2.5V is right on the border of being safe...run more and you will kill your dimms in under a week.
I have seen quite a few quoting they killed DDR2 while benching now, remember you juice it up you take the risk.
T
Printable View
2.5V is right on the border of being safe...run more and you will kill your dimms in under a week.
I have seen quite a few quoting they killed DDR2 while benching now, remember you juice it up you take the risk.
T
why are some manufacturers releasing RAM with 2.4v rated timings.........are they looking for trouble like those old UTT sticks we all killed at some point
can you tell us more about electronic migration you mentioned in the other thread
i must have looks somewhere else but i haven't seen all that money dead ram sticks reported in DDR2....certainly not even close to UTT deaths we had early this and all last year
once I tested my 8000UL @ 3.15V w/no fan ...
result - mem is alive :woot: but OC @ low volts dropped :(
UTT dejavu:)
So does that mean that 2.4V with active cooling still isn't really safe for 24/7 usage?
Quote:
Originally Posted by higgins
Considering companies release dimms with stock voltage of 2.4v and tony said 2.5v is the limit of safeness i would assume 2.4v is safe.
Just a extreme Off topic:
I got my DDr2 @ 4.1v... they burned... literally. The side sticker was totally burned. Was matter of mobo :S
:rolleyes: cant imagine burning up a 500$ ram :slapass:
Got pics ?Quote:
I got my DDr2 @ 4.1v
always thought 2.5v was safe as 2.4v is being used by many manufacturers so adding .1v wouldn't be much. It's what I was planning on running my mem at.
fhpchris also said that he runs his ram at 2.65v his fatbodies. Think he's been using them for more than a week...
Fatbody is different story i think..Quote:
Originally Posted by Praxis1452
Those chips are lot bigger and produced at larger nm rate than 'regular' D9 (D9GKX, GMH etc).
The lower the produced nm is, the less volts it can take... Especially for longer times.
:D well I should be getting 4 sticks of 5400UL sometime soon so it's nice to know that I can run them at higher voltsQuote:
Originally Posted by El Snorro
Is this true even for low clock value modules ? :confused:Quote:
Originally Posted by Tony
Quote:
Originally Posted by Praxis1452
Ive been running 2.58v on my GMH for more than a week with active cooling, just because I was too lazy to lower the voltage to ~2.48 ish! LOL :(
I ran my fatties at 2.67 for a couple of days, and never used them below 2.58v, but I threw them in Yonah and the aopen board only goes to 2.15v...
heh we all get lazy sometimes.Quote:
Originally Posted by fhpchris
Darn and I was hoping I could crank 20v into them and not have to worry about them dying
Above 2.5V (extended use) and your ddr2 ram will degraded overtime, it Will needs more and more V just to run stable.
The better way imo, is to run 400ish and oc'd your cpu to the max...Unlinked preferably;)
Fatty D9 dies just as fast....voltage shows no mercy ;)
Infact its not the voltage that kills it, you need the increase in voltage to get the memory switching faster, but a by product is more heat and an increase in current used. The increase in current is what kills it as its quite easy to control the heat.
Micron is mad expensive again at this time, if you want to keep it working reduce the vdimm ;)
I have found that the modules will sometimes overheat and stop working with high VDimm. People often assume the RAM is dead. In some cases, but not all, taking the modules out of the board and letting them cool off will solve the problem.
by testing my Kingston PC 4200 between 3.3v and 3.6v during approximately 5 hours, I saw well that 4 chips D9DCD were affected by the voltages, the ram functions but it is necessary for more volts to him so that it starts correctly (2.6 / 2.7v instead of 1.8v of origin) :cool:
I would not test with the other kits out of PC 8500 which I currently have, I dont want of to damage even more the chips in D9GKX which are absolutely not the same price as D9DCD :rolleyes:
Quote:
Originally Posted by boblemagnifique
DCD? you sure its not GCT?
I sure :DQuote:
Originally Posted by Revv23
http://www.xtremesystems.org/forums/...d.php?t=122539
I've some DCD also, running at the moment 2.6v for 24/7 although I did run them for about 5 days at 2.9v (lazy me). Not noticed any degredation yet. One thing I did notice was that at 2.9v the dimms were barely warm to the touch with no fan, whereas my GKX got fairly toasty at just 2.45v
Incidentally, my GKX died after only about 2 hours benching at 3.15v :(
EDIT: To be fair, my 24/7 voltage on those was pretty high...
