safe micron DDR3 voltage

[fredyama]

thank you very much for sharing this information here fredyama

and, correct me if im wrong fredyama, but the sticks were not degrading, both sticks just suddenly stopped working completely.

You are quite welcome, saaya-san
Yep, unfortunately the both sticks were suddenly dead in my case

[Stelios]

I have also one stick degraded with 2.40V . The stick lose about 70Mhz within 2-3 minutes. The other stick from the pair continued to work properly.
So far i found 2.35V to be the max for benches .

[saaya]

I havent had first hand experience confirming so, but the direct operset, further I've hurt of ppl going even further then myself..

you mean you heard of people going further then yourself with vdimm. yes, i have heard of 2.8v, but then again, every chip acts different, and look at all the people posting they did damage or kill memory at 2.4v... i dont think they are making it up

stelios, so now the sticks max speed is 70mhz less than before you tried 2.4v?
what board was it on? did you use a maximizer or the mainboards vdimm?
did you meassure real vdimm? how did you cool the memory?

[saaya]

here, i made a new thread with a list of the vdimm overvolting results:
http://www.xtremesystems.org/forums/....php?p=2887032

please post your results if you have time, thanks in advance!

[Stelios]

stelios, so now the sticks max speed is 70mhz less than before you tried 2.4v?
what board was it on? did you use a maximizer or the mainboards vdimm?
did you meassure real vdimm? how did you cool the memory?

I have been using a maximus extreme board .No maximizer but rams were cooled with a modded corsair dominator airflow .
The fans were changed with 3 delta 9500rpms each one (As a result the rams were always cold) .

I have to admit that it's the first mainboard i haven't used a multimeter to measure the real voltage . Applying 2.3V to bios shows in hardware monitor page (in bios) 2.37V approx .
In the first 2 months i never set the voltage more that 2.30V . The clocks were always stable .
The only time i put 2.35V in bios was enough to degrade one of the two sticks . And we're talking for a few minutes run ... less than one 3dmark loop and one stick degraded.

According to your new thread , the real voltage when you give 2.35v is 2.39v ??? I have to measure it to check my mainboard.

[saaya]

I have been using a maximus extreme board .No maximizer but rams were cooled with a modded corsair dominator airflow .
The fans were changed with 3 delta 9500rpms each one (As a result the rams were always cold) .

nice!
any pics?

I have to admit that it's the first mainboard i haven't used a multimeter to measure the real voltage . Applying 2.3V to bios shows in hardware monitor page (in bios) 2.37V approx .
In the first 2 months i never set the voltage more that 2.30V . The clocks were always stable .
The only time i put 2.35V in bios was enough to degrade one of the two sticks . And we're talking for a few minutes run ... less than one 3dmark loop and one stick degraded.

so... were you using only one stick to test more vdimm? or did you use both sticks, but only one degraded and the other was still ok even after 2.35v bios?

According to your new thread , the real voltage when you give 2.35v is 2.39v ??? I have to measure it to check my mainboard.

well, as i wrote there, it changes from board to board and depends on what vdimm you select. for example, i have played with sevral p5k3 boards, they all overvolt slightly different.

1.90v bios on p5k3 dlx =
made in taiwan = 1.95
made in china = 2.00

but im not sure if it really depends on where the board was made... i think there are probably boards made in china that overvolt 1.9v to 1.95, and boards made in taiwan that overvolt 1.9v to 2.00...

and it depends on what vdimm you set, on my p5e3 non deluxe, setting 1.9v results in ~1.950v, if i set 2.3v it results in 2.31v...
on one p5k3 i tried its the other way around, 2.25v results in 2.30v, 2.0v results in 2.03v...

the numbers in that thread are simply the highest reported overvolts. So basically the numbers mean, if you use this board, and set this voltage, it will not overvolt more than xxx, whatever it says in the list. So you can be sure that you are still below a certain voltage. But if you want to know the real actual vdimm, you should always meassure it, and dont trust the onboard monitoring in bios and in windows!

