high vdimm kill phenom?

[All4AMD]

can running high voltage to memory kill a phenom? the other day my system hung and would never post after that. so today i took my system to a buddys house and put his dual core athlon in and it boots up. put my phenom in his board and system won't post.

the memory i have is geil 1066 ddr2 @ 2.35-2.45v

[jimmyz]

I believe there was a thread awhile back from Tony about this very thing, I did a little looking and didn't see it. Don't know what they found out but I believe that several phenoms were dead after high vdimm.

[All4AMD]

i just did a search and it seems this is the problem as i set the ram to 2.35 so i could run 1066mhz

i also just saw that thread

[LIKMARK]

It were only the early rev b1/ba barcelonas this applied to iirc

[YukonTrooper]

Well at least you aren't an AMD fanatic or anything...

[LIKMARK]

Please explain further YT

[jimmyz]

Yeah YT explain why you need to Troll AMD threads. Don't like them then stay out.

If you can't help with the thread than butt out and keep your childish B.S. to yourself. The man lost a cpu and feels bad about it and you come in here flamebaiting, you are exactly whats wrong with this forum.

[informal]

Well at least you aren't an AMD fanatic or anything...

Well at least you actually posted something related to the topic title in AMD subforum...Oh,wait you didn't...You contributed zero to the problem/solution.
Why post if you have no intention of helping out??I am yet to see a post of yours that contributes something to the actual solution/problem in an AMD forum/thread..I guess i am an optimist for life. lol silly me




@All4AMD
edit:sorry i didn't catch u tried that in 1st post .Got distracted by flamebait post by YT :S:

I will search the forum for a possible solution in which you would need to freeze your Phenom in your fridge for 15 mins.Was KTE who tried it,IIRC. He left the forum unfortunately..

Found this thread.Seems many users got issues with C1 errors on a DFI mobo which killed a few phenoms while running Sisoft Sandra/Speedfan apps.Can you confirm you didn't try those apps yourself,just before system failed boot?

edit2:
Also found this post by our friend KTE:

Just to update this;

When a new phenomena arrives, it is difficult to actually understand or assess it because of the lack of experimentation and experience. As time builds, you develop a much clearer picture of the situation.

I noticed I had a non-booting Phenom like some of you were saying, what you would call "dead Phenom", but I didn't return it even after 3 days although I was about to, rather I experimented and repeated it. It was caused by just going past 626MHz on RAM. Voltage didn't matter whatsoever, I've had 5 Phenoms (7-43/7-44/7-45/7-47 etc) now and ran 1.8-2.9VDIMM on them varyingly since around 20th November '07 at low and high VCore and at low and high NB VIDs, no mem. degradation whatsoever nor did those affect it... but speed did. Phenom did 626 4-5-5-15 at 2.2V BIOS/2.18V real fully stable but failed to boot 627 even at up-to 7-8-8-22 2.9V. It did not matter what IMC/CPU volts were fed nor the CPU/RAM temp (up-to 10C idle tested).

The same RAM I swapped in a P35/C2Q setup and it ran 2.2V 676MHz 2.3V 5-5-5-5 SPi 32M with ease 4mins later.

Guess how it all happened?

When choosing high mem speeds on 1:2.66 divider, Phenom would get stuck in a reboot cycle even at low volts. No POST, no matter what you did even after 2 days. By checking the various voltages, I found CPU voltage had dropped to sub 1.19V at whatever setting you chose before that, it didn't matter and that caused a failed bootup -- didn't matter which CPU/NB/HT speeds/volts you ran previously. I usually test under cold and as you'll see documented in the 9500/MSI K9A2 Plat. thread, I had to take the CPU subzero (below -15C) and leave it at those temps for 30mins, then bootup with only one DIMM installed and it would always work fine and restart up thereafter. So "dead CPU" perfectly alive again.

If it were anyone else not trying sub-10C, or even me not trying it, it definitely was like a dead CPU and an RMA was needed.
KTE is offline Report Post Reply With Quote

This post shows what could be happening with SisoftSandra/speedfan running on DFI mobos:

My 9500 also died on the first day in that mobo. DFI support says tools like Sandra/Speedfan can cause the bios to be corrupted, which results in C1 errors.
I got the bios working again with an BE2400 and had no problems with an 9600BE and an 9850BE so far, but the 9500 did not yet post and i tried three different boards. As C1 errors are memory related i guess it's the IMC on the cpu whom stopped working.
I'd recommend bios versions 130/307 or 331 on the DFI M2A and be carefull with those low IO accesing tools like Sandra/Speefan/Evereset. Best is to run only one of those at a time.

So, basically, low level I/O tools corrupt the BIOS, and when it corrupts, it breaks the CPU in the process. I just tried my 5000 BE though, and the board shifted into 88 code. Back it goes, and I'm gonna try to get rid of the thing permanently.

