DDR3 damaged

I recently removed a matched pair of OCZ DDR3 4Gb modules from my system, and sold them to a friend. It was shipped safely in it's original packaging, with oodles of padding, and arrived without signs of damage.

They're 1,066Mhz RAM modules, and all four strips reported exactly as shown below in the screenshot before shipping them off. I have anti-static rubber matting for workshop bench tops and the floor area, and use a ground connection on the chassis whenever working on internal components.

These units run happily at 1.333Mhz on a Gigabyte board in fact, and have been stable for over two years at that frequency, with no system freezes, data corruption, or BSOD's.



Safely received by my friend, her teenage son then installed them to a base level Asus mainboard in which they are specifically listed as compatible. But he did so without any precautions against ESD shocks and did not upgrade his board BIOS before fitting them (his BIOS was something like three revisions old as compared to the most recent edition).

His system would not boot with both sticks installed. He's running W7 x86 Ultimate Edition.

With one stick installed it would boot. However the timings and frequency the board, and I'm assured at default BIOS values, showed were not as per the screenshot above.

I got his mother to test the RAM in another board, and the module which would not boot on her sons Asus board did boot on her ASRock X48 Turbo board (although she did not run memory diagnostics or test it thoroughly).

Advising they return it to me I replaced it in the Gigabyte board it had come out of and ran it at 1,066Mhz and standard timings, but got several freezes and then a handful of BSOD's in the space of a few hours.

The dump files show the errors were due to the memory.

I removed the offending stick. Everything ran perfectly, all timings were identical and as per OCZ specifications.

I reinserted the offending module and ran Windows memory diagnostics at the harshest level for four cycles, taking some five plus hours. Multiple errors arose.

The three decent sticks were removed. I then manually set the suspect RAM timings to OCZ defaults in the BIOS. The board would not boot.

Allowing the BIOS to automatically set and optimize the RAM settings with solely the suspect module in place resulted in a successful boot, and I got the second screenshot below. You will notice the timings are askew and the RAM frequency is off key. The system freezes and is unstable.



Could anyone speculate what has happened to the module behaving as per these readings?

I believe that her son may have tried running it on an Asus BIOS which was written without compatibility for these OCZ modules, and that he could have over-volted the RAM, or that static electricity damaged it when it was fitted.

OCZ no longer make these units. If I were to RMA anything it must now be returned to a service centre in Europe, at my expense. The warranty only covers defects in manufacturing. Does anyone know what OCZ are likely to say, or whether they could determine why the RAM suddenly failed when fitted to a different board?

If RMA'd is it likely that OCZ reject it under warranty? Has anyone had any experience of the standards of customer service OCZ provide?

Any advice appreciated. For the moment everything's running stable with 6Gb of RAM onboard.

have you tryed setting the timming and voltage yourself ? also is there any reason you used slot 3 and not 1?
try setting the timming to ( 9.0 - 9 - 9 - 20 - 27 ) and voltage (1.50) and see what happens. if that don't do it try bumping up the voltage just a little bit and try it again...

scrooge

A problem that many folks make is not to kill the power going to the board by pulling out the AC cord (or flipping the switch off on the back of the P/S if it has one) and then hitting the case power button to drain any remaining charge off the capacitors. MB, even when turned off still have power going through them and removing/installing memory without draining this power can damage the memory.

scrooge said:

have you tryed setting the timming and voltage yourself ? also is there any reason you used slot 3 and not 1?
try setting the timming to ( 9.0 - 9 - 9 - 20 - 27 ) and voltage (1.50) and see what happens. if that don't do it try bumping up the voltage just a little bit and try it again...

I've not dared to increase the voltage, but I did force in the correct timings in the Gigabyte board BIOS. When I did this the board would not boot. This was attempted with the one RAM strip in situ, in slots one through to three before I gave up.

The RAM had previously just been working fine at the right timings and at a standard setting of 1.5v. It's possible to push it to 1.65v but I'm reluctant to do that because it could exacerbate any damage. I tried it in various slots, just so happened one of the screenshots was taken of it when it was in slot three.

mgp1964 said:

A problem that many folks make is not to kill the power going to the board by pulling out the AC cord (or flipping the switch off on the back of the P/S if it has one) and then hitting the case power button to drain any remaining charge off the capacitors. MB, even when turned off still have power going through them and removing/installing memory without draining this power can damage the memory.

