing is if you are looking at Vdimm with USDM, you notice it floats around a bit... but if you have a DMM hooked up to it it is rock solid? :confused: I don't know what's going on there.

[muzz]

I dont either...... but I really don't care much about it........ :D
Glad to see you stop by with your stamp DD.

Thanks Man.

muzz

[Bushboy]

"Thanks all for the input. I will ask about the function with our design team. But I cannot make any promises."

-Epox Tech from amdmb thread.

So it payed off :) Hopefully they will add some VDimm, and make us all very happy people.

[muzz]

We shall see.

[Jeff]

Finally got somewhere after doing the Vdd(chipset) modification(here (http://www.xtremesystems.org/forums/showthread.php?s=&threadid=7215)). ;) I can now say that FOR CERTAIN, the Vdimm mod works.

[muzz]

Dude thats awesome....... maybe this board isn't the piece of crap alot are saying it is........;)

[Jeff]

Originally posted by muzz
Dude thats awesome....... maybe this board isn't the piece of crap alot are saying it is........;)

I've heard people bash Epox before and that doesn't really matter to me since I still think they offer some quality boards that perform pretty damn good(for the most part). ;)

BTW... I'm using Corsair XMS memory for all the Corsair bashers out there. :D

[Holst]

I think my old samsung is gona be getting the 3.5v treatment once I get home.

hell... you only live once :D

[RoydRage]

Originally posted by Jeff
Finally got somewhere after doing the Vdd(chipset) modification(here (http://www.xtremesystems.org/forums/showthread.php?s=&threadid=7215)). ;) I can now say that FOR CERTAIN, the Vdimm mod works.

Jeff,

Congrad's on the mod.... If you get a chance since I see you have long leads... to the VR.. Could you measure the VR @ where you have it set... I would like to use just a resister... Since you can't monitor it from Windows.... I'd rather have something solid, and immovable... In there...

I've abandoned using the A7N8X For my new project after losing the 2nd one for no apparent reason... They just stop booting,

I'm ready to give the EPoX a try... I loved my 8k9A2+ a lot.... Very stable, and fast, maybe not as fast as an Abit Clock cycle for clock cycle, But I found I could clock higher, and with NO DRAMA!

Thanks,

Royd

[Jeff]

Royd, I'll try to measure the resistance needed for a .3v gain later today. That way the 2.9 setting will give 3.2v. BTW, this board maxes out at 3.2x volts. Randi discovered during his first mod attempt and I confirmed it when I was playing around the other day.

And as far as measuring Vdimm from Windows... I hope someone does the mod by lifting the leg on the chip and adding the VR there. I'm curious if doing it this way will make Vdimm monitoring possible.

Like I said a few pages back... my mod may be neglecting a trace that goes under the chip. :confused: I'm not sure.

[RoydRage]

Originally posted by Jeff
Royd, I'll try to measure the resistance needed for a .3v gain later today. That way the 2.9 setting will give 3.2v. BTW, this board maxes out at 3.2x volts. Randi discovered during his first mod attempt and I confirmed it when I was playing around the other day.

And as far as measuring Vdimm from Windows... I hope someone does the mod by lifting the leg on the chip and adding the VR there. I'm curious if doing it this way will make Vdimm monitoring possible.

Like I said a few pages back... my mod may be neglecting a trace that goes under the chip. :confused: I'm not sure.

Thanks Jeff,

Yeah... I was getting 3.28v. On the ASUS, which would have been nice.. But at this point, I'll take what I can get..

Royd.

[Crozet]

Hello,
i made this mode on my M/B but i have a problem now.
The BIOS is showing only 2.43V for the Vdimm!!
Is this normal or did something went wrong?
I suspect that the trace isnt cut and so the Ohms are added!
A response and help would be appreciated!
Thx
Crozet

edit:but when i separate the circuit my comp dont boots.so the cut must be correct!

[RoydRage]

Originally posted by Crozet
Hello,
i made this mode on my M/B but i have a problem now.
The BIOS is showing only 2.43V for the Vdimm!!
Is this normal or did something went wrong?
I suspect that the trace isnt cut and so the Ohms are added!
A response and help would be appreciated!
Thx
Crozet

edit:but when i separate the circuit my comp dont boots.so the cut must be correct!

Croz..