[All4AMD]

Well at least you actually posted something related to the topic title in AMD subforum...Oh,wait you didn't...You contributed zero to the problem/solution.
Why post if you have no intention of helping out??I am yet to see a post of yours that contributes something to the actual solution/problem in an AMD forum/thread..I guess i am an optimist for life. lol silly me




@All4AMD
edit:sorry i didn't catch u tried that in 1st post .

yeah him and i swapped chips. his worked in my board and my chip didn't work in his. with my chip it was getting stuck on C1 which indicated memory.

[informal]

I added a few could-be-helpful quotes in my second edit.I hope you resolve your issue before you decide to RMA the chip(since it could actually be alive!)

[All4AMD]

I added a few could-be-helpful quotes in my second edit.I hope you resolve your issue before you decide to RMA the chip(since it could actually be alive!)

i can confirm it is a dead chip. it still doesn't work in my system or other boards.

i had just set the vdimm to 2.35 and ddr1066. started a game and it locked up after a few minutes.

[YukonTrooper]

Please explain further YT

Just a joke. His username is All4AMD and he has an AMD avatar, so I said "Well at least you aren't an AMD fanatic or anything...". Maybe it came off a bit brash, which I didn't mean it to. Just a joke.

Yeah YT explain why you need to Troll AMD threads. Don't like them then stay out.

If you can't help with the thread than butt out and keep your childish B.S. to yourself. The man lost a cpu and feels bad about it and you come in here flamebaiting, you are exactly whats wrong with this forum.

Sorry, just a joke. I wasn't flamebaiting, although it might have seemed like it, as my humour can be too blunt for some. And TBH, you make it sound worse than it probably is. I'm sure the man is OK and isn't in the corner in the fetal position.

@ All4AMD - Didn't mean it as flaming. BTW, I RMA'd my brothers Phenom because it stopped booting up, regardless of the RAM in use. I knew something wasn't right because he said it just stopped working. I swapped RAM (thought it was that for sure), PSU's, GPU, and nothing helped. Never really looked into the problem any further because I just thought it was a dud CPU. His replacement 9550 has been fine for awhile and he also now runs Corsair XMS2 DDR2-800, which takes less voltage than the D9 Mushkin he had in before. One of the Mushkin sticks started to get errors and my identical Mushkin D9 kit started dying recently as well. I sold his and will by selling mine when they get back from RMA, because I've lost trust in D9 chips. I purchased a pair of OCZ Plat Rev.2 as replacement sticks, which use Epilda chips and also use less voltage. However, I'm running an E6420, so I guess that's moot.

If I were you, I would just RMA the processor and perhaps use memory that takes less voltage, or decrease the voltage use on your chips. I never thought that my brother's Phenom died from too much vdimm, but by twist of fate it's probably a good thing those D9 chips started to die and he's now using different IC's with less volts. It could have been a totally different problem, but after reading this thread it sounds like it might have been the cause.

If you have to freeze a $250 chip to -10C to work, or you have to adjust all the settings in your BIOS just perfect (which will obviously handicap your OC ability), that's absolutely ridiculous. Hey, maybe AMD should ship ducktape with their processors! Us Canadians use that to fix everything.

If you think my joke was bad earlier in the thread, you should see how bad I give it to my brother.

[dball3]

now that's some good info there YT, thanx for contributing to the problem. I always think twice before i pump the voltage higher than 2.2. I always get a feeling that the IMC might freak out on me.

[JumpingJack]

@OP -- there is some underground info floating around that suggest too high Vdimm can kill a phenom, I have only seen reference to it in regards to B2's not B3's. Informal linked some info above, there is also another thread buried somewhere in this forum to suggest as much. Myself, I have not killed a Phenom yet with Vdimm, but I have not breached 2.1 V either. Lost Circuits also has reported killing Phenom's with too much Vdimm:
http://www.lostcircuits.com/cpu/amd_phenom/

One thing that is clear, though is that there is quite a bit of performance to be gained by running the memory at 1066 MHz frequency. However, there is a caveat in that AMD disallows any DDR2 supply voltages above 2.0V. We tried 2.2V and it lasted for about 3 days after which the CPU was dead. We tried some more and had more dead CPUs (courtesy of some of our anonymous contributors).

Perhaps B3 has solved some of these marginalities ... It is hard to say for certain, I have looked over AMD's specs, and could not find a specific electrical specification for Vdimm (but I did not look very hard either, it could be there).

jack

[YukonTrooper]

Myself, I have not killed a Phenom yet with Vdimm, but I have not breached 2.1 V either.

As a note, D9 chips run at 2.2V stock. I'm sure we all know this, but interesting to point out nonetheless.

