Seagate AirFlow & HDA Temperature Attributes

[fzabkar]

The following thread shows the SMART report for a Seagate ST31000333AS drive:
http://forums.seagate.com/t5/FreeAgent-Products/ST31000333AS-Reallocation-Sector-Count/m-p/47724#M10271

Several aspects of the report aren't clear.

(1) Attribute 194 appears to be suggesting that the current HDA temperature is 50C, whereas the maximum is 40C.

(2) The report does not differentiate between max/min values for the current power cycle, and the max/min values over the lifetime of the drive.

(3) Lastly, I don't understand how attributes 190 and 194 can represent different temperatures when their normalised values appear to be related as follows:

Attribute 190 + Attribute 194 = 100

Here is an excerpt from HDDScan's SMART report:

190 Airflow Temperature 050 040 50C 045
190 Airflow Temperature Minimum 050 040 19C 045
190 Airflow Temperature Maximum 050 040 53C 045
190 Airflow Overheated Count    050 040 440 045

194 HDA Temperature 050 060 50C 000
194 HDA Temperature Minimum 050 060 11C 000
194 HDA Temperature Maximum 050 060 40C 000

[Doomer]

For some reason drive reports Maximum HDA temp incorrectly
Other attributes seem to work fine

As far as I remember HDA temp stores lifetime values but airflow temp stores power cycle values

[fzabkar]

IIUC, you are saying that Attribute 190 and Attribute 194 in fact refer to the same temperature sensor, despite their different names.

So the max/min/current values for the current power cycle would be 53/19/50C, and the max/min lifetime values would be 60C and 11C, with the reported value of 40C being in error?

Elsewhere you have stated that the read heads are themselves used for temperature sensing, by virtue of their thermoresistive properties. This brings me to the following thread where it appears that a particular Seagate drive has a faulty temperature sensor:

http://forums.seagate.com/t5/Internal-ATA-and-Serial-ATA/2-Barracuda-7200-11-failing-miserably-HELP/m-p/47552#M18828

190  Airflow Temperature      45     47     37     069135300035
194  Disk Temperature         0      53     63     000F00000035

1    Raw Read Error Rate      6      119    99     00000C2DE92D
5    Reallocated Sectors Co.. 36     100    100    000000000000
197  Current Pending Sector.. 0      100    100    000000000000
198  Off-Line Uncorrectable.. 0      100    100    000000000000

My first thought is that attribute 194 is not reporting any value for the maximum lifetime temperature. I'm therefore wondering whether you have misinterpreted the structure of the data???

Secondly, if the read heads are responsible for the faulty temperature measurements, then shouldn't we expect this to be reflected in the read performance?

[fzabkar]

I wonder if you are reading the maximum lifetime temperature from the worst case normalised value for attribute 190 instead of the worst case value for attribute 194? The two numbers are not independent. Instead they appear to be related as indicated previously, at least in some models. I believe the reason that attribute 190 is expressed as a difference from 100 is to convey the impression that higher numbers are better, ie higher numbers represent lower temperatures. This can be observed in Western Digital models, some of which use the following relationship:

normalised value of attribute 190 = 150 - temp in deg C

I would think that one way to test the validity of your interpretations is to retrieve the drive's temperature history via SCT. SCT temperature data are documented in the ATA standard.

[Doomer]

Min and Max temperatures come from raw values

[fzabkar]

Min and Max temperatures come from raw values

This is obviously true for the current power cycle, but what are the min/max lifetime temperatures in the second example?

[Doomer]

If they not there then they wouldn't be shown in HDDScan
not all the drives support this feature

[fzabkar]

I'd give you the raw data for the subject drive, but the owner has not replied to my request. :-(

Instead here are HD Sentinel (?) data for two other drives, one of which is an identical model:
http://paste2.org/p/701656

Hard Disk Model ID . . . . . . . . . . . . . . . : ST31000333AS

190  Airflow Temperature      45     69     43     002A1F19001F
194  Disk Temperature         0      31     57     000F0000001F


Hard Disk Model ID . . . . . . . . . . . . . . . : ST31000340AS

190  Airflow Temperature      45     42     37     03D13A34003A
194  Disk Temperature         0      58     63     000F0000003A


I don't see any applicable raw values for maximum lifetime temperature.

Once again ...

Attribute 190 + Attribute 194 = 100