11 Replies Latest reply on Jan 15, 2016 4:26 AM by factotum

    Intel 520 SSD SMART Power-on Hours Count behaviour

    Snowman1

      I have a 240GB Intel 520 SSD and have been puzzled by the behaviour of the SMART statistic 'Power-on Hours Count' (Attribute 09) displayed by the Intel SSD Toolbox utility. Others have also noted that this Power-on Hours Count statistic is strange and so this is not a fault that is specific to my SSD.

       

      Intel have acknowleged that there is an issue with the SMART statistic 'Power-on Hours Count' -

      See: http://www.intel.com/support/ssdc/hpssd/sb/CS-034131.htm

       

      It would seem useful to know:

      1. Do Intel plan to fix this Power-on Hours Count problem?
      2. Do the Power-on Hours Count values shown have any useful meaning for a given Intel 520 SSD?
      3. If so, then does this meaning also apply to other Intel 520 SSD units?

       

      A response from Intel would be appreciated with regard to Question 1.

       

      To address Question 2 I have done some investigation into the Power-on Hours Count behaviour for my 240GB Intel 520 SSD.

       

      My PC was built with the Intel 520 SSD and two rotating disks. I have owned it for over a year and the Power-on Hours times for all three disks should be almost identical (the Power-on Hours Count values for the two rotating disks agree to within an hour).

       

      Looking at snapshots of the Power-on Hours Count values taken at various times over the year it became evident that:

      1. The Power-on Hours Count values for the two rotating disks were increasing at the same rate.
      2. The Power-on Hours Count value for the Intel 520 SSD was also increasing, but at a somewhat slower rate.

       

      Over a long period (around a year) the increase in the Power-on Hours Count for my Intel 520 SSD was 0.926 times the increase in the Power-on Hours Count values for the rotating disks. This seemed rather strange!

      If the two rotating disks were correctly incrementing their counts every 60 mins then the Intel 520 SSD was incrementing every 64 min 48 sec - not a very obvious time period for an hour!

       

      The other issue is that there appears to be a large base offset applied to the Intel 520 SSD Power-on Hours Count. Using the 0.926 ratio and extrapolating backwards it would appear that the base value when my SSD was first powered on was probably approximately 894,813.

       

      So for my 240GB Intel 520 SSD the formula for converting the 'Power-on Hours Count' displayed by Intel SSD Toolbox (PoHC) into the actual Power-on Hours seems to be:

      Actual Power-on Hours = (PoHC - 894,813) / 0.926

       

      Does this formula (or a similar one) also work for Intel 520 SSDs owned by other users?

      (See Question 3 above).

       

      Some follow-on questions arose:

      1. Do my rotating disks actually increment the Power-on Hours Count every 60 mins?
      2. Does my Intel 520 SSD actually increment the Power-on Hours Count every 64 mins 48 sec?

       

      This was rather difficult/tedious to investigate using Intel SSD Toolbox so I used a utility called System Information Viewer (SIV) from rh-software.com, which made the investigation much easier (I used Menu -> Tools -> Disk Status and left it running for a number of hours).

       

      This showed that:

      1. My rotating disks do increment the Power-on Hours Count every 60 mins.
      2. The interval at which my Intel 520 SSD increments the Power-on Hours Count varies by several minutes. The shortest I have seen so far is 62 min 49 sec and the longest is 67 min 4 sec. (The mid-point between these values would be 64 min 57 sec, which is reasonably close to the long-term average of 64 min 48 sec mentioned above)

       

      Do Intel 520 SSDs owned by other users also show this significant variation in the time taken to increment the Power-on Hours Count?

       

      So in summary I have learnt that:

      1. The interval at which my 240GB Intel 520 SSD increments the Power-on Hours Count varies by several minutes, with a long term average of about 64 min 48 sec.
      2. My 240GB Intel 520 SSD Power-on Hours Count has a large base offset of approximately 894,813.

       

      Do other Intel 520 SSD users see similar behaviour?

        • 1. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
          joe_intel

          Please allow us more time to check about your inquiry.

          • 2. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
            Snowman1

            I now think that if the base offset is the same for all Intel 520 SSDs then the value is probably a few hours lower than my initial estimate above - possibly around 894,803.

            • 3. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
              Snowman1

              I recently had access to the SMART statistics for another PC that has both a 240GB Intel 520 SSD and a rotating disk.

               

              The Power-on Hours Count for this Intel 520 SSD increased by 544 in the same period that the Power-on Hours Count for the rotating disk increased by 551.

               

              The Power-on Hours Count for the rotating disk seems to be incrementing correctly every 60 mins.

              So the Power-on Hours Count for this Intel 520 SSD seems to be incrementing at a long-term average of

              60*551/544 = 60 min 46 sec.

              The time period between successive increments for the SSD varies - I observed successive periods of

              61 min 33 sec, 61 min 3 sec and 61 min 43 sec (all measured to 10 sec granularity).