[All4AMD]

hmmmm well i gues ill find out if this is a b3 problem or not as ill try a high vdimm with the replacement chip.

[jimmyz]

As a note, D9 chips run at 2.2V stock. I'm sure we all know this, but interesting to point out nonetheless.

By default d-9 chips conform to jedec stds which dictate voltage of 1.8v. some manufacturers agree to warranty their sticks up to 2.2 and above that so that they can market "faster ram". std DDR2 is 1.8V.

[Brother Esau]

Quite simple here guys OCZ Tony as well as RGone from D.F.I both state that High Vdimm to Phenom with kill the CPU. I don't know about you fellas but considering one source works for a motherboard manufacturer and the other for a Memory manufacturer I would say that considering they do this day in and day out as a profession and also Tony as Well as Robert are both highly respected people on the Internet and in the business I would be inclined to believe what they both say , which I do personally and have applied their wisdom and knowledge to my tweaking and tuning and have yet to have issue one with PHENOM Period!!!!. >>>>>>>.There is a Reason for that wouldn't you think? and it all has to do with listening to professionals that know more than I do and incorporating t into what I do>>>>>>>>>>>>THE PROOF IS IN THE PUDDING

[JumpingJack]

As a note, D9 chips run at 2.2V stock. I'm sure we all know this, but interesting to point out nonetheless.

This sorta makes it hard to use DDR2-1066 as many modules, not just D9 chips, require more the the JDEC standard voltage to reach those speeds.

In fact, I have been playing around with the Phenom for about 3 months now, both B2 and B3 and have not pushed to DDR2-1066 speeds that often for this reason, ultimately I will go there and stay ... see what happens.

[Oldguy932]

Here's a 2x1gb set that does 1066 at 1.8v.

http://www.newegg.com/product/produc...82E16820134655

Granted it's not speedy timings, but for $80 that's a steal. I'll probably be picking some up soon because of my memory situation.

[JumpingJack]

Here's a 2x1gb set that does 1066 at 1.8v.

http://www.newegg.com/product/produc...82E16820134655

Granted it's not speedy timings, but for $80 that's a steal. I'll probably be picking some up soon because of my memory situation.

http://www.valueram.com/datasheets/KVR1066D2N7K2_2G.pdf

Yeah, this would work however it;s 'value ram' at CL7 ... it loses some of it's benefit to go to DDR2-1066. OVerall, though the price as you mention is a good deal.

Just a quick scan, there are a quite a few choices in the 2.0-2.1 range with better timings.. this would probably be safe.

EDIT: http://www.newegg.com/Product/Produc...82E16820145033
This is what I am using in the 9850BE right now, and it runs great at 2.0 V ... no issues, MEMtest86 stable and passing.

[All4AMD]

im not gonna be able to get any more ram. i already invested in 8gb. being a college student and an airman is tough

i have 2 of these kits. http://www.newegg.com/Product/Produc...82E16820144224
i know newegg says 2.0V - 2.25V but on the sticks they say 2.35-2.45

[JumpingJack]

im not gonna be able to get any more ram. i already invested in 8gb. being a college student and an airman is tough

i have 2 of these kits. http://www.newegg.com/Product/Produc...82E16820144224
i know newegg says 2.0V - 2.25V but on the sticks they say 2.35-2.45

That is one thing I have learned -- never trust the newegg spec sheet on memory particularly, go straight to the vendor and download the PDF.

I have 8 gig of this http://www.newegg.com/Product/Produc...82E16820145197 and, in this case, the newegg spec matches actual.

Jack

[LIKMARK]

Just a joke. His username is All4AMD and he has an AMD avatar, so I said "Well at least you aren't an AMD fanatic or anything...". Maybe it came off a bit brash, which I didn't mean it to. Just a joke.

If you think my joke was bad earlier in the thread, you should see how bad I give it to my brother.

Hehe, that's okay, I suspected it was something like that, therefore I didn't swallow the "Bait" (unlike someone else here, obviously feeling the need to defend their rights ) JK

@ Brother E: I have read somewere ("respected", I'll look for link) that it was the high "delta"(not sure I'm using the right words here..), difference between low vcore and high vdimm that killed early models of Barcelona/Phenom. That should be BA/B1 models, IE prelaunch Phenoms.

[SocketMan]

High Vdimm to Phenom with kill the CPU.


They probably said "might" kill the CPU ? Brother

I've been feeding 2.2V to my 9600BE for about 4 month now, it's still alive.
I read that post of KTE's about freezing the cpu,a while back and the only place that comes to mind is one of these "industrial" freezers they have in Supermarkets. Ask some girl there if you can stick your unit in for an hour
Have the board ready at home. Bring it in frozen, and see what happens.