Here is extremely important info for DFI NF4 Vdimm

I just talked to Oskar on the MSN, and we discussed a bit regarding the Vdimm source.
All the story is here, I already translet the conversation to English version.
This is extremely important that you will need to know if you expect to get DFI NF4 series board.

hope this will help!! :toast:

OPB
************************************************** *******
opb: Is that true that LP nf4 vdimm use 5v rail?
Oskar: no, it's not directly from 5v rail. It got jumper(JP17) can switch vdimm source!
OPB:oh?
Oskar: If you see the board, on the lower right hand side corner, there is a 6x2 jumper set, that's the
vdimm source switch!
(OPB: please see the attached pic)

oskar: so if user directly switch the vdimm source from 5v rail but use actual vdimm value under 3.3v,
then that would cause serious heat on Mosfets. That 's why the factory default vdimm source is set from 3.3v rail.

**(OPB: SO, if you are not going to do vdimm over 3.3v for 24/7, I and Oskar would not suggest you to switch vdimm source to 5v rail to get that heat on your

MOS)

oskar: However, even though user switch 5v rail source to do vdimm above 3.3v, the heat is still unbelivable hot there. therefore, that's why we use
the 4000~10000 hours 105c very good caps on that area. I would definitely suggest users can give that area some air flow to reduce some heat from there.


Board pic, the red circle shows the Vdimm source switch for 5v or 3.3v rail.

http://www.devilsword.com/lpnf4/lpnf4vdimm.jpg

awesome stuff.. all i need is the board lol ..ask him when did DFI send their NF4 shipment to Australia while you have him on MSN :)

Cute information OPB :toast: , thx for both of you

Quote:

---

Originally Posted by eva2000

awesome stuff.. all i need is the board lol ..ask him when did DFI send their NF4 shipment to Australia while you have him on MSN :)

---

He is going out with is girl friend now..:D:D:D..J/K
I will ask around! Be patient. :stick:

Nice! thanks for the info :)

Quote:

---

Originally Posted by Onepagebook

He is going out with is girl friend now..:D:D:D..J/K
I will ask around! Be patient. :stick:

---

silly girls, theyre always getting in the way of OCing :p:

Thanks OPB, I'll update my "Ultimate DFI thread"...

Thank D_o_S..you did a great job!:D

interesting info...very nice to know. So basicly your saying put sinks on the Mosfets and have a fan run on 'em. Sounds like what some do to there CPU Mosfets :D

jjcom

Quote:

---

Originally Posted by jjcom

interesting info...very nice to know. So basicly your saying put sinks on the Mosfets and have a fan run on 'em. Sounds like what some do to there CPU Mosfets :D

jjcom

---

Yes, and especially at the lower corner of right hand side ;)


OPB

neat trick

Great idea..

DFI probly should have made some kind of Moset cooling ala Abit.. but with the mosfets on the right side of the board its kind of hard.

off topic, but has anyone tested 4 dimms all being used on this mobo?

Thanks

The put sinks on all other mosfets on the board why not those as well? For a $300 board I am sure they could have put a couple more sinks on there.

Either way I would like to see someone switch to 5v rail to see what results they get. Hopefully not a melted board.

yeah, but it ain't that hard to put sinks on the board...just time comsuming. O'well.

Its still a nice feature...but it seems like maybe the DDR booster would be better for those that want more than 3.3v on the memory 24/7...OCZ VX users for example. Or would the booster still heat those up?

jjcom

Quote:

---

Originally Posted by jjcom

yeah, but it ain't that hard to put sinks on the board...just time comsuming. O'well.

Its still a nice feature...but it seems like maybe the DDR booster would be better for those that want more than 3.3v on the memory 24/7...OCZ VX users for example. Or would the booster still heat those up?

jjcom

---

ok, I am sorry, I forgot the DDR booster thing. I did ask him if DDR booster works and Oskar did mention about that.
He uses the same idea as OCZ ddr booster does. They are all same concept of using 5v linear. Therefore, as long as any motherboard can apply booster, it got same heat amount as LP NF4 does. ;)

OPB

4.0V Vdimm is confirmed and working, just tried!!! :banana:

picture del.

