Yellow blocks are areas that have been skipped. Sounds like the drive has some media damage/contamination perhaps? When I get a drive like this that has possible media damage and does some knocking when it tries to read these sectors I increase the jump size and sometimes block reading size. Also have single sector access enabled. Get it to read the good areas first and come back and try the skipped areas once you've got all the easy to read areas.

I'm sure users of DE have many different tactics and options they use to put as little strain on the drive as possible. If I find a big area of damage I'll stop the copying and then start it from say 1,000,000 in front of that area then come back to that area later, probably in reverse. Also image by head using heads map, always good to do one of those after a HSA swap.

If you open the same task then it won't overwrite the 12% you already have, it will skip that area as it's been read already.

I never leave a potentially dodgy drive, or drive that's had a head swap, to image overnight if I'm not there to babysit it. Have had it many times where it will image along happily and then half way into the drive hit some bad areas and knock horribly. If you don't stop it then it's just going to damage itself and maybe kill the heads. Last week I had a drive in really bad shape, had to baby sit the imaging process for hours and hours every day til I got pretty much all the data. Had I left it to it's own devices it would have destroyed itself.

I'm sure the gurus have better ways of doing things Anyway that's my 5 cents.

_________________
Death is nothing, but to live defeated and inglorious is to die daily.
Data Recovery Cape Town

Thank You Nick for kind advice.
I've just changed "jump size" to 1 000 000 and inform You soon about results.
Drive behaviour is like mentioned here:
- almost every minute drive stays BSY,
- power off/on
- becomes DRD DSC
- reads some sectors
- terminal output is full with errors "Switch to full int"

I think HSA could be weak, but I don't have a new one at moment.
I can purchase a same one but with different Site Code: WU (mine is TK), however tell me if someone experienced success with head swap in Seagate with different site code?

Likely not to be compatible

_________________
Hard Disk Drive (HDD), Solid State Drive (SSD, SATA, NVMe, etc), USB Flash Drive and RAID Data Recovery Specialist in Massachusetts

Dear friends,
I think have found important thing.
During copying sectors when LBA is listed let's say: 11 222 333 everything is ok (drive is DRD DSC)
but when LBA appear's like this: 11 222 444(1) it goes in BSY state, then power's off/on by itself and the procedure repeats.
In my opinion this situation shows that is a problem with "head 1".
In this respect shall I disable this head in ROM and let it read only from 0,2,3?
I will take a look to manual right now to find it out, however I will appreciate any advice from experienced gurus
Thanks in advance.

If I'm thinking about the same (1) that you're talking about, it is referring to the number of attempts at reading that particular sector. When it refers to heads it will actually say Head-0 for example.

There is a tab in the settings to set what the drive does when it loses readiness, eg goes BSY. You change those settings.

If you want to only read by certain heads then you should first create a heads map, then once that is done you can select which heads you want to image with.

I think you need to read the manual before you go any further

_________________
Death is nothing, but to live defeated and inglorious is to die daily.
Data Recovery Cape Town

If I understand your post, then Nick_CT is correct.

These settings can greatly change the outcome of your recovery and becoming familiar with them is a must.

Most of the time I do not have "turn off/on drive's power" selected. Some times it is required though. Also if it's your first pass you may not want to have it retry the bad areas much.

First hit with Google:
http://www.google.com/search?q=%22ST320 ... structions

Barracuda 7200.7 Serial ATA Product Manual, Rev. N (ST3200822AS. ST3160827AS. ST3160023AS. ST3160021AS. ST3120827AS ...):
http://www.seagate.com/support/disc/man ... ata_pm.pdf

_________________
A backup a day keeps DR away.

I have no experience with hardware imagers. However ISTM that it may be possible to use the imager's PSU control signal to control the HDD's POR (power on reset) signal to the MCU.

Without seeing the electronics, I'm assuming that there are high-side MOSFETs which switch each of the +12V and +5V supplies to the HDD. If the gates of each MOSFET were placed under manual control, then their on/off control signal could be redirected to the board's POR via a suitable buffer transistor wired as open collector or open drain. In this way the board would undergo a hard reset without the trauma of having to spin down the motor.

