So many A64 deaths

Quote:

---

Originally posted by Zeus
Was Vdimm set to auto in bios while you turned the pot?

I really want to do the Vdimm mod but all these dead CPU's scare me. :(

---

Yes, after reading this thread and Mickey Mouse suggestion, I didn't change the Vdimm value from bios... was set to auto before I started turning the pot.... and I was turning it real slow.... 3.0, 3.05, 3.1 and suddenly, both the multimeter and the bios screen showed 3.66....just for a sec or two, without touching the pot, it came back to 3.1.... The psu is a pc-power&cooling 475, with modded lines which are solid no matter the load....

Quote:

---

Originally posted by RootX
Yes, after reading this thread and Mickey Mouse suggestion, I didn't change the Vdimm value from bios... was set to auto before I started turning the pot.... and I was turning it real slow.... 3.0, 3.05, 3.1 and suddenly, both the multimeter and the bios screen showed 3.66....just for a sec or two, without touching the pot, it came back to 3.1.... The psu is a pc-power&cooling 475, with modded lines which are solid no matter the load....

---

Scary stuff mate! :eek:

Looks like anything over 3.5V can kill the memcontroller, so even the "mickey mouse method" isn't safe....?

Strange how some ppl can run over 3.5VDimm for over months and how some CPU 's get killed instantly by spikes.

Could it be a combination of high Vdimm/high HTT?

I don't mind doing the Vmod but i don't want to take out a €450,- chip for the sake of 0.2V more Vdimm.

Man, i just wish there was a safe way of Vmodding this board. :(

Quote:

---

Originally posted by Zeus
Scary stuff mate! :eek:

Looks like anything over 3.5V can kill the memcontroller, so even the "mickey mouse method" isn't safe....?

---

Well, this is what happened to me, if others don't verify it then something went wrong to my setup and MM's method should be considered safe.

Quote:

---

Originally posted by Zeus
Could it be a combination of high Vdimm/high HTT?

---

The HTT was default when I upped the Vdimm... another strange thing is that the voltage don't seem to fluctuate at any other point, mem@load or not, only when adjusting the pot and/or change the setting from the bios....

Haven't clk'd my 3200+ much 227fsb max and about 2.9 - 3.0vdimm on a kv8 max3. CPU has also been on a k8nnxp but even less clk and volts. On the possitive side it's been running strong for 4months or so now.

This vdimm thing sounds like a surge from changed the vdimm via bios or pot, and oscillating before dropping back or up and down. Seems that the reg or mosfet is on the limit or maybe even the CPU mem controller is overheating or transistor junction breaking down and drawing high current/ voltage subsequently killing itself :confused: .
I like the idea of the onboard mem controller improving performance, but it does seem to have some drawbacks.

Just curious, has anybody tried overclocking an A64 board without doing any extreme volt mods and just using the voltages in the bios?

Before you throw things at me for asking that, I'm just curious from a testing standpoint if anybody has killed anything just using bios settings.

Thanks,
Kosmic

Not sure, but truth be told, you're asking the wrong forum since most everyone runs a vmodded vdimm. We can't hold a piece of hardware longer than 24 hrs without slapping the solder gun on it.

Lol yea, I figured that was the case, but I just wanted to see if anybody had tested it at stock to see if any deaths resulted. I personally have a bit of solder-phobia so if I decide to go with 754 I can report back in here. Might be a while though...

Meh I'm running my 4Th A64 right now un modded.

Stable @ 2.3Ghz 1:1 ram no effort, priming 8 Hrs so far no errors.

Will be voltmodding come early AM :D

Quote:

---

Originally posted by Kosmic
Lol yea, I figured that was the case, but I just wanted to see if anybody had tested it at stock to see if any deaths resulted. I personally have a bit of solder-phobia so if I decide to go with 754 I can report back in here. Might be a while though...

---

I would love to get the chance to try overclocking one of these. My first died during a bios flash.

I got my RMA today and it died within an hour. Changed the CAS rating of the mem from 2.5 to 2 and away she went . I'm getting sick of this. Both stock, both auto with no mods.

I'm using Corsair XMS3500. What are other folks using when they're dying?

Quote:

---

Originally posted by dod
I would love to get the chance to try overclocking one of these. My first died during a bios flash.

I got my RMA today and it died within an hour. Changed the CAS rating of the mem from 2.5 to 2 and away she went . I'm getting sick of this. Both stock, both auto with no mods.