MOD

and get 5V^_^

They should've used a switching voltage regulator, not a linear one. Runs MUCH cooler and it would've eliminated those jumpers, as it could use the 5V line all the time.

EVA2000 I can answer that one. The 26th January is the answer.
http://www.hardwarecentral.com.au/pr...roducts_id=388

Quote:

---

Originally Posted by TinTin

4.0V Vdimm is confirmed and working, just tried!!! :banana:

---

:banana4: * looks at his VX sitting on the shelf.

Great info, OPB!

seems like gotta mod a 60 or 80mm fan to the 5 1/4" bays to get some cool air through that area. Might be a little tough with all the power cables located beside it.

yeah, I think I might have a reason as to why the 5v rail isn't used all the time. I wonder if its PSU related, like certain models may not be able to supply the power needed.

jjcom

It's in OPB's post, using the 5V causes too much of a heatload on the mosfets in the RAM voltage regulator. That's why they prefere to use it only when it's really necessary, like when you need more than 3.2V Vdimm.

maybe its so people with 20 pin psu's can still use the board but with up to 3.3v only?

Dave

I'm hoping some brave soul will try the 5v rail menthod and see what they come up with. I can't see anyone using the full 4v anyhow and most people with VX are getting pretty good results with only 3.3-3.4v. Max I see used is about 3.6 though while high is still not a full 4v.

The option to use 3.6+ can be a good thing with BH-5, they still don't give as much heat as TCCD on 3.1 v when they are on 4 V.
And the perform better for benching.

yeah, having the increased Vdimm is good for benching. But 3.3v maybe enough for a 24/7 computer. But still a nice option to have.

jjcom

Such a simple solution... HOW COME I DIDNT THINK OF IT??? :brick:

Hopefully they will use this on their NF3 939 board as well

Quote:

---

Originally Posted by DaGoof

Such a simple solution... HOW COME I DIDNT THINK OF IT??? :brick:

Hopefully they will use this on their NF3 939 board as well

---

I doubt they will have 939 board. :p:

OPB

... you sure? I've heard a lot recently that they were finally going to put it out

http://www.xtremesystems.org/forums/...fi+nforce3+939

http://www.xtremesystems.org/forums/...fi+nforce3+939

Quote:

---

Originally Posted by DaGoof

... you sure? I've heard a lot recently that they were finally going to put it out

---

Yes, Vivian told the same thing @ AMD Fall Tech tour.

so this board would be definitely the best choice for OC players? And DIMM area still ought to be actively cooled?

It should be, from the looks and from what DFI has done in the past. The DIMM area should be activly cooled when using the Vdimm from the 5v rail but otherwise, no more than on anyother board it would seem.

jjcom

3.96V max ;)

That's cool, can you run some benches with the vdd set to 3.8v-4.0v and feel if the mofsets are getting warm?
Or are you afraid to blow your mem?, I can understand that offcourse.

Quote:

---

Originally Posted by Onepagebook

I doubt they will have 939 board. :p:

OPB

---

Aaargh :stick: Are you sure of that? i can't live with the neo2

Quote:

---

Originally Posted by alexio

That's cool, can you run some benches with the vdd set to 3.8v-4.0v and feel if the mofsets are getting warm?
Or are you afraid to blow your mem?, I can understand that offcourse.

---

3.8V tested with 512*2 Mushkin BH-5 is fine for benching, with a fan blowing on MOSFET... :)

Quote:

---

Originally Posted by TinTin

3.96V max ;)

---

Wonderful! :banana:

Quote:

---

Originally Posted by Onepagebook

I doubt they will have 939 board. :p:

OPB

---

So NF4 ultra isn't 939?

you know what he meant...



how about booster on the dfi nf4 on the 3.3V line?

how about booster on the dfi nf4 on the 5V line?



