Max safe D9GMH 24/7 volts?

Printable View

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

02-05-2008, 05:21 PM
railmeat

Quote:

Originally Posted by BTK

id say without a fan 2.1V
2.2V if you have good airflow

2.3V + benching quick runs with a fan

2.7V+ - dont do it- unless your going for some insane bench

the fact that micron IC die at high volts, but needs it to overclock (and they such good oc'er) does not help......

i run my D9GMH no fan @ 2.1V - 24/7
4-4-4-12 450MHz

6 months and still perfect

your the most realistic member of this forum....;) agree %100 2.2v with fan at good clocks-timings last forever.anything above(2.3v+) that is rma time sooner or later.
02-05-2008, 05:51 PM
R*r

i run my d9 @ 2,4v 27/7 +2x6cm fan,room temp ard 24-26'c its been abt 6mnth

i think im just lucky :shrug: the stick still doin alright :yepp: bench at 2,6v dont know how long it would last...
02-05-2008, 11:22 PM
Rammsteiner

Im running 2.3V trough them with Weak DRAM Drive Strength.

CKE, CS/ODT, Addr/Cmd and MEMCLK Drive Strengths all at 1.25x (only MEMCLK is higher than default here, the other three or 0.25x lower than default. EPP has one of them even at 2.0x).

DATA and DQS Drive Strengths at 1.0x.
02-06-2008, 03:04 AM
nzbleach

2.0-2.1 24/7. Benching is fine at 2.5-2.6 Volts.

I've gone through 3 sets of Crucial D9 @ 2.2 and 2.3 volts for 24/7 across 3 different motherboards. Never again.
02-06-2008, 03:22 AM
Chosen.

Quote:

Originally Posted by Fr3akyPete

I would love to run my 6400 GBHZs (2x1gig) with more voltage than 2.1V.
But my Gigabyte GA-965P-DS3P always reboots after I set +0,4 (2.2V) in BIOS. After I go back in BIOS..PC Health Monitor show 1.8V DDR Voltage "FAIL".
I don't know what to do!

Buy a man's motherboard? :p:
02-06-2008, 07:26 AM
Jaco

max safe voltage 24/7 for D9GMH.

that's easy : 0 volt :p:
02-06-2008, 08:07 AM
neouser

however I have a friend here benched at 3.5v with cell shock d9gkx. They done 1066mhz 3 3 3 5
02-07-2008, 06:16 PM
initialised

Quote:

Originally Posted by WeStSiDePLaYa

What are your opinions. Ram in sig.

See the thread I just posted on OCZ Flex @ 850 3-2-2. I run them comfortably at 2.7V (set to 2.6V in BIOS) on water. Anything less isn't enough for good overclocks or low latencies. They were sold to me as 6400/800 but CPU-Z and SPD-Z show them to be 9200/1150. Like it says on the tin (well Aluminium and Copper) they are very flexible. Nominal voltage is 2.35+/- 0.05.

I was told by a colleague yesterday that the 9600 uniits will do 10400/1300, watch this space and I'll see if my 9200/1150s can get close...
02-07-2008, 06:31 PM
initialised

They may not be dead!

Quote:

Originally Posted by WeStSiDePLaYa

Well, just an update, my FlexXLC has died running only at the volts specified in the EPP profile(slightly undervolt on the board). @:

I have had the odd problem with mine when going for high speeds or low latencies (or both) where the board wouldn't post even after clearing the BIOS. Solution was to install some cheap RAM, boot, shutdown, reinstall the Flex, boot, shutdown, remove cheap RAM, boot, tweak until it happens again.

Were you water cooling the Flex?
02-07-2008, 06:52 PM
WeStSiDePLaYa

Quote:

Originally Posted by initialised

I have had the odd problem with mine when going for high speeds or low latencies (or both) where the board wouldn't post even after clearing the BIOS. Solution was to install some cheap RAM, boot, shutdown, reinstall the Flex, boot, shutdown, remove cheap RAM, boot, tweak until it happens again.

Were you water cooling the Flex?

Nope no watercooling. And yes, they have problems booting, but one stick is 100% dead. no booting.
02-07-2008, 06:52 PM
initialised

Here comes the science

I case you're not sick of me yet a little tip on electromigration: it is a function of voltage and temperature. At high temperatures resistance goes up leading to higher current, without sufficient cooling the heat stays on the chip so you need more current so you make more heat and need even more, positive feedback until something breaks down. This is why active cooling is important, and also why components that run hot overclock well when cooled well as heat is the limiting factor so shifting it raises the limit.

Also electromigration is much less of an issue in copper interconnect than it was in aluminium, due to the higher atomic weight of copper atoms they are harder for the electrons to push around, I'd go into the equations but it's years since I worked in a fab or on semiconductor research.
02-07-2008, 06:57 PM
initialised

Quote:

Originally Posted by WeStSiDePLaYa

Nope no watercooling. And yes, they have problems booting, but one stick is 100% dead. no booting.

You have what sounds like a decent water set up and RAM with built in waterblocks and you didn't get it wet?

As I remember the stock 2.4V is for watercooling. I didn't like taking them over 2.2V on air. I'd RMA it and get the replacement on water, Swiftech do F-splitters which is almost purpose built for this RAM, get some.
02-07-2008, 07:14 PM
WeStSiDePLaYa

Quote:

Originally Posted by initialised

You have what sounds like a decent water set up and RAM with built in waterblocks and you didn't get it wet?