               

              The estimated base offset for this Intel 520 SSD is around 894,811.

              This is very close to the estimate of 894,813 given above for my Intel 520 SSD.

              Both estimated figures could be out by a few hours depending on how the two PCs were initially built

              (the SSDs and rotating disks could have been installed a few hours apart).

               

              So it appears that the base offset may be the same (or very similar) for all Intel 520 SSDs, but the average period at which the Power-on Hours Count increments can vary between individual Intel 520 SSDs.

              • 4. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                Snowman1

                It would be useful to get some information from Intel about how to interpret the SMART statistic 'Power-on Hours Count' (Attribute 09) for a 520 Series SSD.

                 

                • Is there a definitive value for the base offset for all Intel 520 Series SSDs?

                • What causes the variation in the period at which the Power-on Hours Count increments? (e.g. SSD read/write activity? Power Management?)

                • Are there any upper or lower limits to the period at which the Power-on Hours Count increments? (e.g. Is the lower limit 60 minutes?)

                • Is there likely to be a fix for this problem?
                • 5. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                  joe_intel

                  I hope the following information answers your questions:

                   

                  The description of Power-on hours count (09h) as explained in the specifications document:

                  http://download.intel.com/newsroom/kits/ssd/pdfs/intel_ssd_520_product_spec_325968.pdf  (page 17).

                   

                  “The raw value reports two values: the first 4 bytes report the cumulative number of power-on hours over the life of the device, the remaining bytes report the number of milliseconds since the last hour increment. The On/Off status of the Device Initiated Power Management (DIPM) feature will affect the number of hours reported.

                   

                  If DIPM is turned On, the recorded value for power-on hours does not include the time that the device is in a "slumber" state. If DIPM is turned Off, the recorded value for power-on hours should match the clock time, as all three device states are counted: active, idle and slumber.”

                   

                  Therefore power management really affects this attribute. I know it is a different model but we have a similar discussion with more information here: https://communities.intel.com/thread/48978

                  • 6. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                    Snowman1

                    I had wondered about DIPM, however my Intel 520 SSD is installed in a desktop machine and I believed that DIPM was disabled.
                    According to the specification quoted above the recorded value for the power-on hours should therefore match the clock time.

                     

                    I have just confirmed that DIPM is disabled using the Intel SSD Toolbox and the instructions given in the discussion thread that you referred to - Word 79 Bit 3 is set to zero.

                     

                    So I am still puzzled...

                     

                    • 7. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                      factotum

                      Bit late in the day, but I have a system with two Intel 520 480GB SSDs in a RAID config, and two Western Digital 1TB hard disks in a separate RAID config.

                      The power-on hours counts are as follows:

                      WDC1: 25821

                      WDC2: 25606

                      SSD1: 904333

                      SSD2: 904320

                       

                      The WDC hours average to 2.93 years

                      The SSD hours, using (base 894811 and 0.926 divisor), average to 1.17 years.

                       

                      So I guess the base is not the same for all.

                      The SMART stats from the SSDs don't look all that plausible. I got this system in April 2013 and the SSDs still say normalised Media Wearout Indicator 100 (= no wear).

                      I wanted to know when to replace these SSDs before they fail on me. But it seems the stats are not much help.

                      • 8. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                        jonathan_intel

                        Hello factotum,

                         

                        As you may know the Power on Hours Count of the Intel® SSD 520 Series has an offset (894794) that has to be taken off the reported value to obtain the actual power on hours count.

                         

                        The Media Wearout Indicator is actually the best way to determine the expected remaining life of your drive, as the normalized value declines from 100 to 1 as the average erase cycle count increases from 0 to the maximum rated cycles. Once it reaches 1, the SSD may continue to work well, however, it would mean that it reached the maximum rated cycles for the NAND memory.

                        • 9. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                          factotum

                          Thanks. Yes, well as I clearly stated in my post, I have adjusted for the offset.

                           

                          What do you make of the fact the the normalised Media Wearout Indicator is still 100 after being used every day, but off at night, since April 2013? Do you think that figure can be trusted?

                          • 10. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                            jonathan_intel

                            Please keep in mind that the Intel® SSD 520 does not count power-on hours when it is in low power mode (standby), this is probably the reason why the Hard Disk drives show a higher amount of hours.

                             

                            Regarding the Media Wearout Indicator, we consider it to be the best method to estimate the wear of the NAND chips. We are not aware of misreporting issues in this attribute of the Intel® SSD 520 Series.

                             

                            It is worth to mention that the amount of writes is the main factor in the wear of the drive, if a system does mostly reads, then we would expect the SSD to have a longer lifetime.

                            • 11. Re: Intel 520 SSD SMART Power-on Hours Count behaviour
                              factotum

                              Thanks. My read/write ratio is 68% on one drive and 64% on the other. I will assume all is well for some time to come.