For my 24/7 system I keep the ram under 2.2v.
Only for benching in a controlled AC environment will I bring the voltages up over 2.5v and that is for a short period of time and that is on my bench memory only.
I use seperate semi expendable bench ram when pushing voltages. I recently killed a good set of DDR2 6400C3 on the xbx2, one module still works.
That is why we love the Hipro 5 maximizer :D
We can lower voltage from 2.7 - 2.5 and still get same results :D
Or 1 dude could
Speaking of that, just got my maximizer, and i havent overclocked one bit yet - BUT my memory previously would not boot spd with less then 2.1v and im running 100% stable at 1.85v right now. ;) A good start to say the least.Quote:
Originally Posted by Nosfer@tu
so 2.2v is deemed safe but is there any guestimate to how fast the ram will die when exposed to 2.4v-2.5v 24/7 when cooled adequatley?
yeah i run mine at 2.1v but the temptation to go to 2.7-2.8v for SPI benching is strong...........i wouldn't want to kill these sticks so they are staying up to 2.5v :(Quote:
Originally Posted by FUGGER
true, but i'll say 2.6v is the edge
That is NICE :DQuote:
Originally Posted by Revv23
This is why i love my bh5 sticks they want more and more vdimm :D
We should get manufacturers on this.
and say what?Quote:
Originally Posted by HungryForHertz
Quote:
Originally Posted by Tony
Haha, yea really. Get them to increase their voltage tolerances on their memory? Not really anymore possible, as the precautions have already been taken to their extent.
Btw Tony, you forgot to send me a sample of your new watercooled stuff. Did you lose my shipping addy? :p: :rolleyes:
increase capacity and decrease latency.Quote:
Originally Posted by Tony
I was just using my mem at 2.25v without fan for the moment, and it really gets hot... more than my ol BH5...
I've taken ram up to 3.5 volts... test ram, crap ram that I used for exactly that kinda thing... testing.
I killed it in less then 2 minutes. boot up, set the voltage, reboot, get back into bios, take the 3.5 volt bios shot, shut down.... and ram was dead.
I didnt think the ram would be that fast to die, even for crap ram.
but it did.
I always keep a set of crap junk ram for this kinda thing, so this only cost me $40 to do. that ram probably had about 100 hours of total testing on it, in the 3 volt range.
(testing voltage mods and stability and such)
I took my Mushkin 6400 to 3.47 volts to get the Cas 3-3-3 WR at 1060 mhz.
would I do it again? hellllllllll no.
I said that in my post, you want the WR, its yours.
once at 3.4 is enough for me ;)
I am just glad the ram is alive, OC'n well and seems normal.
but I dont intend to do anything like that to this ram again.
while I've obviously went above and beyond the call of duty when overclocking my ram...
I know 1 thing... 2.3 volts on DDR2 is alot, and gets hot fast.
I keep it below 2.25 everyday no matter what.
I will go beyond the limits for a OC.
but at the end of the day, I still need ram that works, and I think it's foolish to go beyond 2.4 for daily use, unless you have no intention of having that ram live long.
Hope everyone who's bought Corsair UL's in the past few weeks are reading this. Many user's planned on running 2.5-2.6v to get 4-3-2-X DDR2-1000. I was one of them, ran my 8000UL's 4-3-2-4 @ 504mhz 2.55v for over 12 hours with memtest. They still work, but I guess I won't do that again.
Do you know the chips of your memory stick ?Quote:
Originally Posted by Tony
Is it micron D9 ??
Because some memory don't like volt at all (like samsung chips) but Micron D9 loves it ^^
My D9GMH runs at 2.45v with 80mm fan for all days computing :) without any problem :/ I see better OC recently :/ certainly the winter xD
My infineon chips was running at 2.45v (mobo max) without any fan, but this ram sucks :slapass:
I don't know the max for all day computing of D9 chips :( but that's seem to be lower than 3.0v :D But better than 2.50v
Sorry for my bad english :)
Sincerely,
Just like TCCD/5 mostly topped out at about 2.7v and BH-5/6 around 3.7v for 24/7 use..
Some DDR2 will do 2.8v probably, rest will be closer to 2.1v..
2.5v is the new 3.5v :D
I do not see how 2.55v is that bad, sure your ram can die at this voltage, but if they have decent cooling I would think they would be fine.Quote:
Originally Posted by The Nemesis
I ran my fatbodies @ 2.67-8v to get 400 3-2-2-x ;)
My D9GMH has seen 400 @ 3-3-2 with that voltage.
