New Kingston Hyper-X 240GB SSD died after 5 days

[mrjoltcola]

Just thought I'd post a note here for information purposes.

For my new Sandy Bridge build, I included a Kingston 240GB HyperX from Microcenter. It ran great for 5 days on SATA 6Gb port with Windows 7 Ultimate. On the fifth morning I sat down at work to a black screen saying "Drive read error". Drive shows in BIOS but wouldn't boot.

After a while the drive mysteriously came back, and I booted once. Within an hour and a couple of more reboots, it stopped booting for good. Read error, nothing. No matter how many retries, no luck.

Kingston worked with me on the phone, and were very helpful and friendly, but I'm not happy! RMAed it Friday, waiting for replacement now. Side note, advanced replacement is charged at the full MSRP of 720 bucks or whatever, so I opted out of that!

Restored backup to a Velociraptor until I get replacement.

My other rig has an Intel X25-M 160GB SSD with Win7+Oracle+SQL Server, no problems for over a year.

The performance on this Kingston was "off the meter", and I had waited for this new Hyper-X after reading Anand's review, and because I am somewhat conservative and Kingston is on my list of conservative companies. I've had good experience with their products; I do outsourced software development and can't afford to cut corners. I've had my share of DOA platter drives. In the past 3 years I had a DOA Western Digital 2TB Green and a 300GB Velociraptor. But the replacements that run good past a week have never failed. My failures have always been DOA, never do I have drives die days later. The problems with SSD aren't mechanical (unless they were banged around in shipment) so I'm pondering whether to revert to Intel X25-M if I can find one available, for their 0.5% RMA rate. Intel and Kingston were the two lowest failure rates based on the French company study of the unnamed online retailer (also noted on Anandtech I think), but that was pre-Hyper-X SSD.

Motherboard is an ASRock Extreme4 Z68.

I hope this is just a fluke. Anyone else with this drive have any comments?

 

[amanoai]

I don't have a Kingston but did you check if there was a firmware upgrade prior to installing the ssd?

 

[mrjoltcola]

Yes sir. Tech support said I had the latest firmware.

 

[VirtualLarry]

That drive uses a SandForce 2xxx controller, doesn't it?

 

[mrjoltcola]

Yes. The SF-2281

 

[beginner99]

I would avoid Sandforce at all cost. They show the best performance figures but reliability? No. Just look around here how many people have issues with sandforce drives.

for reliability -> Go intel.

EDIT:
Note that Kingston also had re-branded intel ssds which explains their better reliability numbers...

 

[mrjoltcola]

I wish I had known this when I spent my 500 bucks. I only use Intel, Western Digital and Kingston if I can help it and didn't realize there were known reliability issues with Sandforce.

Can you give me some references for this?

Also isn't Intel due to release a new SSD this year? I read an article that said they would release the "fastest SSD" on the market but don't remember where I read it or whether my memory is accurate even.

 

[scajjr2]

I have the 120Gb HyperX and have had 1 BSOD and one freeze up of Win 7 in the 5 weeks I've owned it. Kingston told me after the freeze up 3 weeks ago that they hadn't had any reports of issues. I'm betting now that their drives have been out for a month or so that they start having the same amount of issues other SF2281 drive mfgrs have.

Visit the OCZ, Patriot, any SF2281 maker's support forums and you'll see it's across all mfgrs.

Patriot is supposed to be releasing a new firmware for their drives today.

Sam

 

[amanoai]

I have Crucials, Intel, and Corsair. Though I have no problems with any of the SSDs, I would recommend Intel and Crucial, before the Corsair. Here's the way I see it, majority of users will not have problems with their SSDs. The users that do, more often than not, it is a SSD with a Sandforce controller. Also, as more users buy SSDs due to lower prices and word of mouth, we are getting more of the less knowledgeable crowd that does not keep their system, firmware, programs, and drivers up to date.

For the Intel's new SSDs, I think you are referring to Cheryville.

I wish I had known this when I spent my 500 bucks. I only use Intel, Western Digital and Kingston if I can help it and didn't realize there were known reliability issues with Sandforce.

Can you give me some references for this?

Also isn't Intel due to release a new SSD this year? I read an article that said they would release the "fastest SSD" on the market but don't remember where I read it or whether my memory is accurate even.