I too suspect the power issue, at installation, is a significant factor.

There may be no way to definitively tell what happened to it; at least with the diagnostic tools at my disposal, and with the OCZ heatshields it's tough to tell visibly if anything had fried (and removing them to look could cause damage of course, as well as knocking any warranty in the head!)

Thank you both for your replies :)

Does anyone else have actual experience of OCZ's RMA policy in relation to memory?

badger906 said:

I have only ever updated my or a bios once in my many years of building and repairing computers, and that was only to allow my mobo to accept a core 2 quad cpu.
If it aint broke don't fix it!

I highly doubt the bios would be to blame, and to my knowledge no motherboard at default spec would have a ram voltage at such a different level to that of another board, within the same family of ram (ddr3) and voltage ranges are vast as it is.

Static electricity... I'm a sceptic. Yes when I was really young I had the anti static bracelet and made sure I was grounded to at least 2 sources. But no matter how hard I've tried I can't make components die! Even scuffing around with rubber shoes till I actually transfer a static bolt to components lol.

Compatibility issue gets my vote. Some things just don't go together!

--

Yeah I have, limited is an understatment! I had some of there ddr2 gold plated things back in the day, and had a couple go bad.
I got no where. It was basically my fault so I was told. But then I contacted ebuyer who I got them off and I had a carrier at my door the next day, it was collected and I was refunded :)

Thanks for the RMA input in particular. These strips also came from ebuyer originally oddly enough! It'd be worth asking if I can RMA them to ebuyer I guess, however the invoice was raised in late 2008, so I have some doubt about that now being acceptable to the store!

The budget Asus board these were going into had around six BIOS updates in as many months; and was a new build. It came as a package deal evidently, with a G620 (crippled i5) Intel CPU and basic 2Gb single Kingston RAM strip.

Apparently it was working fine, even though the BIOS loaded was some 2/3 revisions behind the latest.

I suspect the board may support a bus clock speed of 1,333Mhz. However the CPU does not support more than 1,066Mhz.

Unfortunately I wasn't present when it was fitted, but I do not believe the PSU capacitors were drained, and I've no idea whether the BIOS settings were correct.

I suggested a BIOS update after the buyer reported non-boot; in case the memory compatibility list was based on a particular level of board BIOS, and because there had been so many releases in a short time scale. Suffice to say this proved not to solve the problem.

On the ESD front I've known static electricity to kill both sound and ethernet chips on the edge of motherboards, and have sometimes suspected it might have played a role in damage to other components. In my experience it does seem to be linked to people, rather than hardware, simply because some folk do seem to have ESD issues more than Joe Average might!

As far as I'm concerned the strip cannot be trusted and what I actually get left from the situation is a single piece of 2Gb DDR3 which I could perhaps sell to someone with a mini HTPC style board.

I've asked the buyer to consider carefully what further costs could arise in sending it back to OCZ in Europe, versus the likelihood of an RMA claim being rejected.

Qdos said:

...
Does anyone else have actual experience of OCZ's RMA policy in relation to memory?

Yes, not good at all, unless you push very hard. Their hope & intent is that you'll get frustrated & go away. It took 2 months to get a check fo $5, after sending back the defective flash memory, product card & proof of purchase from approved vendor.

I will never buy another OCZ product again & recommend no one else does either. I will continue to give them bad reviews as long as they adhere to their policy of bad customer support & warranty. Their waranty marginally meets the letter of the law without meeting the intent of the law & they will continue to do what they do until we, the customers, no longer purchase their products & they finally go out of business & the executives disappear from the scene.

badger906 said:

...Thing is, RAM is so cheap these days its almost pointless to try and get repaired or RMA'd...

That's exactly what they are hoping you will do and that is your right, but not in your best future interest. I say, if the have a warranty, hold them to it. The reason being, if you don't, they have no incentive to produce a quality product. This is why all electronics is now junk, the customers don't hold the manufacturer's accountable.

I'm not trying to 'dis you in particular, just think about what I'm saying. I've been dealing with electronics (designed computers of all kinds) all my life and have seen a drastic decline in quality in recent years. It will only get worse until customers say, "no more".