BH5 KILLER READ IF YOU HAVE DFI

Printable View

Show 100 post(s) from this thread on one page

05-25-2005, 10:07 AM
Millyons

Why don't ppl with dead ram post, what mobo rev they got........... and from what i can remember wasn't there something concerning hardware that was changed in the early revision (not sure of this, don't quote me)
05-25-2005, 10:38 AM
hovo73

I think I read something on DFI-street from Oskar that said there is only one revision. The only difference is the BIOS it ships with.
05-29-2005, 04:06 AM
LostInTennessee

This is the most interesting thread I have read in a long time, and I don't even own a AMD 64bit MB.
But I think the answer is "out there" in the vapor.....
05-29-2005, 06:19 AM
Snark42

did this trhead just lose it's 'sticky' status? (At least...I thought it was stickied)
05-29-2005, 08:49 AM
psycho303

hello all,

just want to say that this mystery happend to me too. i killed two corsair bh-5 with 3.5v , each´died in the 2nd slot, one OCZ VX UTT already with 3.4v and one muskin bh-5 with 3.3v :mad: 512MB each stick=2 Gig RAM dead :mad: . my board was faulty, i had A00 rev. and i could not use the onboard nvidia lan, always machine freezes. now got my second DFI NF4 and using TCCD, no more problems, dont wanna use anymore highvolt chips on that board.
05-29-2005, 09:44 AM
Snark42

Quote:

Originally Posted by psycho303

hello all,

just want to say that this mystery happend to me too. i killed two corsair bh-5 with 3.5v , each´died in the 2nd slot, one OCZ VX UTT already with 3.4v and one muskin bh-5 with 3.3v :mad: 512MB each stick=2 Gig RAM dead :mad: . my board was faulty, i had A00 rev. and i could not use the onboard nvidia lan, always machine freezes. now got my second DFI NF4 and using TCCD, no more problems, dont wanna use anymore highvolt chips on that board.

were you useing the 5v jumper in all cases? I only ask because it could be read 2 ways when you said...
"one muskin bh-5 with 3.3v"
do you mean 3.3vdimm or while it was on the 3.3v jumper option?
05-29-2005, 03:42 PM
burningrave101

Quote:

Originally Posted by muzz

IMO that bogus 480W rating is just DFI trying to cover their ass from folks that buy $25 600w power supplies with 18A on the 12v rail.
I have the PS420, and it's more than enough to cover my board, and the menial current draw I am asking it to supply.

Yea i've disputed it over at the DFI forums but it doesn't seem to sink in that wattage rating doesn't mean crap. A power supply should have a beefy +12V rail and be able to provide a combined load of 400w+. If it can do that and its a quality brand then thats good enough for the majority of systems.

There are 480w+ PSU's that are easily outdone by the PowerStream 420w yet because of its wattage rating isn't higher its not recommended.
05-29-2005, 03:44 PM
burningrave101

I was considering buying some OCZ BH-5 memory to replace my OCZ PC3200 Platinum Rev2 TCCD but now i'm sorta scared to give it the higher voltage if this board is killing DIMMs.
05-29-2005, 03:51 PM
ReelMonza

Rev2s may not be TCCD anymore...
05-29-2005, 04:50 PM
Kosmic71

yea, I've heard of that. rev2 might be TCC5 now. :slapass:

he he my rev2 sticks just rock, must be TCCD. :p:
05-29-2005, 05:38 PM
GregP24

Quote:

Originally Posted by ReelMonza

Rev2s may not be TCCD anymore...

I got my Rev2 a few days ago and IT IS NOT TCCD! I cannot get it stable above 230mhz no matter what I do.

I have a question regarding VX. Since th 5v jumper is known to kill the mem, wouldn't using the DDR booster with the 3.3v jumper work just fine? One of my VX sticks died a short while ago(I"m not sure if it was the stick on slot 2 or not). I bought the Rev2 thinking it was still TCCD. I was planning on running 280x10 and calling it good. But since I"m limited to 230x11, I'm leaning towards putting the VX back in once it gets back from RMA. I just don't want to lose another stick.
So, wouldn't using the booster bypass the problem with the Ultra-D?
05-29-2005, 06:43 PM
wtz54321

Quote:

Originally Posted by psycho303

hello all,

just want to say that this mystery happend to me too. i killed two corsair bh-5 with 3.5v , each´died in the 2nd slot, one OCZ VX UTT already with 3.4v and one muskin bh-5 with 3.3v :mad: 512MB each stick=2 Gig RAM dead :mad: . my board was faulty, i had A00 rev. and i could not use the onboard nvidia lan, always machine freezes. now got my second DFI NF4 and using TCCD, no more problems, dont wanna use anymore highvolt chips on that board.