_________________
A backup a day keeps DR away.

can you build map of critical data? What timeouts are you using? Change read timeouts to 50 m/s, seagate respond well to this, it will image much quicker. Fill in the gaps after the first complete pass.

Guys, huge thanks to all of You for advices, especially for Nick_CT.
The recovery process isn't finished yet. I've made head map according to Nick's instructions, selected proper parameters in task window. So in summary, I've made three passes with one head at a time including 1 000 000 jump size.
H0 and H2 read everything perfect. H1 had about 30 red blocks. During copying from mentioned heads I've head BSY state (then power off/on) just a few times.

The weakest one is H3, it is very slow and if meets red block during copying it goes BSY and power's off/on - mentioned restart of HDD happens almost every 1 minute ! At the moment I've got over 1 000 red blocks.

Currently I've got about 80% sectors read successfully from H:0,1,2 and about 70 000 000 sectors to process.

Tell me one thing guys, if I find another HSA donor and make another head swap, will it be possible to read the red blocks again (so it change to green) assuming that reading error was not related to media damage?

@HDD Spaz:
Thanks for info about timeouts, After the fourth pass with weak H3 finishes, I will change UDMA timeout to 50m/s, decrease jump size and provide another one pass using all heads.
Or maybe, is it good idea to change the timeout right now? and let Head-3 finish the job with smaller timeout?

I would lower the head read count and the jump size if you are just trying to read head 3. Is it necessary to repower the drive everytime?

@ Jar:
Repowering is performed by default, after drive gets BSY and holds in this state for a few seconds.
I think it's some kind of security option to wake the drive up after it goes sleep, anyway, I'm not sure if shall I change it, but if You are, would You tell me where I can find this option?

Options are found under the "task parameters" button

_________________
Hard Disk Drive (HDD), Solid State Drive (SSD, SATA, NVMe, etc), USB Flash Drive and RAID Data Recovery Specialist in Massachusetts

Yes, you will read a lot more. Then create submap of skipped (yellow) sectors and repeat process. Once all yellow sectors are read, create submap of read sectors, clear map and repeat. Keep doing this until you have read the maximum amount possible. Once you have read maximum you could try upping the read retries then read ignoring ecc.

Sorry if it's a silly question but wouldn't imaging via MFT have been a possible option? Assuming it's NTFS? Wouldn't that put less strain on the drive & heads? Also your saving time buy only imaging what you want to recover & not the whole drive.

Loki

Yes, unless the customer wants a clone.

Hi NIck_CT, it wouldn't be better decrease reading size? ( less blocks->less stress in reading for heads )
i agree totally about larger jump size.

Hello again, my task didn't finish yet, I've got ca. 85% of data and about 50 000 000 LBA to process.

Meanwhile, I want to ask You the way which DataExtractor makes data copy from ATA0 to ATA1, because I couldn't find it out in manual.

Disk connected to ATA0 has 390 721 967 LBA (about 200GB)
Disk connected to ATA1 has 976 773 186 LBA (about 500GB)

After the copy is made, I want to make an image of ATA1 to file. I don't need the whole 500GB drive image (only data copied from ATA0), so I would select to copy sectors from 0 to 390 721 967. However I'm not sure which way "make data copy" algorithm works?
I mean, is it reading data from sector let's say: 123 456 from ATA0 and ===> copying to the same corresponding sector 123 456 on ATA1 ?

What if disk connected to ATA1 wasn't erased before "make data copy" process, and sector 123 456 contains some data?
During copying process, does the algorithm overwrites sector 123 456 with new data coming from ATA0 or maybe it looks for the first free sector ?

DE copies sector to sector from ATA 0 to 1. So it would copy sector 123 from ATA 0 to sector 123 on ATA 1. It doesn't look for free sectors first it just overwrites whatever is there. If your destination ATA 1 is not clean (zero'd) then enable the "clear not reading sectors" option which will write zeros to ATA 1 in areas where ATA 0 can't be read. You don't want data from an old job mixed up with the image you are currently making.

_________________
Death is nothing, but to live defeated and inglorious is to die daily.
Data Recovery Cape Town