If you have a multimeter, Try measuring the voltage right on the board

RoydRage

[Crozet]

Hi,
i did just now. Its 2.43 and it cannot be changed from the BIOS!
I cant understand where the problem is....

Crozet

[RoydRage]

Wow...

Could you verify... That the trace was cut?

I would suggest using some junk RAM till you get this sorted... I have not done the mod yet, and am waiting for my board to arrive today.

Let me know how you make out

Royd

[Crozet]

OK i verified it now.The trace was CUT!
I reversed all changes and my comp is working like before.
Im pretty sure i did everything correct, so im start having doubts if this mod is working!

bye
Crozet

[Jeff]

Originally posted by Crozet
OK i verified it now.The trace was CUT!
I reversed all changes and my comp is working like before.
Im pretty sure i did everything correct, so im start having doubts if this mod is working!

bye
Crozet

You should say... working for you. I'm 100% sure it works for me since I'm measuring Vdimm right at the dimm socket and I get a better o/c from my RAM when I bump this voltage up to 3.2x volts.

Take a few pictures of what you are doing and maybe someone can help you out.

What resistance are you adding across the cut trace? You should be starting at ZERO Ohms and then working slowly up. Vdimm will top out ~300 Ohms at 3.2x volts and if you add too much resistance, strange things could be happening. I know my BIOS, MBM, and USDM all don't like when I have too much resistance.

[Crozet]

I added a 300 Ohm resistor at the beginning.The comp didnt bootet with this one.I lowered to 150 Ohms.The result was what i described above!
Vdimm was 2.43V measured at the Q32.BIOS was saying the same and it couldnt be changed!
Separating the circuit would result in no boot!
So i connected the trace again and everything is working like before.
I dont say the mod isnt working, but it sure didnt worked for me.

I have done many mods in the past including the vdd on this one.So i can say im handy with such stuff.I really dont know where the problem is.

bye
Crozet

maybe i will give it another shot tomorrow with lower resistors.see if this make the difference.

[Crozet]

OK ,i used a vr and it is working now! Dont really know the why since i didnt changed anything else but its cool!
Jeff , GREAT WORK!

bye
Crozet

[Jeff]

Glad to see it worked with the VR. :toast:

[Crozet]

Jeff, :toast:
What you would recommend for maximum voltage on the RAMs ?!
I have ~3.00 V right now.Should i give it more ?!Is it safe?
Im using 2 x OCZ PC3500 sticks.

thx
Crozet

[Jeff]

Don't know what OCZ likes. :confused: You have to be careful with some memory because it just doesn't like the higher voltage. With that said, my Corsair XMS loves 3.1v-3.2v.

[xtreme]

hmmm plz tell me which resistors and poti's to use... im stuck at 225 (corsair pc3500) / 223 (kingston hyperx) :(

[Bushboy]

Originally posted by xtreme
hmmm plz tell me which resistors and poti's to use... im stuck at 225 (corsair pc3500) / 223 (kingston hyperx) :(

EDITED, sorry released this was VDimm not VDDR (thanks muzz :D)

[muzz]

Wooooooooooo boy!!! We are talking vdimm here, not the vdd.

Here is what Jeff said, and he's done it.

"What resistance are you adding across the cut trace? You should be starting at ZERO Ohms and then working slowly up. Vdimm will top out ~300 Ohms at 3.2x volts and if you add too much resistance, strange things could be happening. I know my BIOS, MBM, and USDM all don't like when I have too much resistance."

[KS1]

Jeff,
Is it also possible to connect one end of a resistor to Q32 output pin (STR1) and the other end to GND? It's a simple voltage divider circuit. It will drop the volt across the circuit without doing any cutting on the board.

[Jeff]

Don't know. You'll have to wait for someone more intelligent to come around for the answer to that question. ;) My gut says no, but I really dunno... :confused:

[KS1]

Hmm.. I thought about it again.
Unlikely since that is a true drop in Vddr at STR1. This will force volt regulator to increase Volt. End up running volt regulator at 3.2V but get real 2.8V output.
I know now. The trick is only to fool a volt regulator but not to do the actual volt drop.

[KS1]

I am trying to understand this mod.
If anyone have any idea, please help.

What Jeff has done is to use a resistor to connect the output pin of MOSFET (STR1) to Vsense pin of the volt regualtor (after cutting the trace).