Err ... Why you still choose to kill the 2nd , 3rd , 4th dimm after the previous one died ? :confused:
05-30-2005, 02:28 AM
burningrave101

Quote:

Originally Posted by Kosmic71

yea, I've heard of that. rev2 might be TCC5 now. :slapass:

he he my rev2 sticks just rock, must be TCCD. :p:

It will say v1.1 on the sticker if their TCC5 instead of TCCD. I got mine last year so i know their TCCD.

Quote:

Originally Posted by GregP24

I got my Rev2 a few days ago and IT IS NOT TCCD! I cannot get it stable above 230mhz no matter what I do.

I have a question regarding VX. Since th 5v jumper is known to kill the mem, wouldn't using the DDR booster with the 3.3v jumper work just fine? One of my VX sticks died a short while ago(I"m not sure if it was the stick on slot 2 or not). I bought the Rev2 thinking it was still TCCD. I was planning on running 280x10 and calling it good. But since I"m limited to 230x11, I'm leaning towards putting the VX back in once it gets back from RMA. I just don't want to lose another stick.
So, wouldn't using the booster bypass the problem with the Ultra-D?

You may just be having a compatbility problem with that San Diego. I can't get my OCZ PC3200 Platinum Rev2 w/ TCCD to run worth a crap with my Venice. I have to use a divider and 2T just to get it stable in windows at higher speeds. If i try to run 1:1 1T it just locks up and reboots every time when windows starts to load. I think we may just need a BIOS update to address the problem because i'm not getting errors in Memtest86.
05-30-2005, 02:36 AM
burningrave101

Quote:

Originally Posted by DFI Forums

Memory Death:

We use two stages of linear regulators to generate the VDIMM on the DFI NF4 motherboard.

1st regulator from the psu’s 3.3v or 5v line to VDIMM sets the target voltage to 4.1v, or 3.2v (depending on what how the J17 jumper is configured). The 2nd regulator controls VDIMM voltage to the memory modules themselves.

Explanation as follows:

In 3.3v source jumper mode, the 1st regulator will output 3.3v to the 2nd regulator. The 2nd regulator handles the voltage output to the memory modules. (2.5v-3.2v)

In 5v source jumper mode, the 1st regulator will output 4.1v to the 2nd regulator. Then the 2nd regulator will handle the output to the memory modules. (2.5v-4.0v).

ATTENTION:
The DRAM voltage as originally supported by the spec is up to 2.85v only. Jumper J17 allows you to select beyond the defined spec. Although the function is supported (at the hardware level), selecting a higher voltage may cause unstable current supplied to the system board which may subsequently cause damage to the CPU and DIMMs

(copied from the warning sheet supplied with the DFI NF4 motherboard concerning the J17 5v jumper and also found on page 9 of later revision NF4 motherboards (A02 and later))

Using this design, the 5v source voltage will not exceed the VDIMM source from the 1st regulator (4.1v). This circuit will not supply DRAM voltage higher than 4.0V(4.1V-0.1~0.15V) no matter what kind of user error or bios error.

Using this type of linear voltage regulation is very resistant to any voltage spiking and in the DFI labs and in independent testing no spiking at the memory sockets has been observed.

To conclude, there has been no evidence that the DFI NF4 motherboard has been responsible for any memory death. You can choose to not believe this if that is what you want to do. We have done extensive testing and found no evidence that the board kills memory modules though, and that is the truth, no matter how much some of you might want to flame and shout that you 'know for sure' that the board does kill memory.

http://www.dfi-street.com/forum/showthread.php?t=12687
05-30-2005, 04:10 AM
[XC] 4X4N

Quote:

Originally Posted by burningrave101

It will say v1.1 on the sticker if their TCC5 instead of TCCD. I got mine last year so i know their TCCD.

You may just be having a compatbility problem with that San Diego. I can't get my OCZ PC3200 Platinum Rev2 w/ TCCD to run worth a crap with my Venice. I have to use a divider and 2T just to get it stable in windows at higher speeds. If i try to run 1:1 1T it just locks up and reboots every time when windows starts to load. I think we may just need a BIOS update to address the problem because i'm not getting errors in Memtest86.