As I remember the stock 2.4V is for watercooling. I didn't like taking them over 2.2V on air. I'd RMA it and get the replacement on water, Swiftech do F-splitters which is almost purpose built for this RAM, get some.

it would affect my cpu temps too much to add all the extra restriction and complication. Plus I like to change mem.
02-07-2008, 07:34 PM
RyderOCZ

Quote:

Originally Posted by WeStSiDePLaYa

Well, just an update, my FlexXLC has died running only at the volts specified in the EPP profile(slightly undervolt on the board). @:

Kind of disappointing that their own specs caused them to die.

They always stayed cool to the touch, but died anyways.

One stick is 100% dead, other can't even run stock EPP.

Hopefully Tony sees this and helps me out. Never had to RMA any of my many OCZ sets yet.

Maybe D9's just have poor reliability at any voltage?

Tony doesn't do RMA's... I do.

Email is in my Sig. :)
02-08-2008, 03:48 PM
initialised

Quote:

Originally Posted by WeStSiDePLaYa

it would affect my cpu temps too much to add all the extra restriction and complication. Plus I like to change mem.

What are you running that Wolf at? I had one stable at 3.6GHz with no thermal issues (4 hours 3D mark, overnight Prime95) and a mate had one stable at 4.2GHz on air, both were on 780i boards. Unless you are at 70C+ loaded, have a weak pump (<400lph) or an under cooled radiator I doubt you would have any issues besides modifying the loop.

Water cooling the RAM might actually improve CPU temperatures by dumping the heat from the RAM in the radiator instead of dumping it onto the motherboard near the CPU and the surrounding air.

As for complication two F-splitters and a foot or less of 1/4 inch hose.

When I change memory I'm going to keep the coolers. I think OCZ should sell the coolers separately (hint, hint @ RyderOCZ).
02-11-2008, 10:23 PM
WeStSiDePLaYa

Well just an update people.

I got my ram exchanged in store for a new set.

The new set is running 1200mhz at 5-4-4-18 w/2.2v 100% orthos stable for 9 hours.
02-12-2008, 02:08 AM
bf2142lol

One can expect ram to degrade In overclockability after prolonged periods of high voltage use. My previous dims, ocz plat rev.2 did 900-950 cl4 or 1100 cl5 @ 2.3v, I did a lot of suicide runs after than at around 2.6v and they were never the same again, lol. They could hardly do 900cl4 with 2.35volts
02-13-2008, 01:16 AM
The0men

Well I look at it this way... This is XS, and it's hardly worth having D9 unless you want 2.3v+ in it daily. If you want low voltage you have qimonda which is probably better around 2.1v mark.
02-13-2008, 01:57 AM
zsamz_

Quote:

Originally Posted by The0men

Well I look at it this way... This is XS, and it's hardly worth having D9 unless you want 2.3v+ in it daily. If you want low voltage you have qimonda which is probably better around 2.1v mark.

why 2.3v+ i'm running 1200 mhz 24/7 with only 2.12v
02-13-2008, 02:19 AM
R*r

Quote:

Originally Posted by zsamz_

why 2.3v+ i'm running 1200 mhz 24/7 with only 2.12v

run it 2,3v so can get higher n higher :clap:
02-13-2008, 05:49 AM
RyderOCZ

Quote:

Originally Posted by WeStSiDePLaYa

Well just an update people.

I got my ram exchanged in store for a new set.

The new set is running 1200mhz at 5-4-4-18 w/2.2v 100% orthos stable for 9 hours.

Have you tried lower voltage? These are quite possibly not the same as you had before, too much voltage is bad on these new modules.
02-13-2008, 09:39 AM
zsamz_

Quote:

Originally Posted by R*r

run it 2,3v so can get higher n higher :clap:

i can do 1300+ with 2.3v but no need to push it for 24/7
i can do 4-4-4-13 1120 with my mushkins @ 2.3v but after losing 6 gigs @ 2.26v lol ,,,
i'll stick with 2.13 lol i dont need to got above 1200mhz
02-13-2008, 10:48 AM
WeStSiDePLaYa

Quote:

Originally Posted by RyderOCZ

Have you tried lower voltage? These are quite possibly not the same as you had before, too much voltage is bad on these new modules.

No I haven't yet.

These do seem slightly different though, the blister pack was spot welded closed unlike my previous set.

Do these no longer use D9?
02-13-2008, 07:50 PM
R*r

Quote:

Originally Posted by zsamz_

i can do 1300+ with 2.3v but no need to push it for 24/7
i can do 4-4-4-13 1120 with my mushkins @ 2.3v but after losing 6 gigs @ 2.26v lol ,,,
i'll stick with 2.13 lol i dont need to got above 1200mhz

whats your memory brand?maybe ill try my luck finding one like your one :up:

my balistic only make it until 630@5-5-5 2,4v :shakes:
02-13-2008, 07:54 PM
WeStSiDePLaYa

New update.

Now fully 100% orthos stable at 5-4-3-15 at DDR1200 with only 2.2v.

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

All times are GMT -8. The time now is 04:40 PM.

XtremeSystems