DFI nf3 - Vdimm OVP (can it be removed?)

Quote:

---

Originally Posted by ][nSaNe

yes i see that mod and i think i will have to do that this weekend. im going to move my setup around and face the OCZ Powerstream pots towards me. and go to town.. i wish someone could really figure out this OVP thing though...
do you loose some stability or some cleanliness of the power if you bypass and feed the ram directly form the PSU and not through the mofsets and caps and such?

---

You'll gain in stability and gain in mem OC, the OVP will still remain though ... :mad:
That is if the +3.3V rail are strong and stable [ OCZ Powerstream have strong and stable rails ].

For me the OVP kicks in at Vdimm higher than ~3.65-3.67V [ 3.70-3.72V on +3.3V rail ] but today when I was shooting for 10x280 at 1.5 2 2 6 2T [ 1T is a no go at this speed ] I adjusted my +3.3V pot to 3.75V and successfully booted up.
It was the first time I could use a Vdimm of 3.70V !!!

But Memtest v1.55 refused to even load ! :nono:
A poor history with a crappy ending I know. :slapass:

Congrats to twilius_basic with a 38K s754 '01 score! :clap:

I run my 3,3v line at 3,9 all the time without problems on this board, however bios only reads 3,75...
Using a booster I can boot with a vdimm of 3,7V, if I want higher I just raise vdimm when in windows, works fine all the way up to 4,2v.
I`ve had a few dfi nf3 boards and they all behaved the same way.

Oh, almost forgot, before the Persivor booster-modd, I could only boot at 3,5v, raising vdimm from bios was no problem though.

yea its not worth borking bios' like 10 times trying to mod a bios to allow over 3.1vdimm
so i broke down and soddered more wires to my stuff... 3.3+sense to pin 7..
got 275HTT or ddr550 5-2-2-2 to memtest for over an hour... not a single error..
it takes the 3.3rail cranked to 3.65 after it boots of course.. previously i couldnt hit that high of a fsb with this ram... my vdimm dropped alot when i was running it that fast. it would be @ 3.6vdimm.. and loaded it would drop to 3.3 or 3.4.. not anymore..
this was a very easy mod.. it took 10 minutes.. i didnt even pull the ram out of the slot. just flipped it up.. connected the dots.. and boom... i got a nice ram clocking board now.. and also.. i dont have to wait for people to mod bios files.. just flash anyone and get alot of vdimm... its pretty sweet... xtreme is totally where its at.

I have to boot @ 3.5v or lower and dial-up on booster. Then OVP hits my board @ 3.7v

Here's another chip not too far away from Dimm's for the guru's to look at.

http://img.photobucket.com/albums/v3...UTNF3dimm2.jpg

Quote:

---

Originally Posted by G H Z

I have to boot @ 3.5v or lower and dial-up on booster. Then OVP hits my board @ 3.7v

Here's another chip not too far away from Dimm's for the guru's to look at.

http://img.photobucket.com/albums/v3...UTNF3dimm2.jpg

---

I checked out the chip, and it looks like its just the I/O controller for overclocking, so it probably isn't directly connected to the Vdimm OVP issue :( . Can someone try to get a clear, hi-res picture of the the Dimm slots and surrounding area please.

Quote:

---

Originally Posted by persivore

I checked out the chip, and it looks like its just the I/O controller for overclocking, so it probably isn't directly connected to the Vdimm OVP issue :( . Can someone try to get a clear, hi-res picture of the the Dimm slots and surrounding area please.

---

OK, here you go. Non-broadband users beware ;)

http://www.fh-augsburg.de/~madocer/D...imm%20area.jpg

If I'm not mistakin' the chip responsible for Vdimm is the Philips LM358.

Quote:

---

Originally Posted by celemine1Gig

OK, here you go. Non-broadband users beware ;)

http://www.fh-augsburg.de/~madocer/D...imm%20area.jpg

If I'm not mistakin' the chip responsible for Vdimm is the Philips LM358.

---

The saga continues :D

Hi guys,

I didn't have much time lately, but today, I had a free afternoon and so I decided to go ahead and finally start with initial modifications of my DFI LP NF3.

I removed both of the Mosfets that are responsible for Vdimm. I'm going to replace them with much higher rated units, and I'm planning to use the 5V rail to feed Vdimm. I'll build my own little circuit.

Here's a pic of the initial mods, only mosfets removed (the cap that is normally in between the mosfets was also removed to have more space for soldering; I'll solder it back in, as soon as I have the new mosfets on the board):

http://www.fh-augsburg.de/~madocer/D...nnections2.jpg

Again, non-broadband users, beware.

OK, here's a little update:

Yesterday I felt like it was time to make the 5V Vdimm mod. I built my own little PCB fed by the 5v rail, with two high amp (30V 40A) diodes to drop the voltage slightly and three caps (low ESR 2200µF 6.3V) to filter the voltage. Then I soldered in the new mosfets and didn't connect the former 3.3V-input pin (source pin of the mosfet on the left, directly under the dimm slots). I had that pin bent up, so I could connect it to my little circuit that would feed it with ~4.8V.
Well, the good thing is, nothing burnt up when I powered the board up, BUT:
Something did go wrong, as the Vdimm mosfet leg (source leg of the mosfet on the right, directly under the dimm slots) that normally showed ~2.7V vdimm (auto) before the mod, now showed 2.04V (2.29V with nothing but the blank board, no CPU or RAM. :D ). OK, that didn't scare me, as I only had my old 256MB BH-6 RAM in there. :fact: :eek: . I'd say that is kinda low and I wasn't surprised when the system refused to boot.
Perhaps I only have to add a old fashioned vmod and up the volts to about 2.5V or something like that for boot up.


In a few weeks I'll have more time to look into that prob, but for now, I'll have to put it aside.
Just wanted to let you know.



Edit: OK, got something mixed up. Corrections made.


Edit 2: Got it!!! :toast: :banana: :

This is the chip you were all looking for:

http://www.richtek-ic.com.tw/ch/coun...173A-14(1).pdf

It's the regulator chip which also controls the protection mechanisms!
Well, and it looks like it's not OverVoltageProtection kicking in, but rather OverCurrentProtection. ;)

Update:

Pic of the RT9173A, right next to the DIMM slots

http://www.fh-augsburg.de/~madocer/D...%20RT9173A.jpg