To my understanding, a resistor works by limiting CURRENT flow. Putting a resistor between those pins limiting the current to Vsense. So I don't understand how this mod work.

To my understanding, a volt divider circuit between STR1 output and Vsense pin will actually drop the volt (together with cutting the trace that Jeff did).

STR1 (MOSFET output)
|
VR1
|----------Vsense
R2
|
GND

Then volt will be adjusted by VR1. Starting from 0 Ohm. Raise the Ohm until the desired volt is achieved.

Any idea?

[Jeff]

All the resistor is doing is dropping some voltage. Pin21 monitoring Vdimm and if/when it gets a little low, the chip will compensate. What we are doing is faking it out. We are making it think that it's still trying to hold the voltage to 2.9v by stealing .2v or .3v. Pin21 is still getting 2.9v but everything else is getting 3.1 or 3.2v.

Everything except the Vdimm sensor. :D

[eliteone]

call me a newb guys but do i use a 1/2W resistor or 1/4W?

[Crozet]

Originally posted by eliteone
call me a newb guys but do i use a 1/2W resistor or 1/4W?

A 1/4 W.
But like Jeff already pointed out,it is much better to use a Variable Resistor since the "space" you can play with is very small.
From my experience with this mod i can say that 20-30 Ohms can make a huge difference on the resulting voltage!

bye
Crozet

[eliteone]

does this apply to the VDD mod as well? Using a 1/4W i mean.

[Crozet]

Originally posted by eliteone
does this apply to the VDD mod as well? Using a 1/4W i mean.

yes always

bye
Crozet

[muzz]

NP BB.........;)

[Learn]

hi,


the v-ddr MOD work fine :D THX Jeff

timing: aggressive 7,2,2,2 3.1 v



http://membres.lycos.fr/ftpad/images/3702.png

[Jeff]

Boy oh boy is that some bandwidth!!! :eek:

:toast:

[KS1]

I came across this Vddr mod guide.
They did Vbt as well to improve stability.

In the guide, they said that Epox is likely to include Vdd adjustment into BIOS as well.

Vddr mod guide (http://www.vr-zone.com/guides/EPoX/8RDA+/)

[cpulloverclock]

http://membres.lycos.fr/cpulloverclock79/vmem8rda+.JPG

[Jeff]

You can get Vddr to 3.48V??? Lets see a screeny with that!!! :eek: ;)

[cpulloverclock]

Originally posted by Jeff
You can get Vddr to 3.48V??? Lets see a screeny with that!!! :eek: ;) sorry
the voltage is fixed at 3.2v :mad:

[Jeff]

Ok, that's what I thought.

Any idea what the Vbt mod actually does? And do you really think it stabalizes anything? :confused:

[Jeff]

Also, do you think the Vdimm measurement is measuring that Vbt and times that by 2? I've only done the Vddr mod and I can't get accurate software measurements.

[Jeff]

'k... starting to think out loud again... :D

And I'm ignorant as to why you would ground the other lead of the VR? :confused: Anybody know why you'd do that? Does it actually change the measurement of the two pins you used? Or is it just a good practice to do so?

[Jeff]

Text is wrong for the Vbt mod. That second blue wire(which goes to the top transistor) is NOT connected to ground. That's where Vbt is... definitely NOT a ground.

The only ground wires are the TWO brown wires. Each of the blue wires is acting as a sense line... one for the Vddr and one for the Vbt.

[KS1]

In that guide, he could have USDM measured Vddr volt.

[Crozet]

Originally posted by Jeff
Text is wrong for the Vbt mod. That second blue wire(which goes to the top transistor) is NOT connected to ground. That's where Vbt is... definitely NOT a ground.

The only ground wires are the TWO brown wires. Each of the blue wires is acting as a sense line... one for the Vddr and one for the Vbt.

Cut another trace again and connect BTsen with Q33 is the VBT mod right?!
But is this neccesary?! Or is it just for reading the correct values in the BIOS :confused:

bye
Crozet

where to find the datasheet of this chip?

[Bushboy]

Excuse my ignorance, but what is the VBT mod for? It says that it is necessary to complete the VDDR mod, so is it just so that BIOS reads the correct voltage, seeing as a couple of people have successfully dont the VDDR without doing the VBT? Seems like cuttting teaces and such is not really worth it to get BIOS to read the correct voltage...A multimeter will do fine!