 

[mrSHEiK124]

Don't get an X25-M - if you really want one just buy the Intel 320 it's the same thing.

Intel 320 = SATA II, but win.
Intel 510 = SATA III, win, but expensive. Potentially less reliable than the 320, but probably not.
Crucial M4 = Intel 510 with a cheaper price tag. Lots of win here.

 

[mrjoltcola]

I guess I'm stuck with the Kingston unless the replacement dies too, then I'll pursue a refund. Do you know if the manufacturer does refunds, or would I have to return it to Microcenter (assuming it failed a second time).

The tough part is feeling like I've got imminent doom waiting on me even if it is running stable, it could fail at anytime; it'll keep me on my toes and ensure I keep my backups going, but failures seem to happen at the worst times, like in the middle of a hot project.

 

[mrjoltcola]

Can anyone quantify: What exactly are the Sandforce issues? Are they controller hardware faults due to component / fabrication quality, or are they microcode bugs that are fixable by a flash?

 

[mrSHEiK124]

Can anyone quantify: What exactly are the Sandforce issues? Are they controller hardware faults due to component / fabrication quality, or are they microcode bugs that are fixable by a flash?

Seemingly unresolvable firmware issues with the controller itself. The thing spazzes out and locks the drive down if you look at it the wrong way.

 

[tweakboy]

I think a firmware might have prevented that sudden death. But lets get real, Kingston isn't a top notch company. Next time buy a Intel or Sand Force. So whats the deal you going to get a replacement on it ? You were saying about some thing you opted out of... You payed 720 dollars for this ? Kingston should RMA and youll be set,, next time you wont buy a Kingston,,,

 

[mrjoltcola]

I paid 499 for it. They wanted MSRP (720) for the advance replacement credit card hold (shipping replacement before receiving mine back). I said scratch that.

Not sure what you mean by "Sand Force", that is what I bought, SandForce 2281 controller.

In my experience Kingston RAM has always been solid for me, I've bought tons of it in the past, and always viewed them as a conservative company. The word on the street with the HyperX SSD was that Kingston waited a while to release their sandforce SSD to work out the kinks, while other vendors were rushing theirs to market (OCZ problems with the Vertex)

Next time I _will_ stick with Intel, that is for sure. I'm considering moving my X25-M from my old rig right now, however, I don't want to get all scared up without giving the drive a chance. I keep good backups.

 

[scajjr2]

In my experience Kingston RAM has always been solid for me, I've bought tons of it in the past, and always viewed them as a conservative company. The word on the street with the HyperX SSD was that Kingston waited a while to release their sandforce SSD to work out the kinks, while other vendors were rushing theirs to market (OCZ problems with the Vertex)

That was why I went with the HyperX because I thought Kingston's late entry possibly meant they weren't going to have the same SF2281 issues. My other SSD is an OCZ Agility 2 120Gb, have had for 6 months with not 1 issue but when I was researching SF228x\SATA III SSDs, their forums were full of users with problems so I waited.

From what I read the main issue involves the drives having a problem when in sleep or hibernate or as in my case, just having the computer sit idle for a while. The drive "drops out" and isn't seen by the OS or by the BIOS if Windows restarts after a BSOD or if you hit the reset button on your computer. Only a power off/power on cycle of the computer brings the drive "back" in the BIOS\OS.

I would hope this is fixable with a firmware fix. Patriot pushed their firmware update back a day to do more tests so maybe the final fix hasn't been found yet. OCZ is testing a new firmware too. Hard to find any info on Kingston. Even their own website doesn't even show they sell the HyperX SSDs, only thing I found was the press release from a few months ago.

Sam

 

[nanaki333]

Don't get an X25-M - if you really want one just buy the Intel 320 it's the same thing.

Intel 320 = SATA II, but win.
Intel 510 = SATA III, win, but expensive. Potentially less reliable than the 320, but probably not.
Crucial M4 = Intel 510 with a cheaper price tag. Lots of win here.

there's a pretty significant performance different between the intel 510 and m4. intel also uses 32nm vs 25nm in the m4.

 

[mrjoltcola]

Has anyone started a "If you have a Sandforce 2xxx SSD post here" thread to track failures, models, and scenarios?

