cancel
Showing results for 
Search instead for 
Did you mean: 

Uncorrectable Error Count - Intel 520 SSD

idata
Esteemed Contributor III

Hi,

I was surprised to find out how big the raw value of Uncorrectable Error Count for my drive is: 220147178. The rest of the attributes in SMART seem to have normal values. I've checked the value both in Windows 7 64-bit using Intel SSD Toolbox (3.0.2) and GNU/Linux using smartmontools. I've also noticed it grows quite rapidly when using the ssd.

According to the documentation Intel 520 SSD Product Spec:

The raw value shows the count of errors that could not be recovered using Error Correction Code (ECC). Full Diagnostic Scan in Intel SSD Toolbox passes without errors. Furthermore in linux I have successfully copied many big files (checked their md5sums). The funny thing is that after running the Full Diagnostic Scan the raw value is smaller: 169588940 (overflow?). I also found some screenshots from anandtech's review of Intel 520 SSD and his Uncorrectable Error Count raw value was 0. Motherboard: Gigabyte P35-DS3L (yes, SATA2...) Any ideas what's going on?

Thanks in advance!

10 REPLIES 10

idata
Esteemed Contributor III

I found some information that if true, would possibly explain why the UEC count is wrong or a strange value (if it actually is wrong or strange.)

As we know, the 520 is the first Intel SSD to use the SandForce 2281 SSD controller. You may also know that Intel has been working with SandForce (now LSI) for over a year to verify the 520's dependability, and implement some proprietary (to Intel) 2281 firmware changes or fixes. A recent review of the 520 at XbitLabs mentioned that one area of the firmware that Intel changed was in it's error correction.

To quote from the XbitLabs 520 review: "The key firmware feature – compression of data written to flash memory – has of course remained intact but some algorithms (for example, for error correction) have been revised especially for the Intel solution."

Beyond that, regarding what the raw value for UEC should be, or normally is in a correctly functioning 520, I really don't know. I've never owned a SSD with the 2281 controller besides the 520, so I have no experience with the UEC SMART attribute on other SSDs using the 2281 controller. I know nothing about the UEC attribute, so I'm in no position to make any comments about it. IMO, we need information about it before we can have any meaningful discussion about it.

I installed Windows 7 on my 520 earlier this week, and have restarted, shutdown, and rebooted that PC many times, and my UEC value remains non-zero. I'll keep checking it, but I can't correlate the UEC value with anything right now.

idata
Esteemed Contributor III

interesting reading .. thanks parsec

PCali1
New Contributor

I have the exact same problems. please anyone help me !!

idata
Esteemed Contributor III

It's really not a problem IMO, my 520's are working fine, including pairs in RAID 0. I noticed that value will reset to 0 after a reboot or cold start. My 520's seemed to get a non-zero value when they were newly installed, and that subsided over time and reboots. A pair in RAID 0 would both have a non-zero UEC value, although not equivalent values. Quite a coincidence that both SSDs would have the same issue if it was a drive issue. It may be a side affect of something else occurring, like a driver having a problem, or anything among the thousands of things happening on a PC.

Calin, are you having problems with your PC or your 520?

IMO, we should keep in mind how complex PC systems are, and there is so much that the average or way above average user is not aware of or understands, which certainly includes myself.

PCali1
New Contributor

Well, i have no problems in terms of speed ( ATTO benchmark benches the ssd proper). The uncorrectable error count is reset to 0 after the PC is rebooted, but after some minutes the raw values rises again to huge numbers. I've talked to a few people with the same SSD, and they have the same problems. So i guess this is a Intel "problem" or their custom firmware could not be read by software proper.