I have always ran 2.4-2.45 for 24-7, so benching with more just seemed natural. I would not recommend going over 2.45-2.55v for 24-7 use even though I have never heard of people killing D9DQT/D9DQW.
I don't know, i wouldn't push D9's nor Aeneons over 2.30V for 24/7. RAM is too expensive at the moment to afford yourself ending up with a faulty ram stick :D
You do know he works at OCZ right?Quote:
Originally Posted by Progerien
Fatty D9 is 90nm whereas the newer stuff is 80nm so like with cpu's wouldn't the voltage hurt the Newer ones more atleast a little bit? So fatties do take a bit more voltage? jwQuote:
Originally Posted by Tony
in theory, yes. in practice? who knowsQuote:
Originally Posted by Praxis1452
My D9GMH ram is set to 2.3V in BIOS and 2.25V with Ai Suite. Is that a safe voltage to run them at 24/7 with a fan 20cm away from it?
I plan on keeping them for 2 years.
For 24/7 on D9GMH, D9DCD I used 2.3V-2.45V not more ( with 80mm fan on stick's )
For bench 2.7V is good for D9GMH ( Ballistix set )
Also I set 3V with this Ballistix and ram is live, but P5B Delux not :(
Regards
Martin
My 8000ul (fatties) ran 5 months without problems @ 2.65V and no active cooling... (beginn q1 06 - mid q2 06)
After that extreme 24/7 voltage I used only 2.35V, because it was enough for good frequency and latencies...
The only thing to worry is that this increable ram don't run good with p5b del -> only with spd timings... that sucks...
With p5wd2 I didn't have a single problem!
may be told this: once ran one of mine 8000UL (FattyD9) through 3.15V (was testing Vmem mod) without any cooling at all - the mem survived, but now OC degraded @ low volts by 20-30MHz :(
Thanks for your input =DQuote:
Originally Posted by DEVIL K-ce
I thought this was a load of crap at first but apparently not. My 2*512 PC5300 ballistix d9gmh ran 2.5v 550+ 4-4-3-5 benching for a while (30-40 hours?) on AM2 then spent about 10 hours @ 2.8v benching 3-3-3-5. Been running them 24/7 at 2.3v for a month with no fan, but they stay cold to the touch.
Last night I BSOD'd and set back to 2.5ghz figuring CPU was having an overheating fit again. Then 2.5 started BSODing so I shut down, threw a fan over the rams, and all was ok for an hour until I went to bed. Came home to the dreaded reboot loop and now even 400mhz wouldn't boot (was running 500mhz before) with stock CPU. Swapped the ram into the black slots on my C51XEM2AA and put voltage at 2.325v and it was stable for about 2 hours of gaming before BSODing. Running 2.35v now.
Unless my CPUs IMC or my foxconn board just crapped out randomly, I'm pretty sure running high voltages for any period of time will degrade d9 eventually :(
Those of you who have their RAM "failing" like this, suddenly running only low clocks - put it on a shelf for a couple weeks.
I had components recover from this condition. Most notably, after that Athene 850W PSU fiasco my magic CABNE had lost 200 MHz overclock. When I wanted to throw it out some time later I re-tested and all was fine, strong as before.
Tony,
Just wondering how effective some active cooling like those twin 60mm fan OCZ XTC's would be for lowering 24/7 temps.
For example if I had memory I normally ran 2.2 24/7 would those coolers let me go 2.3 or 2.4 with the same longevity?
Its not temperature that kills them. It's electron migration.
Still, nobody has answered the major question people have here.
How far can you go 24/7 with an active cooling system on it? Be it a 60MM fan, Corsair Dominator airflow system, water cooling etc.
Personally I've been running 2.4 for a while on my Crucial 10th Anniversary with the Corsair airflow system and no problems.
So Hipro provides added stability w/o even racking up the voltage?Quote:
Originally Posted by Revv23
Dredd - 2.3v ish.
Speederleander - cleaner voltage supply, so yes.
Hi everyone :)
IS THIS POSSIBLE to run D9 GMH (Ballistix) 3,75v for benching up to 4,0v (for few hours)? We have in Poland one "pistol" who claims that it he is running that voltage on vmoded Biostar Tforce 965PT (but he didn't notice any boost of MHz/timings). He claims that on 4,0v computer don't boot...