 

[nanaki333]

Has anyone started a "If you have a Sandforce 2xxx SSD post here" thread to track failures, models, and scenarios?

that would be a looooooooooong thread

 

[Coup27]

http://forums.anandtech.com/showthread.php?t=2187601&highlight=official

This is one of many many threads on the subject of the SF 2281 controller.

Go over to the OCZ support forums for Vertex 3 and have a look around. OCZ has the most active forum but the problems are by no means limited to just them.

In short, there is a widespread and serious fault with the SF 2281 which after months, SF still haven't even gotten to the cause of it, let alone a fix.

For absolute reliability, X25-M G2 is probably still king of the pile. the 320 should replace this but it has itself only just recovered from a show stopping firmware bug.

M4 seems quite reliable, although I personally dislike its all out approach for the fastest sequential speed regardless of random performance or how well the controller can manage its speed over time.

 

[groberts101]

funny thing is about all the assumptions being made about this controller and that it's ALL on Sandforce's shoulders is that Intel themselve's practically invented the partial slumber spec.. but then wen't ahead and disabled it on their drives(as did all the other controllers.. or at least the few that I've seen the data from.

Then there's the fact that not every system is completely stable to begin with. Saying crap like.. "well.. I don't have problems with my HDD on the same system" doesn't really mean squat when it comes to how much added stress a regular sata2 SSD adds.. not to mention one that can almost completely saturate the sata3 spec.

A quick look around at the "P67/Z68 freezing and bsod" issues will tell you that some system are seeing issues and they don't even have ANY SSD's installed. People who just love to rant/hate fail to dig for facts and are far too quick to really look to see who else may be involved in this fiasco.

OCZ is testing beta-firmware right now that will disable the partial slumber modes(such as the other controllers have done) even though the friggin' Intel drivers themselves have it inherently designed right into Windows itself. Think that it's possible to screw the pooch when the rug has been pulled from under you and you are designing firmware for a specific function.. only to find out that it has not been implemented as you were originally told? Think that could also screw the ACPI tabling between the bios.. OS.. and SSD controller? Ummm?.. yeah.. it can.

Then there's the whole flashing firmware while some mobo's are setting the ports to external configurations problems but to most that will just equate to another Sandforce controller's failure to make sense of. Which is actually probably more correct than most of the other assumptions being made in this and other threads about this subject.

Then that leaves the question of.. will Intel get all their issues and partial slumber specs implemented when the "sure to use Sandforce" Cherryville model comes out? Who want's to take bet's that they won't release it until it is? lol

Then again.. maybe the newest firmware that's soon to be released by OCZ(and surely others as well) that has killed the slumber mode availability will actually cure every known freeze and bsod known to man, eh?

and as a last little dig at Intel here.. ever wonder why Intel is axing so many of the sata3 ports of the X79(down to 4 from the original 10)? Intel is not infallible and those that put them on the highest pedestal(SSD or otherwise) will probably blame something else as the cause or bury their heads in the sand until issues pass. This is just typical hardware and software growning pains and assuming that even a Zillion dollar company can't possibly have issues is just being naive.

 

[mrjoltcola]

@groberts101: Yes, I'm all for factual info, which is why I suggested a SF2xxxx thread. I'd like to see full system config for people with failures, (overclock, motherboard, etc.). I do recall my problem was in the morning, but I do have all my sleep modes disabled, so I assume the slumber mode you refer to is all self-contained in the controller?

Anyway, I just got the replacement, about to install it.

 

[mrjoltcola]

The replacement drive is a no go so far. Windows won't install to it!

Same motherboard, same cables.

Windows 7 Ultimate w/SP1.

Partition shows as 223.6GB

Attempt to install to it says "Setup was unable to create a new system partition or locate an existing system partition."

I can go to Advanced and Delete/Create/Format, but cannot install to it.

 

[mrjoltcola]

I'm calling Kingston. I may move forward on the refund now and get an Intel.

So what is the Intel X25-M G2? Is this the original X25-M, or if not, how do I differentiate it from the X25-M I already have.

 

[mrjoltcola]

Sorry to spam post, could it possibly be the motherboard?

I also installed to a Velociraptor 300GB in the same rig, as well as the first Kingston same model.

As I said, I have an ASRock Extreme4 3.0 with latest firmware update.