Samsung HD204UI/Z4 massive CRC errors

[Chungalin]

I have a problem with a Samsung HD204UI/Z4 that throws CRC error trying to read any sector in these 3 areas:

from 1.654.856.800 to 1.686.890.495 (~15GB)
from 1.833.449.344 to 1.864.833.599 (~15GB)
from 2.009.776.216 to 2.039.768.519 (~14GB)

Quite strange, because it doesn’t give any error outside these ranges, and the Good/No Good transition is sharp. It reads fine to the last sector before these zones, and then it’s CRC error for every sector.

I tried the firmware patch F4EG (since it was manufactured on 2010/08) with no improvement.

No strange sounds inside the drive.

What could be the diagnosis based on this error pattern?

[jermy]

can we see the terminal log when trying to read those sectors ?
what happens when reading in reverse ?

[Chungalin]

At this point my main interest is to see if this HDD can be reliable or fully formatable in this state. No data recovery required. Oddly enough, I cannot access SMART info via my external USB enclosure, while I can do this for most of other HDDs.

I don’t have any pro tool. I used WinHex to determine unreadable areas.

[fzabkar]

You can dump the firmware modules with SHT demo:
http://www.hddoracle.com/viewtopic.php?f=149&t=1565

How to connect a terminal adaptor on a Samsung drive ?
http://www.hddoracle.com/viewtopic.php?f=116&t=189&p=432

ISTM that the drive has a [firmware?] problem with one particular head. The three bad areas appear to correspond to 3 consecutive zones. They are separated by error-free areas, each with a size of 75GB. This appears to be consistent with 5 x 15GB zones. This would suggest that the drive reads through one bad zone on the problematic head, then traverses the corresponding zones on each of the other 5 heads before returning to the next zone on the bad head, and so on.

[Chungalin]

Thanks, fzabkar, that’s an interesting analysis. At this moment I don’t have a low voltage serial adapter, but I’ll get one and check what you say.

[lcoughey]

You have a 200GB drive with 4 heads. Samsung uses very large zones for each head. 1/4 of the drive (44GB) is not readable because 1 head is dead. At least that is how I see it when I read through the OP quickly.

[lcoughey]

Nope, my theory is out...it is 2TB drive model. Could just be coincidence and is just 3 bad zones due to media damage. It does seem odd, though.

[UhClem]

I agree with fzabkar's analysis ... with a couple of refinements. There is no doubt that ~15GB of consecutive sectors/LBAs are unreadable. Now, that # of tracks (~12k) seems a lot for a single serpentine pass, but Mfgrs/models will employ varying strategies ["Different strokes ..." ]. But those 3 (x15GB) runs might comprise a single Zone, and, maybe, a single ancillary, but critical, low-level formatting/servo-related (factory-calibrated) parameter for that one head/surface for that one Zone, has become glitched.

Considering that the head must be OK (reads every other area), and it's hard to imagine that single band of that surface being so consistently, and totally, damaged, without the corresponding head being damaged; suggests a non-hardware cause.

Out of curiosity, did you find this drive by the roadside? or do you know its history prior to this fubar discovery?

I'd feel safe partitioning it as 825 | 15 | 75 | 15 | 75 | 15 | 975 , not using the 3x15s.

--UhClem