[Jeff]

Yeah, but Bushboy... if what you are saying(and what I'm asking) is correct, then you need to manually adjust this Vbt to about half of Vddr... so it's really not an accurate measurement of Vddr(using MBM or USDM or whatever). You know what I'm saying?

I know this mod has got to be doing something else(I hope), but it almost seems like it's just for the software monitoring programs... which is strange. :confused:

Oh well... I'm happy with doing just the Vddr mod. I'm going to do the Vbt mod this weekend and see if it helps 'stability' but I'm not holding my breath. ;)

[muzz]

Yeah been wondering about this myself since I saw it..........
What is the difference? Maybe DD can tell us.

[muzz]

BTW CPU.... how are you showing it in the USDM screenie? Did you do the vbt mod also?
Sorry if I missed you saying that you did.
EDIT: Last but not least... did you measure c32 and come up with the same voltage as what is pictured.

Thanks

muzz

[Bushboy]

Count me in on one of those who successfully did the mod ;) I'm not sure the exact voltage yet, but I'm quite sure it worked perfectly! Gotta give a big thank you to Jeff, with his work on both the Vddr and Vdd mod, it's great to have someone like him around here. I'll be seeing how much of a difference .3v makes with regard to FSB tomorrow...it's getting late.

[KS1]

Lifting a pin:

I have one experience to share with you, perhaps it might be of useful.

While I was doing a Vdd mod, I accidentally soldered a wrong pin (worked late at night). I tried to de-soldered that one out using a solder gun for a very short period of time intermittenly and tried to detach the wire out using a bit of force. Suddenly, it was not the wire that was detached. They were the wire together with a pin. After correcting my mistake, I soldered that pin back onto the PCB and it still worked OK.

If you don't want to cut the trace on the m/b, you might try this method. Use the smallest wire available. Primer the wire with solder then use soldering paste. Next solder it to the pin. Then pretend to undo the soldering work.

[Crozet]

The datasheet from this winbond chip say the following about the functionality of the BTsen:

"Bus termination controller STR2-1.25dual for high speed bus termination to sink and drive redundant current"

Maybe someone can translate this so me can understand this to ?!! :) :rolleyes:

bye
Crozet

[sharpbw]

How do I measure the gain at Q32? Sorry, but don't quite understand the connection points to test.

[muzz]

Pin closest to the middle of the mobo is where you put positive( on Q32 ), case or mobo ground is neg.( Using multimeter)

Look where Jeff has the wire attached to Q32( in his pic).. thats the point for the pos. on MM.

Unless someone corrects me, thats where I saw the increase.

[KS1]

what about the other 2 mosfets very near (nearer) to DIMM sockets, have you tried measured thems?

[muzz]

I think I did, but not 100% sure...... I did see the increase on the pin stated......... didn't seem to do alot for me here, then again I only went up to 3.1 for my testing.

[KS1]

if you look at the trace of the mosfet Q34 and Q33, the trace looks like they are serving Vddr to DIMM as well.

if you got +3.1V (when set at 3.1V on Q32) on other pins Q34 and Q33, it appears that there are more than 1 mosfets responsible for converting the volt output.

if you got <>+3.1V, it seems that doing just Q32 mod is a job half done?

[Jeff]

Q33 is for Vbt and Q34 is for ???.

[WildKard]

Ok what wires do i connect to the Trace and to the Q32 using a VR do i use both hotpins(the outside pins) and leave the ground out of the equation?(thats what it looks like to me just double checking)

im going to attempt this mod tomorrow afternoon im hoping to get 235-240 5-2-2-2 stable...thatll be ultra nice :D

[bowman1964]

well modded mine last night.before max was 215 dual memory,stable prime at that.did mod now looks like i am able to run 225 dual with prime stable @ 3.05.:)
but i noticed while i was stressing ram last night i am seeing a drop to about 3.042 under load.
so maybe the second part of the mod is needed VBT mod ?
this might stablize the voltages.

also i am not sure if you guys noticed but.while i have my meter connected to the mofset measuring the dimm voltage.if i shut down in windows i can see the meter drop to no voltage like it should but.........if i disconnect the power suppy for a second.then reconnect it i can watch my memory voltage rise to 2.77 with the motherboard still off.
:rolleyes:

i dont understand that at all?when i power up the motherboard the voltages rise like they should back to 3.05.
any ideas or is that normal you think?