I'm using Corsair XMS3500. What are other folks using when they're dying?

---

Try another motherboard, it might be the CPU'S aren't dieing but the motherboard playing up from a bad flash. Either way I'd be trying a new CPU/ MB combo.

Sorry to hear that dod, that really really sucks. I don't see how it is possible for the CAS rating to affect the chip. Sounds to me like you have a bad mobo that's doing the killing, but that's just a guess.

Quote:

---

Originally posted by RootX
Yes, after reading this thread and Mickey Mouse suggestion, I didn't change the Vdimm value from bios... was set to auto before I started turning the pot.... and I was turning it real slow.... 3.0, 3.05, 3.1 and suddenly, both the multimeter and the bios screen showed 3.66....just for a sec or two, without touching the pot, it came back to 3.1.... The psu is a pc-power&cooling 475, with modded lines which are solid no matter the load....

---

could it be that your VR is not a 'precision VR'?
i mean, normal pots are supposed to stick to the before set resistance, but they don't guarantee linear resistance while actually TURNING the VR!!!
Only precision VRs are supposed to act in a linear way while changing.
So - if you change the resistance on the VR under current you take the risk of wrong values for a short time.

just a thought...

This is what I'm running right now:

http://www.gonad.org/upload/images/240x10.jpg

LDT @ 2.5, Ram 1:1 slot 1&3 :up:

schroet, you could have linked to your 2nd post.
no offence.

Quote:

---

Originally posted by Goldlocke
schroet, you could have linked to your 2nd post.
no offence.

---

I'll do that from now on, thanks for the heads up :toast:

Any idea if this affects Socket 940 Opterons and Athlon FX'es?

I'm going to buy:
Opteron 240
ASUS SK8V
512MB PC2700 Regd'ed ECC.

I'd prefer Opteron so I can move back into Dual heaven later on, and having a server processor makes you feel so much more badarse :)


Last thing you want is your machine crashing at 12 midnight in a middle of a game

Quote:

---

Originally posted by Goldlocke
could it be that your VR is not a 'precision VR'?
i mean, normal pots are supposed to stick to the before set resistance, but they don't guarantee linear resistance while actually TURNING the VR!!!
Only precision VRs are supposed to act in a linear way while changing.
So - if you change the resistance on the VR under current you take the risk of wrong values for a short time.

just a thought...

---

U're probably right mate, this could be an issue but I've been using the same pots for alot of mods, including my 9800's ones and something like this has never happened before...
Better safe than sorry though, so no more "live" tuning for Vmem, :)

Never encountered any problems whilst raising the voltage, i just turned the VR's very very slow and the voltage came up nice without ever spiking or anything.

I used 2x 1K 15turn pots and had to give them a lot of turns to start off but then the increase in voltage happend faster and faster.

So once it goes over 3V turn it very gently, it will raise fast.

Quote:

---

Originally posted by Jeff

Sorry for being on the rag today but I know of at least two other threads that would have benefitted from some A64 "brainstorming" where basically nobody replied. If people are affraid of sounding like an idiot, get over it... I'd rather hear from an idiot than from nobody at all(or someone after the fact saying "yeah, I had that happen to me a few weeks ago"). ;)

---

Hey Jeff I didn't have that happen to me last week ( heheh)...
I dunno why it's dying bro, but I would hope someone that had it happen would chime in as Randy and Bowman have......

Good to see you guys.

m

Anyone killed an A64 specifically by raising the 3.3v rail too high? My vdimm is chillin at 3.38v right now, but I'm still having issues getting my memory bus up very high, i have to run a 4:3 divider or higher, my HTT bus doesnt have any problem going up now, but still not much luck on the mem bus. Only getting about 210mhz at 3.38v 2-2-2-5 and this is bh-5?? what gives?

Try and take 1 block of memory out and only use 1 stick in slot1.

Ya i already know that that allows me a higher mem bus now...but I'd rather take a hit on my mem speed then have to go down to only 512megs...Planetside would die...

just wanted u guys to know, the 3000+ CG stepping boxed is available by now

edit: its already a newcastle

Quote:

---

Originally posted by BMORIN
Sounds to me like you have a bad mobo that's doing the killing, but that's just a guess.

---

You were right. RMA finally got approved yesterday. Don't know what the fault was though.

Well, glad to see that you isolated it. A mobo killing chips is not an easy thing to isolate.