about chipset voltage, im running 1.7v in bios on all p5k3 boards for months without any problems.
real chipset voltage is 1.75-1.80v depending on the board.
The nb heatpipe heatsink thing is cooled by a fan though, we use scythe katana heatsinks that blow the air directly over the mosfet area, very nice design from scythe btw

[dinos22]

m.beier dinos22 and massman, sorry in case i sounded a bit rude, i am a bit dissapointed since i got the impression that you, instead of beeing happy about the infos i posted here, only ask for more and only try to disprove things instead of trying to work on something that would benefit all of us. if thats not your intetion then i hope you accept my apology.

i updated the first post and changed the range, 2,2v seems to be safe as well, but for 24/7 opteration over 1-2 years im not 100% sure. there might be some sticks that do fail at 2.2v after a long period of time.

about benching, 2.3v seems to be safe, even 2.35v, but above that the risk of the memory being damaged increases notably. what is your guys experience with 2.4v? has any of you run 2.4v for an extended period of time?
on what board and with what cooling?

maybe it would be useful to add a list of how mainboards overvolt vdimm.
lowest and highest reported values

P5K3 dlx 0.030 - 0.055
P5E3 non dlx 0.025 - 0.055
P5E3 dlx 0.050 - 0.120v
blitz xtreme 0.070
maximus xtreme 0.036V to 0.038
P5E3 WS Pro 0.095V to 0.098

you shouldn't take it personal
that's your problem

info is useful of course

the best advice i would give ppl in terms of overvolting or undervolting boards is to go buy a DMM and measure every board you test on your own so you know where you're at

i pretty much always check...i know johnny would be ty with me but i usually probe some of thespots around the RAM sockets to find the readings and then change values in bios and compare again to make sure they are the right ones

johhny thinks that the best readings are on one of the pins inside the RAM sockets but that's too complicated and difficult

Asus overvolting is shocking....crappy QC i guess




also while we are talking about high vdimm testing....one friend of mine used 2.6v on corsair 1800MHz Dominator kit for two days testing in the lab and the RAM has not skipped a beat nor has there been any reduced OCing

my kit seems to out at around 2.25vdimm or so which is a shame as it would have been awesome to see how it scales at higher volts.........i don't think i'd be sticking 2.6v into them though...that's just crazy....guess he doesn't care since he had 3xkits to use hahaha

this Crucial RAM i have seems to like volts but crucials die so much in DDR2 flavour i'm even freaked out using 2.2v on this ddr3 ram kit lol knowing they will probably die >>> what's worse i think this 2GHz CAS9 kit is based on the new JNL ram or whatever it's called which is made on a smaller process...that would make them even weaker i'm sure eek

[saaya]

well how should i not take it personal? :P
im trying to give everybody a heads up here and people keep asking for more and more and more. i feel like offering a hand to help and getting my arm twisted behind my back asking me to burn a few dozen sticks of memory

D9JNL isnt stable above 2.1v for me, some not even above 1.7v
so i dont know if it degrades at high vdimm, i didnt see a point in even trying it 2.1v works fine for weeks so far...

[KTE]

this Crucial RAM i have seems to like volts but crucials die so much in DDR2 flavour i'm even freaked out using 2.2v on this ddr3 ram kit lol knowing they will probably die >>> what's worse i think this 2GHz CAS9 kit is based on the new JNL ram or whatever it's called which is made on a smaller process...that would make them even weaker i'm sure eek

I'm cautious with these sticks, damaged one trying 2.25v BIOS on them which was less in reality as board undervolted 0.038v at 2.1v the last I measured. Only set those volts for around 3 minutes, booted XMP profile and then rebooted.
Set back to 1.5v 1066 and booted up Memtest CD. Failed Test#4 ~60% and Test#5 was over 1000 errors inside a few seconds.

Took first stick out, tested again, no problem.

Put back taken-out stick, took out put-in stick and same problems again.

Swapped slots, tested again, problems.

Swapped PSU, tested again, problems.

Got a freind to test on his board, same problems.

Second stick lived fine but since then I've backed off to 1.8v for it.

[M.Beier]

well how should i not take it personal? :P
im trying to give everybody a heads up here and people keep asking for more and more and more. i feel like offering a hand to help and getting my arm twisted behind my back asking me to burn a few dozen sticks of memory

D9JNL isnt stable above 2.1v for me, some not even above 1.7v
so i dont know if it degrades at high vdimm, i didnt see a point in even trying it 2.1v works fine for weeks so far...

Well well well

Its a public forum, and nobody has been elected champion of anything, and we have alot of very smart people on different areas - many different oppinions... - Its impossible to tell how much experience each member has, thats why one has to assume its always pro's, and respect one anothers oppinions...
Lets just say, your mobo voltage thread, I find it a bit funny, because, not you, nor myself, nor anybody else has a complete sight, nobody knows, and its almost 100% certain if one seeks on the various forums, the numbers will be very different - so the "highest reported", is a bit off IMHO....

Its important at all times not to state having better knowledge then others, even if one has - hint: Everybody knows that the Japanies SPI'ers are gods in the contest, however, they're always very humble... Same with Sham and 3D

[saaya]

kte, so you killed a stick at 2.2v real pretty much?
so im not the only one who managed to do that!
pheew, i feel better now
what cooling?
what board?
is that kit with d9jnl or d9gtr?

about the vdimm list.
every board overvolts vdimm different.
even at the same setting it depends if you use 2 sticks or 4
double sided or single sided
depends on what chips the mem uses
depends on the psu used...

and if your board overvolts 0.05v at a setting going one step up might change the whole thing. so i didnt see a point in trying to find an average overvolting value. especially since there are china and taiwan models and different revisions that change the overvolting behaviour again.

the idea i had behind that list, is people can check what board they use, and regardless of what vdimm they set, they can be pretty sure it will NOT be higher than what they set plus the highest reported overvoltage for that board. still not 100% sure, but pretty sure...


hmmm a lesson on beeing humble from m.beier

First to break 1000mhz CL6

stay tuned
Ram records coming up!

[M.Beier]

Hey I didnt state myself as being humble... Im an a**hole, thats no secret
Im danish, thats the reason, you know?

For the record, during my marketing competition, the end comment (in a friendly way) "Yeah Marc, now you've just confirmed that your an arrogant pr!ck", was funny as hell =)

My point was, that it would be very quire if everybody agree's, its not very likely, and there's no reason for taking it personal
- And nobody has full sight, its important too respect one another.
... Honestly I didnt even remember signature & avatar, disabled displaying it

[chris.y2k.r1]

... I wish I didn't have to read all the childish trite crammed between the actual information that the author and others are trying to convey and collect. It's really annoying.

[KTE]

kte, so you killed a stick at 2.2v real pretty much?
so im not the only one who managed to do that!
pheew, i feel better now
what cooling?
what board?
is that kit with d9jnl or d9gtr?

What I noticed is, many boards actually increase vDroop and vDrop the higher the voltage goes (without any feature to disable it) and they did similar for vDDR.

The board was GA-P35T-DQ6 with that Cruical Ballistix 1600 (XMP) and the board was undervolting and the higher you went, the more it undervolted.
So if at 2.1v real was 2.062v then at 2.25v real must have been below 2.21v, I would expect around 2.2v real was spot on.