[Bushboy]

Originally posted by WildKard
Ok what wires do i connect to the Trace and to the Q32 using a VR do i use both hotpins(the outside pins) and leave the ground out of the equation?(thats what it looks like to me just double checking)

im going to attempt this mod tomorrow afternoon im hoping to get 235-240 5-2-2-2 stable...thatll be ultra nice :D

Connect a wire from one hot pin to the trace, and then from the ground on the VR to Q32...Using two hot pins will give you a fixed resistance always, therefore eliminating the Variable of Variable resistor ;) Make sure the hot pin you use and the ground have a resistance of 0ohms first, then turn the knob to increase.

[cpulloverclock]

I have 1.85v for VDD and 3.34v for Vmem

It's rock stable at 225-226 cas2 7-2-2 in DUAL DDR with 2*Corsair Pc3200 cas2

[cpulloverclock]

edit : 230 cas2 7-2-2 with 2 adata winbond 5ns now
vdd 1.95v

[KS1]

Vbt is half of Vddr. I think that it is used to terminate the bus of DDR RAM so we could use 1, 2, or 3 DIMM slots(?).
Can someone please check Vbt pin to see if it's actually half of Vddr after the mod (ie. Vddr 3.2 --> Vbt 1.6V)?

[bowman1964]

Originally posted by KS1
Vbt is half of Vddr. I think that it is used to terminate the bus of DDR RAM so we could use 1, 2, or 3 DIMM slots(?).
Can someone please check Vbt pin to see if it's actually half of Vddr after the mod (ie. Vddr 3.2 --> Vbt 1.6V)?

no you have to set it manually to half.with my mod if i set the memory to 3.1 the vbt only shows 1.45 which is only half of 2.9v which is the max the board thinks you can go to.
so you have to start off with 0 ohms and slowly turn the vr resistor untill you reach half of your memory setting.for my 3.1v setting i set the vbt to 1.55v:D

[KS1]

That means Vbt use data from Vsense pin to adj bus terminal voltage.
If we only do Vsense mod, Vbt is not 1/2 of Vddr.

[Jasonxxx]

Well, I want to do the vdimm mod, but do I have to do the vbt mod also to make it stable??

Jeff, muzz anyone: Anyone done this vbt also and shown more stablity or does this mod just show the correct vdimm in software?????

[bowman1964]

Originally posted by Jasonxxx
Well, I want to do the vdimm mod, but do I have to do the vbt mod also to make it stable??

Jeff, muzz anyone: Anyone done this vbt also and shown more stablity or does this mod just show the correct vdimm in software?????

well it doesnt do anything for showing correct vdimm in softwear.
it just stablizes the memory more like it should be.i can tell a differance under load.before mine used to drop from 3.05v while running 3d2001.( i keep a volt meter attached the the mofsets while benching) would drop from 3.05 to 2.96 or so.now it holds within a few points.3.05 -3.03.
but you wont tell this unless you have the system pushed hard.
:D

[Jasonxxx]

hey thanks bowman!!!!:toast:

i will complete the mod over the next few days in between work........:D

I want to be able to complete 3dmark at 230mhz fsb.......:banana: x that by 11.5.........

and get this 9700 up around 490/800 "a far cry" I will get a real good 3dmark score........

[youthemandan]

Will a 500ohm VR work???? If not what is the best one to use???

[Holst]

VBT didnt make my bios show the correct voltage.

And I didnt get any higher overclock with it.

I havent tested the votlages like bowman, but I wasnt getting much drop under load before VBT so I dont think there would ahve been a big change.

I used a 500ohm VR BTW.

[KS1]

I think VBT is used to terminate the VDIMM bus so a PC could be automatically configured to work with only 1, 2, or 3 slots of RAM.

To test this, use 1 slot and try setting VDIMM in BIOS to +2.50V and adjust VR resistor to get +3.4V. I think that a PC won't boot due to 36% discrepancy between +1.25V VBTout and the expected +1.70V (1/2 of VDIMM)

Vbt needs to be 1/2 of VDIMM.

[youthemandan]

So do i need to do the VBT mod?

[KS1]

i recommend.

