Well, Crucial Ballistix has a dark secret...

The problem ONLY occurs with the 1GB Crucial Ballistixs PC4000 DIMMs. The 512MB modules have no problem.

Now, this MAY be coicidence, but every instance of these DIMMs failing that I have seen has been on DFI motherboards. Now, I don't know if that's a case of selection bias, or if there is perhaps more too that. It's just an observation I've made. The only hypothesis I can make about this is that DFI boards actually slightly overvolt their DIMM slots. Perhaps this is why DFI boards are such good clockers, they "appear" to be running at stock voltage when actually they aren't. Just something to think about, but I have nothing to back it up and I'm not claiming it's true, so don't bark at me. Does anyone know if there is anything that can actually test the voltage on a DIMM socket?

I have these DIMMs in my brothers machine, which is a P4 3.0C on an ASUS P4C800-E Deluxe, and they run at 250MHz with no problem. The system has been running at 3.75GHz no sweat since I got the DIMMs...must have been about 6-9 months ago.

I'm wondering if perhaps it was a bad batch of Micron chips that got out, and they are still circulating around. Mine are fine, as are a lot of others people's it seems that people who have problems, their chips run really hot to begin with. Mine don't. They actually run as cool as my older Corsair DIMMs. I guess "cool" isn't the right word. Perhaps they run as hot as my other Corsair DIMMs, but it's nothing out of the ordinary. They seem pretty normal to me.

Or speculating that DFI boards over-volt the mem modules?! :?

Quote:

---

Originally Posted by proxops-pete

Or speculating that DFI boards over-volt the mem modules?! :?

---

No. I suspect it's probably a bad batch of Micron chips. However, DFI boards are the best clockers. They do have problems with memory that seems to run naturally warm. They aren't as good clockers with chips based on Micron chips, which are less voltage tolerant, and James here says that at 2.6V the ram runs fine at 3-4-4-7 at 250MHz, which is still below spec timings even at a lower voltage. Sometimes I just wonder "why" it is that DFI boards seem to generally be the highest clockers. You put the same DIMM in two boards, and...well in the few cases I've done, the DFI board clocks higher. Why? Maybe I'll pick up an OCZ DDR booster and see if there is any substance to my crazy ideas.

Quote:

---

Originally Posted by ImaNihilist

No. I suspect it's probably a bad batch of Micron chips. However, DFI boards are the best clockers. They do have problems with memory that seems to run naturally warm. They aren't as good clockers with chips based on Micron chips, which are less voltage tolerant, and James here says that at 2.6V the ram runs fine at 3-4-4-7 at 250MHz, which is still below spec timings even at a lower voltage. Sometimes I just wonder "why" it is that DFI boards seem to generally be the highest clockers. You put the same DIMM in two boards, and...well in the few cases I've done, the DFI board clocks higher. Why? Maybe I'll pick up an OCZ DDR booster and see if there is any substance to my crazy ideas.

---

A lot of the DFI forums tell you the exact contact points on the motherboard to get true Vcore, vdimm voltages. If you were that interested you could just hook a multi-meter up to those points and see whats up. If I have some free-time maybe I'll try that. I just got a DFI board and I have a multi-meter and some experience using a multimeter.

Quote:

---

Originally Posted by OutKlast

A lot of the DFI forums tell you the exact contact points on the motherboard to get true Vcore, vdimm voltages. If you were that interested you could just hook a multi-meter up to those points and see whats up. If I have some free-time maybe I'll try that. I just got a DFI board and I have a multi-meter and some experience using a multimeter.

---

I'd be interested to know. I don't have a board on hand now, but I'll check out the forums. Thanks.

Also, does anyone know what JEDEC spec is? Is it 2.5V or 2.6V?

Technically, I think it's 2.55v, but we won't go there. ;)

I know for a fact that at least the DFI boards that I have dealt with slightly overvolt the vDIMM. Setting a vDIMM of 2.6 gives you a 2.63v in BIOS. Setting a vDIMM of 2.8 gives you 2.85v in BIOS. Even the BIOS is reporting that the vDIMM is higher than "set".