I know there's no point switching to 5V line AND using booster but just for safety cuz I thought that with all boards + booster that draw Vdimm from 5V eather the booster burns or the board burns out but it never works.

Anyone know wich voltage regulator they used with 5V line on memory ?

Quote:

---

Originally Posted by Cyrus The Virus

Anyone know wich voltage regulator they used with 5V line on memory ?

---

Anyone ??

http://www.silverstonetek.com/products-tj06.htm
This case looks like it could help with cooling CPU/MEM/MOSFET area!

:D O.K. guys - maybe I got somethin' wrong but DFI manual says if jumper is set to 3,3 V rail max Vdimm is 2,85 and OPB and Oscar told us to switch the jumper to the 5 V rail only when using more than 3,3 Vdimm.

But what to do if something between 2,85 and 3,3 is needed ???

Does anyone know ???

:toast:

Quote:

---

Originally Posted by HARDCORECLOCKER

:D O.K. guys - maybe I got somethin' wrong but DFI manual says if jumper is set to 3,3 V rail max Vdimm is 2,85 and OPB and Oscar told us to switch the jumper to the 5 V rail only when using more than 3,3 Vdimm.

But what to do if something between 2,85 and 3,3 is needed ???

Does anyone know ???

:toast:

---

That manual is more than likely incorrect. Using the 3.3V rail, it should at least get you 3.1V-3.2V on your VDimm.

yeah, since it came from Oskar chances are the manual is wrong...maybe an early beta version? Yeah, if its like the other boards...about .2-.15v below the 3.3v rail is the Vdimm. So should be nice :D

jjcom

:D Good news here - my G.Skill LA like to get 2,9 - 2,95 at higher speeds so it will be working fine just from the 3,3 V rail.

Thanks all !!! :toast:

Quote:

---

It's in OPB's post, using the 5V causes too much of a heatload on the mosfets in the RAM voltage regulator. That's why they prefere to use it only when it's really necessary, like when you need more than 3.2

---

Then, why did the DFI Infinity and LanParty nForce2 motherboards draw VDIMM off of the +5VDC rail for 24/7 use? My old Infinity NF2 didn't even heat up very much around the VDIMM Mosfets.

Add some nice copper heatsinks, and you're all set so long as you have good airflow. I don't think that a fan specifically for the Mosfets will be too easy to implement well (look at all the power cables around there), and hopefully not even neccesary for stable operation.

Quote:

---

Originally Posted by felinusz

Then, why did the DFI Infinity and LanParty nForce2 motherboards draw VDIMM off of the +5VDC rail for 24/7 use? My old Infinity NF2 didn't even heat up very much around the VDIMM Mosfets.

Add some nice copper heatsinks, and you're all set so long as you have good airflow. I don't think that a fan specifically for the Mosfets will be too easy to implement well (look at all the power cables around there), and hopefully not even neccesary for stable operation.

---

As far as I knew they didn't...

jjcom

As far as I remember they did infact. You didn't need anything out of the ordinary on your 3.3v rail to run your ram at 3.3v with those boards.

Whoa...thats very new to me...kinda makes me wish I had gotten an Infinity board now...lol.

jjcom

Quote:

---

As far as I knew they didn't...

jjcom

---

A straight-up VDIMM voltmod and I was up to 3.6V of VDIMM, and with my +3.3VDC rail at only 3.55V ;). One of the reasons these boards were so excellent for FSB overclocking.

Both DFI nForce2 boards fed VDIMM off of the +5VC rail :).


And neither of those boards needed active Mosfet aircooling...

939 LP's probably dont use a switching power supply for vdimm

Switching runs cooler, but linear runs cleaner. The RAM will like the linear VR better. Would be interesting to see what happens with the 3.3/5 jumpers totally removed using the booster.

One thing I noticed when playing around with this board was that you can set the jumper for 3.3V and then increase your 3.3V rail on your PSU. Same effect as on the DFI NF3-250. I got the feeling that this was a little more stable then when running from the 5V rail with the same vdimm. This little trick made the mosfets even hotter though.