[Nohto]

Why are they using both sides of the VR on the mods over at VR Zone (http://www.vr-zone.com/guides/EPoX/8RDA+/)?
Edit
It doesn't look like they are going to both sides of the VR, but it doesn't say exactly where the other side of the brown wire was attached. Can anyone verify the exact location of the 2 brown wires coming of the Q34 Mosfet?
Thanks in advance
Nate

[Jeff]

Looks like they grounded the other side out... which I asked about a month or so ago and nobody had any reason as to actually why someone would do that. ;)

[Nohto]

The Mosfet leg is ground. I doubt that they went from a ground to a ground.

[Barracuda]

very nice m8 thank you :)

[youthemandan]

Ok so lemme get this straight which of those wires is a ground?? the one soldered onto the trace right?

[Nohto]

MickeyMouse, I was talking about the 2 brown wires that they used in the picture. Can you or anyone tell me exactly where they go and what they are for would be helpful also.

Thanks
Nate

[youthemandan]

They are not connected to the pot from what i can see

[Nohto]

Originally posted by Nohto
Why are they using both sides of the VR on the mods over at VR Zone (http://www.vr-zone.com/guides/EPoX/8RDA+/)?
Edit
It doesn't look like they are going to both sides of the VR, but it doesn't say exactly where the other side of the brown wire was attached. Can anyone verify the exact location of the 2 brown wires coming of the Q34 Mosfet?
Thanks in advance
Nate
I know I stated that earlier.

[youthemandan]

so do i have to expose the trace to solder that wire on???

[Nohto]

Yes I do believe so. If you go here (http://www005.upp.so-net.ne.jp/octo/pc/) they have better pictures of the mods.

[youthemandan]

Ok so lemme get this straight which wire is ground. On a pot like the one's at the shac how would i go from the dial side down. Would it still go Blue white.

[youthemandan]

Correct me if i am wrong but isn't this how those corresponding wires in that pic should go if i were to use a pot like this
http://mollyockett.com/potentiometer.gif

[Jeff]

I did not connect the brown wires when I did the mod and it works fine for me. I really don't know what they are accomplishing and I haven't heard anyone give a reason as to why you would do it so I don't. :D

And you can use either of those pluses along with the negative if you do it like I did.

[bowman1964]

ok guys from what i am told ..and i havent tried to confirm this because mine work...
but the brown wires are to ground and it is suposed to make for a non varying resitance.the vr resistors sometimes change slightly and that is surpossed to stop that.

but i have no reason to try to find out LOL:D

[morphling1]

http://www2.arnes.si/~mlivak/8rda/vddr%20mod.JPG

I've made both mods vddr in vbt, I uses 5k pots and did the same as vr-zone. I don't know what are the values of both trimmers because when I set them to 500ohm and started computer I only got FF on diag disp. So I take multimeter and measured the voltages. They were both way off, so I first started turning vbt pot to 1.45V (I had 2.9V set in bios) when I set that, I started turning vddr pot. and check for vddr voltages with multimeter. When I find 2.9V I restarted and went into bios (everything was working obviosly) and change vdimm to 2.77V. Then back to post up and press pause, to stay there. Then I again started turning the pots and set vbt to 1.55V and vddr to 3.1V. I wanted 3.1 at 2.77V to have room left. Then I went to bios again and set vdimm to 2.9V, restart and check the voltages, vbt was 1.62 and vddr was 3.23V just as I want them. I didn't try anything further but I definetly could. I'll be playing with 3.2+ to see if Kingstons show any imporvement past 3.2V. My BW is in overclocking section btw.-

[spaceboy]

Just done the mod :i did only the Vdimm mod now.. not VBT
I've set a trimmer to 220omh..
mobo booted fine while setting 2.9 in bios..
i've tried to mesure with multimiter but i always read 3.25 even setting 2-5 in bios

if i set 2.9 or 2.77 mobo doesn't load windows

i'll try to work it out

[spaceboy]

oh yes :D
i work it out 2 minutes later..
i forgot to post it !

i read 220 ohm in another mod guide..
so i set it before soldering
then i lowered the pot while measuring with multimeter and set it to 3.2v max

so VBT isn't worthy right?

[Zoom Proof]

Hi guys :)

Can someone tell me , where i can get proper vddr and vbt
with multimetr ? I'll be very glad to see some pics where
these points are selected.

ps sorry , but my english as not good as i want and
it's so difficult to me to read all these posts :(

[Shroomalistic]

I did the vdd mod, and im wondering if i should do the vmem. After the vdd mod was done i am able to run at 230(460fsb) but cant get any higher with any more voltage which leads me to belive the ram needs more now. Can anyone confirm this for me?

[k1114]

Originally posted by Shroomalistic
I did the vdd mod, and im wondering if i should do the vmem. After the vdd mod was done i am able to run at 230(460fsb) but cant get any higher with any more voltage which leads me to belive the ram needs more now. Can anyone confirm this for me?

Try to run you ram higher then your fsb to see if it really is holding you back. If it won't run any higher, then you might need the vdimm mod.

[gtcrispy]

I want to try this but I think I lack the skills. My 8rda is undervolting down to about 2.83-2.84V when I have it set at 2.9V. It this normal? The vdd is set at 1.9V. I can get it too about 227-228mhz though somehow a month ago i got it to 232mhz and benchmarked a few things.

[gtcrispy]

its from mbm5.3 and the bios.

[lght1]

Hi

*bump*

Just a refresher question concerning these 8RDA's rev1.1,

Does it make sense to vdimm mod this board, ot just jump ship to a NF7 2.0 ?

[OC&Burn]

Originally posted by bowman1964
well it doesnt do anything for showing correct vdimm in softwear.
it just stablizes the memory more like it should be.i can tell a differance under load.before mine used to drop from 3.05v while running 3d2001.( i keep a volt meter attached the the mofsets while benching) would drop from 3.05 to 2.96 or so.now it holds within a few points.3.05 -3.03.
but you wont tell this unless you have the system pushed hard.
:D

Ok, I haven't read the whole thread but tell me if im wrong. I you do a simple Vdimm mod without cutting or soldering. Just using grabbers and a pot, you _will_ get higher Vdimm, but it will not be as stable as it would be if you raise the VBT too? You will get a voltage drop? Anyway, if that's all, a voltagedrop to 2.96 is good for me, it's alot better then 2.83 that im getting now with my 8RDA3+ and 2.9Vdimm in BIOS.

[lght1]

Hi

I've already finished both the vdmm and vbt mod. With a few turns of two screws, I have all the way up to 3.2v :eek:

However, I am not able to test it out as I sold my crappy GEIL PC3200 DC Ultra. This ram wouldnt o'c hardly at all regardless of how relaxed the timings were, or how high the voltage.

I'll likely send for some TwinMOss next week.

Still, although I've had this 8RDA r1.1 since Feb, I still like it a lot.

I've given a NF7 2.0 some consideration, but I may keep this beauty.

[OC&Burn]

Originally posted by lght1
Hi

I've already finished both the vdmm and vbt mod. With a few turns of two screws, I have all the way up to 3.2v :eek:

However, I am not able to test it out as I sold my crappy GEIL PC3200 DC Ultra. This ram wouldnt o'c hardly at all regardless of how relaxed the timings were, or how high the voltage.

I'll likely send for some TwinMOss next week.

Still, although I've had this 8RDA r1.1 since Feb, I still like it a lot.

I've given a NF7 2.0 some consideration, but I may keep this beauty.

i see your using the Vdimm mod. Is my question above your reply correct? Can I use just the Vdimm mod to raise the Vdimm, BUT it will drop some under load? I don't want to destroy the board with the vbt mod (seems a bit hard for me....) but I need higher Vdimm.

[lght1]

Hi

I would recommend both mods, as the loss of vtg stability could possible cause overclocking to be affected as well as some intense 3d apps.

In my case, the 1.1 I spoke of earlier, died last week along with a 9500 Pro . It was a power supply that caused such devastation.

I am now on a v2.1 8RDA+ which will do 230 without any mods other than adding active cooling to the nb.


Rather than taking a chance on wrecking a perfectly good older NF2, I'd go for the Ultra 400 versions out now. I wished I 'd have done that, then I'd have another NF2 to upgrade the other pc in the house with.:doh:

BTW, My 2.1 actually provides a real 2.9v to dimms. However, it tends to undervolt the cpu, so I just set it higher for the same vcore as before. The vdd needed to do 230 is 1.8, as opposed to the 1.92 for 220 on the v. 1.1 .

[OC&Burn]

Originally posted by lght1
Hi

I would recommend both mods, as the loss of vtg stability could possible cause overclocking to be affected as well as some intense 3d apps.

In my case, the 1.1 I spoke of earlier, died last week along with a 9500 Pro . It was a power supply that caused such devastation.

I am now on a v2.1 8RDA+ which will do 230 without any mods other than adding active cooling to the nb.


Rather than taking a chance on wrecking a perfectly good older NF2, I'd go for the Ultra 400 versions out now. I wished I 'd have done that, then I'd have another NF2 to upgrade the other pc in the house with.:doh:

BTW, My 2.1 actually provides a real 2.9v to dimms. However, it tends to undervolt the cpu, so I just set it higher for the same vcore as before. The vdd needed to do 230 is 1.8, as opposed to the 1.92 for 220 on the v. 1.1 .

okie, but I think im just going to do the Vdimm mod, don't wanna ruin the board because of my soldering skills ;)

Do you have a picture of a Vdimm mod?

[lght1]

HI


I would recommend both, as one isnt any more or less difficult then the other.

Here is a good link with pics and a step by step guide.

http://www.motherboardfaqs.com/article.php?35.255

[q149]

Here is a crappy pic of one of my Vdimm mods on an 8rda+ 2.0..

http://q149.netfirms.com/cgi-bin/8rd4.cgi

[STEvil]

You're right, that is crappy! 8-)

EDIT

Has anyone tried running 3 sticks of ram before and after the mod?

EDIT 2

What would happen if we only did the VBT mod?

[lght1]

Hi

All the vbt does is stabilize the current voltage setting. Without the vdmm mod to go with it, there shouldnt be any need.

[STEvil]

Yes, but what would happen if it were adjusted slighlty high or low? Could it possibly play a trick on the vdimm?

[lght1]

Hi

Perhaps I misunderstood.

I thought that you were only going to do the vbt. If that is the case, it wouldnt mean anything as the mobo should not needed it if stock bios settings are used.

VBT settings are exactly half of what the dimm voltage is set for. The mobo does this itself with factory settings. Its only when we exceed the factory specs that this vbt mod is necessary.

[STEvil]

What I am wondering now is if there is a way to disconnect the vdimm and just loop it back to the IC, and run an extra wire right from the 3.3v line on the PSU through a couple caps or something and into where the old vdimm line was going... or through the existing circuitry even.

[STEvil]

Anyone know of a 1.6v source on the board?

If you connected a wire from a 1.6v source on the board to the BTsen pin you would be able to skip the Vbt mod.. may have to lift the pin, though... or cut the trace to it?...

http://www.vr-zone.com/guides/EPoX/8RDA+/VDDR-1.png

Wish I could figure out what pins 1 5 and 24 do as this IC is supposed to be able to overvolt up to .4v over, but gives no documentation in its PDF's on how exactly it does that in stock configuration....

http://www.winbond.com/e-winbondhtm/partner/PDFresult.asp?Pname=866

[STEvil]

ok, after extensive searching I have found some resistors that give me 83.2 ohms when soldered in series...

Going back to Mickey's post that he used a 72-76ohm resistance for 3.1v @ 2.9v, would this resistance give me close to 3.2v..?

http://www.xtremesystems.org/forums/showthread.php?postid=117087#post117087


UPDATE

Did the mod, with an 80ohm resistance I am getting 3.0v @ stock settings. Tad high, but oh well.

[DragonFlight]

I am ready for the Vdimm mod on my 8RDA3+. I will do the Vdimm first and if i have any problems i will do the Vbt also!
Now i have a question: Did anyone tried this mod succesfully on 8RDA3+(NOT 8RDA/+).

[Rub87]

Originally posted by Zoom Proof
Hi guys :)

Can someone tell me , where i can get proper vddr and vbt
with multimetr ? I'll be very glad to see some pics where
these points are selected.

ps sorry , but my english as not good as i want and
it's so difficult to me to read all these posts :(

Here (http://www.motherboardfaqs.com/article.php?33.255)

pic1 of my mod (http://users.pandora.be/rub87/SIMG0614.jpg)
pic2 of my mod (http://users.pandora.be/rub87/SIMG0613.jpg)

[Swiftech]

Ok, I'm pretty confused here. You cut the PCB hard enough until there are no circuit running ?

I've seen another method which is lifting up the leg of the IC. However, do we always need to perform the Vbt mod ?

[STEvil]

i havent done the vbt yet... my voltage is sitting pretty stable it seems (watched it for a few runs of 3dmark, didnt dip farter thane 3.17).