Oh, and the timings on the other set of Ballistix in my friends computer is 3-3-3-7, not 3-4-4-7. ;)

It does seem to be an issue solely with the 1GB PC4000 sticks from Crucial. However, I've seen 2 documented cases in Asus boards. :shrug:

What does your vDIMM read in the BIOS on your Asus, ImaNihilist?

Quote:

---

Originally Posted by James

Technically, I think it's 2.55v, but we won't go there. ;)

I know for a fact that at least the DFI boards that I have dealt with slightly overvolt the vDIMM. Setting a vDIMM of 2.6 gives you a 2.63v in BIOS. Setting a vDIMM of 2.8 gives you 2.85v in BIOS. Even the BIOS is reporting that the vDIMM is higher than "set".

Oh, and the timings on the other set of Ballistix in my friends computer is 3-3-3-7, not 3-4-4-7. ;)

It does seem to be an issue solely with the 1GB PC4000 sticks from Crucial. However, I've seen 2 documented cases in Asus boards. :shrug:

What does your vDIMM read in the BIOS on your Asus, ImaNihilist?

---

Never checked. I don't have access to the rig right now. Might be home on Sunday, I'll check.

i dropped the vdimm down to 2.7 and tightened up the timings a little. should have tried that in the first place. (a8n-sli premium)
3-3-3-7-10-14 250x10 1T

Very nice.

Quote:

---

Originally Posted by OutKlast

A lot of the DFI forums tell you the exact contact points on the motherboard to get true Vcore, vdimm voltages. If you were that interested you could just hook a multi-meter up to those points and see whats up. If I have some free-time maybe I'll try that. I just got a DFI board and I have a multi-meter and some experience using a multimeter.

---

I'm not too reliant on my multimeter readings right now. I've been putting it through quite a bit of stress. Its max setting is 600V and I've gone at least to 2000v.

Eric

Quote:

---

Originally Posted by ImaNihilist

I'd be interested to know. I don't have a board on hand now, but I'll check out the forums. Thanks.

Also, does anyone know what JEDEC spec is? Is it 2.5V or 2.6V?

---

I just checked the VDIMM and VCORE (PITA with a Big Typhoon in the way) on my DFI Lanparty UT NF4 Ultra. Voltages for Vdimm and vcore were both .01V below what Motherboard monitor (with a custom DFI profile) were reporting. The BIOS is reporting voltage's that are .02 higher for the Vcore and .03V higher then MBM and therefore .03, and .04 higher then true voltages respectively.

So in conclusion get MBM and the custom DFI profile and you'll have a pretty accurate picture of the various voltages. And it seems the DFI does OC the Vdimm .03v from your setting; but again with MBM you would see the true voltage with that variation.

It's not just dfi boards. You can read @ XS about new Asus boards eating dimms as well. Maybe it's on the sticks after all.

Are there OCZ sticks that use the same Micron chips?

One other question, that anyone with these sticks can answer, can you run 2-2-2-6 stable at DDR400 speeds?

Quote:

---

Originally Posted by OutKlast

I just checked the VDIMM and VCORE (PITA with a Big Typhoon in the way) on my DFI Lanparty UT NF4 Ultra. Voltages for Vdimm and vcore were both .01V below what Motherboard monitor (with a custom DFI profile) were reporting. The BIOS is reporting voltage's that are .02 higher for the Vcore and .03V higher then MBM and therefore .03, and .04 higher then true voltages respectively.

So in conclusion get MBM and the custom DFI profile and you'll have a pretty accurate picture of the various voltages. And it seems the DFI does OC the Vdimm .03v from your setting; but again with MBM you would see the true voltage with that variation.

---

I stopped using MBM after the NF3, because the developer stated they would no longer build in support for the new chipsets. Did someone else take the project over?

*Edit: Oh, and ImaNihilist, I'll check when I get them back from newegg. :)

Quote:

---

Originally Posted by James

I stopped using MBM after the NF3, because the developer stated they would no longer build in support for the new chipsets. Did someone else take the project over?

*Edit: Oh, and ImaNihilist, I'll check when I get them back from newegg. :)

---

Thats what I also heard but with the custom data profile (not available through MBM site) It correctly reads Temps and voltages the fan rpms occasionally go crazy.