I have the new v1.1 with tcc5. Using 5/10-1 bios. I can bench and do most things I want with 280-2.5-3-3-8, but can't make it all they way through 32m of super-pi or prime for more than 30-40 min. No memtest errors. Even using windows memtest no errors. Ran it to over 1000%. Tried 2.5-4-4-8, same result. Also tried up to 3.1v with no help. I can also bench at 9x300 2.5-4-3-8, again no memtest errors, but fails prime right away. The highest prime stable with 1:1 is 10x265 2.5-3-3-6. I have run prime over 7 hours at 10x280 with 180 divider, so I know the cpu can go to 2.8. Now I'm wondering if this is a problem with venice, bios, or is it just the limitation of the ram. Any suggestions?
05-30-2005, 04:29 AM
TekXoID

Quote:

Originally Posted by 4X4N

I have the new v1.1 with tcc5. Using 5/10-1 bios. I can bench and do most things I want with 280-2.5-3-3-8, but can't make it all they way through 32m of super-pi or prime for more than 30-40 min. No memtest errors. Even using windows memtest no errors. Ran it to over 1000%. Tried 2.5-4-4-8, same result. Also tried up to 3.1v with no help. I can also bench at 9x300 2.5-4-3-8, again no memtest errors, but fails prime right away. The highest prime stable with 1:1 is 10x265 2.5-3-3-6. I have run prime over 7 hours at 10x280 with 180 divider, so I know the cpu can go to 2.8. Now I'm wondering if this is a problem with venice, bios, or is it just the limitation of the ram. Any suggestions?

I know, I can run up to nearly 3.1GHz in memtest without error but 2.84GHz
is the highest that it can actually run with prime stability for 1HR+.

First, load up 510-3 which is the best of the 510's for TCCD.
(Trust me, I have been sitting here for the last two days sitting here experimenting.)

Then make both of your drive strength settings as weak as possible
and go up from there. (Use odd number levels only with TCCD.)

Also, disable interleave, the actual performance loss is minimal and the O/C headroom
gained is well worth the minor loss. Might need some more volts on the core as well?

Edit: That's very odd, my TCCD walls at 265MHz too when I have both sticks in. :stick: ???
05-30-2005, 05:01 AM
[XC] 4X4N

Ok, I'll try the 5/10-3. Have drive strength at 5 and 2 now, I'll see if 3and 1 help. I haven't tried to disable bank interleave, I thought that with it enabled allowed a better clock. I'll see if it helps. I also tried upping the vcore as well with no help. I have been checking around, and it seems most people with this newer version seem to be maxed around 265 as well. :(
05-30-2005, 05:32 AM
Electric Bill

Quote:

Originally Posted by burningrave101

It will say v1.1 on the sticker if their TCC5 instead of TCCD.

Got my RAM (2x512 platinum rev 2) just over a week ago from The Egg. I was sure that it had to be TCC5 at this late date. Well it doesn't have v1.1 on the sticker and I'm running Prime (tested 14hrs) and 3DMark01 stable at 10x280 on air with the stuff in my Sig. Can Andy or someone else from OCZ confirm this so I can steer the folks over at PC Per/DFI in the right direction?

Bill
05-30-2005, 05:35 AM
lizardmech

My vx 4000 died after 9 weeks, is it safe to use them with my DFI nf4 when I get the replacement?
05-30-2005, 06:54 AM
pazza316

My experience with my latest batch is yes you are safe using the 3.3 mod rarther than 5 volt jumper.
05-30-2005, 01:12 PM
wtz54321

Quote:

Originally Posted by pazza316

My experience with my latest batch is yes you are safe using the 3.3 mod rarther than 5 volt jumper.

IF the dram is not killed by high voltage or what we think that 5V jumper kill the dram , be careful to use the dram above 3.0V even you use a DDR booster or 3.3V MOD ... Until now , we do not know if it's the UTT is weak from high voltage + long term usage ...
05-30-2005, 04:38 PM
lizardmech

Quote:

Originally Posted by wtz54321

IF the dram is not killed by high voltage or what we think that 5V jumper kill the dram , be careful to use the dram above 3.0V even you use a DDR booster or 3.3V MOD ... Until now , we do not know if it's the UTT is weak from high voltage + long term usage ...

Thats annoying, my last set needed 3.4v just to reach 250mhz 3-3-3-8. It wouldn't do much over 210 when I had it on the 3.3 jumper.
05-31-2005, 01:18 PM
Millyons

Quote:

Originally Posted by wtz54321

IF the dram is not killed by high voltage or what we think that 5V jumper kill the dram , be careful to use the dram above 3.0V even you use a DDR booster or 3.3V MOD ... Until now , we do not know if it's the UTT is weak from high voltage + long term usage ...

From this thread this is not only UTT dying, its older BH-5 also............
05-31-2005, 01:39 PM
jiff

It would be nice if someone could confirm a death of a stick using either the 3.3rail jumper trick or a booster. ;)
05-31-2005, 02:02 PM
GregP24

Quote:

Originally Posted by jiff

It would be nice if someone could confirm a death of a stick using either the 3.3rail jumper trick or a booster. ;)

We'll know soon. I'll be using my booster with my VX once it gets back from RMA.

Show 100 post(s) from this thread on one page