We are waiting for pictures :D
I think for 100% that its FAKE
3.75V? :eek: Fried memory anyone? :p: they will be crip and long deadQuote:
Originally Posted by Miravo
been posts of people running 3.3v+ hasnt there? Anyways, dead soon.
I wouldn't be surprised if they just died on boot up at that voltage.
Search function would remove your surprise.
There are people running there ddr2 @ 3.5-3.8V for short benches...not every ram dies immidiately.... ;)Quote:
Originally Posted by STEvil
Quote:
Originally Posted by Speederlander
Yes, part of it is the cleaner power, and part of it is because of the more stable power, with a drr maxi there is barely any vdroop at all when the dimms are put under load.
On stock Vdimm circuit, it is a much weaker design, made to run ok at stock but not at higher volts. So when the dimms get under load the voltage drops, meaning you have to raise the voltage even more to compensate. This voltage fluctuation isnt very good for your ram either IMO.
Is 2.55v too much for Corsair 8000ULs @ 400mhz & 3-2-2-4-6 - benchmark stable ?
Just ran some benchies at this level with active cooling on the sticks and all seems fine, but I don't want to be standing too much on the "edge" here...
Will post screenies very soon.
maybe seconds...Quote:
Originally Posted by Miravo
I highly doubt hours yet alone minutes.
probably only BH5 can survive those volts.
2.55 should be fine for benching.
i would not run 24/7 over 2.5 though.
Quote:
Originally Posted by afireinside
My pc have the same symptom as above, my ram is patriot 6400llk d9gmh run it at 2.47 volt with active cool when it happened. My mainboard is asus P5b Deluxe. sometime right after BSOD, i can't even get the machine to the boot
to xp boot screen.
:shrug: I've been running Team Group PC2 5300 for 5 months now 24/7 at 2.4v. Only around 373MHz, but higher voltages make me feel warm inside. And speeds haven't degraded a bit, I can bring them right up 100MHz higher with 2.65v. I think the key is good airflow. Especially when benching, people tend to forget. I know I did. I was very surprised by how hot my ram got even after a few minutes of use with no fan at 2.4v. After its first hour or so, my ram has always had a fan directly above it, regardless.
From reading this thread ive decided ill be running 2.3v-2.4v 24/7 use and benching at around 2.6v. This will be on Gskill HZ. This is a safe voltage from what i have gathered, :)
Quote:
Originally Posted by Miravo
Firstly, not ballistix!
Where I written that working 4.0 V???? :stick: In bios only 3.95V ~10 seconds, but for windows max 3.75V
G.Skill HZ 6400 (SN 605)
It is possible on 3,75V but only 5 Minutes :D
Now memory soils taste :(
I keep my GSkill for everyday use at 2.25v but for benching I will go 2.55. I was going to go crank them up to 3.7 for some benching but after reading this thread I decided that my memory costs too much to kill just to lower my SuperPi32 numbers.
Yeah obviously there was some accelerated degradation as Tony puts it.Quote:
Originally Posted by TaPaKaH
So its current (electron migration) running through the sticks that kills them, but how much effect does active cooling have on slowing the degradation?
I would think the cooling effect would be similar to the effect it has on CPUs. Maybe not completely eliminating extra wear but still effective.
P.S. I like those new Thermalright HR-07 Ram sinks.
Is this the DDR2 Maximizer that will allow you to run lower voltages through the memory?
Someone said he went from 2.2v to 1.85V with same settings. Only difference was the maximizer. If that is true, this is a worthwhile investment.
I have Crucial DDR2-1000 that I am running at DDR2-800 3-3-3-6 @ 2.2V on my P5W (P5W won't do much higher than DDR2-950). The stocks volts are 2.2V. I intend to get the memory up to about 500Mhz with my P5B and try to run it at 4-4-4-8. I will clearly need to run more than stock volts to do this. But with that DDR2 maximizer allow me to do this at lower voltages?
Cooling will only help so much and will only be helpfull below certain temperature ranges.Quote:
Originally Posted by Phosphate
Even if you have 20c loaded temps there could be one spot which is weaker than the rest.