Cooling was a Delta AFB1212SHE 53dBA 152 CFM 3700 RPM, IC serial was covered up by a big Ballistix logo but they were most likely D9GTR and probe temps inside the heatspreader slots was 30C. PSU had stable rail outputs, Corsair HX 520. I only had it for around 25 mins before killing the one stick, I didn't expect it reading here with guys touting 2.3v as 24/7 safe priorly but sure, I won't be giving that again unless I can risk $600 losses

[saaya]

kte, yeah, i had that same board here and... its a weird design...
as soon as i increased the chipset voltage above stock, i could only set vdimm to 1.8v max, above that the board refused to boot. very very weird... lower chipset voltage allowed for more vdimm... quite bizzare...

hmmmm i will edit the list in the first post and will declare 2.2v not safe then...
thanks a lot for the details

[Achill3uS]

nice thread saaya, i never used more than 2.00-2.05v for 24/7 but for testing 2.25-2.30v (aircooled w/ 9cm fan) many times and no degradation expected with 1800 CellShock's
(on P5K3 Deluxe +0.04vdimm real to bios set)

[saaya]

do you happen to know the name of that guy in hungary who ran 2.4v with his ocz micron ddr3 and degraded it that way?
could you maybe ask him for some more details? it degraded but he kept running it at 2.4v for weeks and it didnt die, right?

[Achill3uS]

do you happen to know the name of that guy in hungary who ran 2.4v with his ocz micron ddr3 and degraded it that way?
could you maybe ask him for some more details? it degraded but he kept running it at 2.4v for weeks and it didnt die, right?

yep right, he used for weeks 2.4v and it degraded, finally was stable only at stock clocks/timings, but didn't die (on Blitz Extreme as i know)...

[saaya]

can you get some more details?
like how soon did they degrade, immediatly?
did they degrade a lot quickly or slowly bit by bit?

if he used 2.4v in bios on blitz extreme that means 2.43v real right?
thats about the speed a couple of kits died at so it makes sense...

[Achill3uS]

can you get some more details?
like how soon did they degrade, immediatly?
did they degrade a lot quickly or slowly bit by bit?

if he used 2.4v in bios on blitz extreme that means 2.43v real right?
thats about the speed a couple of kits died at so it makes sense...

im not contact with the guy, he is a jerk sry
but as i remember when talked with him, "this is an OCZ DDR3 it needs to stand 2.4v or not?" nope dude "hmm, i used it on this voltage for ~1 month, but nowdays sometimes i've got freeze, and my memorys are completely unstable above stock speed/timings" like i said to you saaya
short story

[saaya]

hehe, ok then
i dont wanna push you to suck up to some guy you dont get along with to try to get some useful information from him

[Jaco]

What about DRAM drive strength ?
Could that have an influence on the RAM degradation/failure ?

[TheHardCase]

Things that make you go hummm, my viper kit says 1.8 VDimm @ 7-7-7-18 @ 2T.
Funnything is I had to apply 1.9 VDimm real for them to operate at stock.

I also tested OCZ OCZ3P1600EB4GK & it called for 1.9 VDimm, they have problems on this Asus Striker II Extreme motherboard huge, nothing but frezzing, sticks tested fine in memorytest ect. So I tried 2.0 VDimm same problem & when I toched them they were hot! I mean hot. The OCZ are 7-6-6-24@ 2T & could not get stable from 1.90 VDimm - 2.0 VDimm Real.
I also relaxed timming to 8-8-8-24 @ 2T 1.90 VDimm & same problem.
I also moved them to slots 2 & 4 same problem.

As it stands I'm running Viper 4 x 1GB 1600MHz overclocked to 1800MHz on 2.14 VDimm - 2.16 VDimm Real, 100% Stable.
I'm going to run 24/7 & we will see what happens about voltage.
They are very cool to touch not like the OCZ hummm.

[RyderOCZ]

Trfc plays a big role in how much voltage is needed... with DDR3 pushed upwards of 1600+... Try Trfc at 72 with near 90 at 2GHz ram speed. It may allow lower voltages.

[M.Beier]

Trfc plays a big role in how much voltage is needed... with DDR3 pushed upwards of 1600+... Try Trfc at 72 with near 90 at 2GHz ram speed. It may allow lower voltages.

Interesting...
But I must admit, isnt 72 quite high in first place? I've been running 45 trfc DDR3-2000???

In my experience: ras to CAS is the true sinner when it comes to increasing voltage...