Barracuda 7200.7 spinup and identification ok, but no access

[shaun]

Hi all,
I have a Seagate Barracuda 7200.7 here that starts up, identifies itself correctly, the partition table can be read, but when I try to access any sector, no more than 2-5MBytes are transferred before the drive starts spitting out read errors. Repeated seeks can be heard that stops as soon as the access is stopped.
Could the experts probably point me into the right direction what kind of failure I should expect?

[pepe]

Hi

if the symptoms are the same throughout the entire surface, most probably U are having a bad head.
check the number of successive good/bad sectors, if it is ~constant in several periods, then it is head.
U can also run head test.

regards,
pepe

[gsimon]

Hi

if the symptoms are the same throughout the entire surface, most probably U are having a bad head.
check the number of successive good/bad sectors, if it is ~constant in several periods, then it is head.
U can also run head test.

regards,
pepe

pepe,,

How could we start to run head test?
What Software ?

I have the same model/ same problem ...

[shaun]

I would guess that head testing either involves one of the well-known suites like PC3000 or can be done from the serial service console, I'll have to read about this. Maybe someone can point me into the right direction...

I'll do some more careful testing on the disk tomorrow.

I have to arrived at the DR scene so far, and I feel a strong dislike against PC3k and similar, although they are very powerful tools. The work involved would also justify the political price of those tools, but nonetheless they are a bunch of expert tools based on reverse engineering (forbidden in many places) and copyrighted firmware (forbidden to copy in almost any cases, although not justifiable from my point of view - I've paid for a HDD, so I've paid for a usage licence of the embedded firmware; the other firmware files contained are worthless for me before I have a corresponding HDD, so what).

Whatever, I'm willing to learn and although many experts say days of DR are counted, there will always be a need for our skills (yours in DR, my in general electronics), so keep it up

[rameez]

Heads can be tested by using terminal as well .

[shaun]

Good to hear. I've to go through the terminal command description soon.
For now, I've a log from a terminal session I did some minutes ago on the Barracuda.

Maybe you and/or someone else may have a look on it and help me with interpreting what the drive is telling me?
If it is a weak/failing head, would there be any chance to disable it temporarily for reading? From the fact that it's initializing fine and works up to the point when I read the partition table I still have some hope that the service area is not affected, but you may correct me if I'm wrong.

Interface task reset
1024k x 16 buffer detected
ALPINE - 1_Disk M.14

Buzz - Head Mask 0000 - Switch to full int.
Spin Ready
3.06 10-21-03 15:53
(P)PATA Reset

CE Log EC=0 Rtype=36 OV=0 STStatus0
Master with Non-conforming Slave
(H)PATA Reset
Master with (H)PATA Reset
Master with (H)PATA Reset
Master with Non-conforming Slave
AT Er 00 Nwt Er 37 RdWr 05ad.00.0001
ATA St 58 Er 00 Op c8 e, 060003f,08 06 00
Niwot: 0060003f 4a 0000003f.60.000 0000 8000 0d0c 01fe
AutoRd Err 47 at 0005ad.01.02fe

CE Log EC=37 Rtype=5 OV=0 STStatus0
AT Er 00 Nwt Er 37 RdWr 5e0f.00.0002
ATA St d0 Er 40 Op c8 e, 060003f,08 00 00
Niwot: 002c0000 44 00000000.2c.000 0000 0118 01a0 0005
AT Er 00 Nwt Er 37 RdWr 05ad.00.0001
ATA St 58 Er 00 Op c8 e, 060003f,08 06 00
Niwot: 0060003f 4a 0000003f.60.000 0000 8000 0c79 033e
AutoRd Err 47 at 0005ad.01.02fe

[beto]

Thats the typical failure " Autord Err 47" on seagate barracuda´s family , that´s case some time´s reach DRDY DSC, so u can use u tools to tried to stabilize then applied some ATA Algorithms to get some sectors



Best Regards

[shaun]

Could you explan a bit more detailed?
Do you have a clue what is causing this failure?
What category: firmware fault, weak heads, other kind of increasing tolerances, electronics?
What would be the right algorithm to rip some data off the drive? Heavy retrying?

[beto]

In my cases on this models the best for me its tried to get some of sector , how? on first u need to understand "how hdd´s works" many times this is most important than how tools "ace, salvation, hrt etc" works after u understood that how works the hdd´s u can get and applied some settings using your knowledgment and your tools to tried to stabilize the hdd and get sectors , but u need to check contiuilsly the hdd to prevent a head crash, while its getting sectors i had check that, at some models happend that, that´s why some times its need a DR company the experience, and knowledgment

Best Regards

[shaun]

Hi,
thanks for your replies so far,
but I'd still like to know what would probably be the most likely cause for such failure.
Anyone...?

[coolmac]

we called that "common fault" in a certain model,as ST or anyone else brand,so it unnecessary to trace the cause.just to think how to deal with it!
yeah,agree with pepe's opinion,head swap is the final solution to get data back.
if you wanna confirm with you eyes.running "MHDD" under DOS on it,red or green indicator sign will arranged at intervals of a specified number of sectors during scanning the platter surface by pressing F4, and speed slower down in the same time and break in the end.

[shaun]

Hi there,
thanks. Sure, recovery is the goal, but for me (I'm from the electronics side if I didn't mention yet) thinking about a solution for a problem means understanding the problem means knowing the cause.
No, I don't want to give the manufacturer any advice what to make better, but to understand enough to solve it.
For know, it seems that I get every odd 16MB of data, and I'm still looking for an explanation and maybe a fix. (What I mean is that it gives the first 16MB, then from 32-48MB, 64-80MB and so on, as I wrote in another thread)

[craig6928]

sounds to us that you might have a possible bad read head.

[shaun]

Yep, I think so too. I'm just mirroring the hdd with a slightly patched ddrescue program which skips some megs whenever it runs into an error condition, keeping a log on the ranges it already completed. We have about 60% with lots of gaps now, chances are that we will be able to recover some of the excel spreadsheets our customer needs.
The calculation is somewhat simple: an assistant who hacks all the printed sheets into new excel documents will be less than 400 EUR, so I'm doing this mainly for research and improving my knowledge on data structures.