DDR Maximizer doesn't do miracles......YES you can work your rams with lower voltage than the motherboards onboard Vdimm circuit but it also depends on the motherboard itself on how stable is it's circuit for this and how "clean" are it's Vdimm voltages......(not quite good I could say for at least most of them).Quote:
Originally Posted by pauldovi
DDR Maximizer has many advantages over nowadays motherboard's DDR circuits but it does not making miracles..... :)
Yes you can coldboot up higher (or even much higher) with a Maxi coz when we power up, most of the mobos firstly feed the DDR Modules with 1.8Vdimm and within a sec, it goes up to the point we have picked within bios. In such situation, our rams won't boot, therfore out all system won't boot.
Yes you can feed your rams ON-THE-FLY within Windows with the Maxi.
Yes it's dumn stable.
Yes it has TOO LOW Voltage output ripple comparing to ANY other motherboard's onboard design.
Yes it helps motherboard NOT to get hotter around the Ram's area (from it's onboard Vdimm power Mos-Fets) therefore the DDR Modules don't "draw" some of this temperature and they work colder
Yes the motherboard runs a bit colder too.
Yes it helps NOT to burn down (under some strange circumstances) during boot up with very high onboard Vdimm you may have piched, your DDR Modules from "spices" of this onboard Vdimm.
Yes it's easy to use and understand how it works within half an hour of playing with it.
BUT it does not making miracles.:)
Why doesn't ASUS make a high end motherboard that doesn't have fancy lights and cool stickers on it. Instead, they could make a high end motherboard that supply clean power and small voltage droops.... :rolleyes:Quote:
Originally Posted by hipro5
I am a kid, short on money like all the others. I do have an appreciation for electrical engineering (I have a family history of them). I think I will pick one of these up when I get the money. $190 isn't a bad price for your $300+ investment.
yes, exactly pauldovi.
i swear, a asus striker would be totally worth the money if they lost the vdroop. loose the heatsink rollercoaster and the voltage droop (put digital pwm).
Hey, guys. Please dont do like: "Hey everyone my stiks working at 2.6v for 24/7 and not die yet" It is bad. Please write your chips (like GMH, DCD, GKS...), cooling, voltages for 24/7 and benching voltages, degradation your memory(if it was). We need more FULL statistic to make some decisions about degradation. We need to make like: this chips like 2.xV for 24/7 and <2.xV for benching... M`ok?
Quote:
Originally Posted by Drager2
modules: G.Skill 6400HZ ( D9GMH )
cooling: Corsair Dominator fan kit.
24/7 voltage: 2.4V ( max. voltage on P5WDH, no mod. )
current status: D E A D.
Ran at DDR 800 4-4-3-5 with 2.1V from Aug. to Dec. w/o problem whatsoever.
Decided to tighten up the timing ( 3-3-3-6) during X'mas with 2.4V eventhough I hardly noticed any performance gain . About 10 days later, windows kept crapping out constantly until I lowered the speed to DDR320 @ 2.4V but it was short-lived anyway.
Just picked up a new set of Mushkin last Friday but I think I'm gonna keep it running @ cas 4 with 2.2V for 24/7 usage from now on.
Noticed a couple things.
Reports of the Maximizer helping out even while not using it for big voltage boosts...
It may help by cleaning up the Vtt and Vdimm droops (and SPIKES) that alot of these motherboards seem to suffer from.
Remember the DFI NF4SLI-DR boards, with a huge droop/spike on the Vtt signal with BH5? It severely limited the max OC, but at least on BH5 it didnt kill anything. Same Riktek 9715 voltage regulator is used on alot of the boards for the Vdimm/Vtt. My new Foxconn 590SLI board has it, I was quite disappointed when I saw it...ugh.
Now maybe with DDR2 and lower voltages it is ok, but maube not. I wish someone with a digital storage scope could do some captures on the popular motherboards to see if the droop/spike is present.
Tony? You guys have the equipment yes/no?
Im using a set of OCZ 1066SLI sticks and they are awesome, and I set it to 2.1V with active cooling blowing on them. They were doing great at 570 5,5,5,15 with a 5600+ Windsor (12x270Mhz for 3250Mhz). But today it got hotter and suddenly I had a no boot/hang in memory test in the bios. Pulled one stick out, booted fine, later put it in at 230Mhz 2.0V and it is running, but now I skeeeered of the voltage.
PS: Tony, any suggestions for the sub-memory timings with AM2 chips, everything out here now is Intel related.
Quote:
Originally Posted by Dumo
totally agree.i wonder how many micron d9 chips haved been rma,ed since there original ddr800 release date,lolololol:rofl:
voltage mainiacs = rma:eek: