Platter alignment

[Sling Shot]

Currently working a case whereby customer has not only opened his drive, he decided to remove the platters too, usual stuff of finger prints etc. but they don't look damaged. Had a hell of a time getting sector access but long story short, it's imaging. 

Drive is : 2.5 Fblite 

Model : WDC WD5000BMVW-11AJGS1

Firmware : 01.01A01

Capacity : 500 GB (976 707 633)

Currently imaged about 1/3 of the drive and the data in sector editor shows good but theres zero file system and scanning the image for RAW, using DE, R-Studio or UFS doesn't find anything (exception R-Studio finds 1 .mov file).

Part of getting this drive back to life involved regenerating the translator module 0031.

My question is this: Some drives do not require platters to be aligned, however I do not have any access to any resources that demonstrate the models that either do or don't need it. Do you guys think my problem here could be the misalignment of the platters, as to why I can't pull data from this drive or could it be something else ?

Many Thanks.

[Data Recovery Guru]

It's a USB based drive.

Are you cloning via USB pcb? Or SATA PCB?

I assume SATA pcb since translator regeneration took place.

So, via SATA pcb, the data is likely hardware encrypted, hence not finding anything.

Have you tried decrypting in PC3K-DE? 

--
You received this message because you are subscribed to the Google Groups "DataRecoveryCertification" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datarecoverycertif...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/a9d2f1bb-28a8-4a69-bf9a-b28ed53a9adfn%40googlegroups.com.

[Sling Shot]

Outstanding :) 

Thanks for your help man

This drive has put me through hell and that's the one thing that completely left my mind.

Decrypting the image in DE is giving me data access, so glad you suggested it.

I also guess this is one of the drives where alignment ain't so important

[PCLAB]

We started a list here: https://www.recoveryforce.com/forums/viewtopic.php?f=6&t=565&hilit=platter+alignment&sid=55f82ac6cc6b52a7a97a0abac3d34d90

of drives/models already tested and that don't have issues with alignments.

Any new cases confirmed will be welcome here.

[lcou...@recoveryforce.com]

FBLite now added to the list on the Recovery Force forums

[Sling Shot]

Excellent to see such a resource being compiled, happy to add to that as I go and I will be sure to conduct some more tests and add more to it.

I dont know how a misaligned drive would behave as I have never seen one, if a drive was sensitive to alignment then what would it likely do if misaligned ?

[lcou...@recoveryforce.com]

If a drive doesn't work with misaligned platters, you will usually experience one of the following two symptoms:

- detects, but only the bottom two heads read

- clicks and doesn't detect at all

Sometimes there are some tricks to even work with misaligned drives by playing with the headmaps.  But, for such cases, it is a lot easier to just keep the platters aligned when transplanting.

[Data Recovery Guru]

Depending on which surface (system surface/head), surface may not be accessible or drive does not initialize.

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/32a6d155-b898-4bab-8662-12f5fb21a580n%40googlegroups.com.

[t...@desertdatarecovery.com]

Can you have the SA platter misaligned?

 

 

Tim Homer - Lead Engineer

Desert Data Recovery

t...@desertdatarecovery.com

www.desertdatarecovery.com

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/CAHoXmqmQnfS%3DbG2uhCqjVYamSzYiUxwu8HqrhE0_gcchrOwq%2BA%40mail.gmail.com.

[Data Recovery Guru]

If it is drive with only system copies 0 and 1, no. For those with copies on 2 and 3, I guess so.

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/042301d69cbb%24bc1ddf20%2434599d60%24%40desertdatarecovery.com.

[Sling Shot]

I thought data was written through all platters, making the read/write faster than if just to one. Thus, if one platter is moved it becomes misaligned with its counterpart, rendering said file broken, due to the co-ordinates of location being wrong. I guess this is incorrect in some cases ? 

I still haven't finished this case so I don't know if its conclusive yet that 'fblites' aren't sensitive to alignment.

Reading data off heads 0,1,2 quite well but 3 is very poor. 

Head 0 - good

Head 1 - patchy but steady

Head 2 - patchy but steady

Head 3 - bad (more likely to be platter as that was the most contaminated)

Once I've finished securing a decent stack of user data I will revisit the platters and heads issue and try get a better result. 

By proxy that result will help know if alignment is an issue or not with fblites.

[Data Recovery Guru]

You are right.

Tim was just asking for disk 0 with SA copies 0 and 1. Those two surfaces cannot be misaligned in terms of each other.

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/c2f87033-2ae2-46ac-9287-4fb2d784108en%40googlegroups.com.

[Sling Shot]

Yes for sure, I didn't mean to be answering to Tim's statement, I just meant in general that's how I thought data was written to drives.

[t...@desertdatarecovery.com]

@Sling Shot. Data is written ‘through’ all platters but not because of speed. A HDD only ever reads and writes with one head at any one time (although Seagate are looking to change that). So swapping heads would theoretically make the speed fractionally slower. I do not know, but I suspect it is just to spread the wear out over the set of heads/platters to extend the life of the drive.

 

We do not fully understand how manufacturers use servo markers. Some drives can have misaligned platters and read data fine, so we can assume when the drive switches heads to read from a new platter it finds the servo markers for that platter and carries on reading. However some drives cannot have the platters misaligned, so perhaps those drives expect the servo markers in a particular place on each platter and when its not there, it does not read.

 

So theoretically as platter 0 an 1 are the ‘base’ reading for servo markers they can never be misaligned. However there are also other factors such as balance of the platters that can also affect reading after platter swaps.

 

Tim Homer - Lead Engineer

Desert Data Recovery

t...@desertdatarecovery.com

www.desertdatarecovery.com

 

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/75d04297-566b-4af3-9d9d-9d0f395e9e28n%40googlegroups.com.

[Sling Shot]

I didn't know that Tim, I thought they could read/write from all heads simultaneously. So, taking that theory, does that mean that most drives are then filling up a platter at a time, rather than spreading it across multiple platters ?

One of the AceLab guys told me that all drives have servolabels that allows to determine the desired sector addresses, even in case of bad platter alignment, but this affects the reading speed a lot. And if the alignment is completely destroyed than it won't read anything at all, neither file structure nor files in raw recovery.

I am assuming he means by 'totally destroyed', either severe media damage or platter in upside down, in other words unable to be read whatsoever.

[wayne horner]

The slowest action in reading a disk is seeking

once it gets to a cylinder it is fast to read the other heads at the same cylinder address

because no seeking is involved when changing reading from one head to another at the same cylinder

then there is some rotational lag when switching heads

so for efficiency they can map the sectors so that they will be over the needed one right after the switch.

If they miss the sector then they have to wait for a complete rotation to catch it again.

Alandata Data Recovery -  (949)287-3282  
"Cleanroom Data Recovery of RAID, VMware, Network Attached Storage, Linux, Tape, Disk, Forensics"

www.AlanData.com   www.AlanDataRecovery.com

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/199b45e2-9053-4574-81e0-383e0526df62n%40googlegroups.com.

[t...@desertdatarecovery.com]

@Sling Shot: Each drive has zone allocation tables. Those detail which head reads from which platter and when. A good way to see this in action is to create a task in DE. Build a head map and then start imaging the drive. You will see the drive will (for example) read xxx sectors from H0, then H1, then H2, then H3, then H2, then H1 etc etc. So if in real life you save one large file to a hard drive it will normally be stored on several platter surfaces and written by several different heads.

 

It was believed until a little while ago that any misalignment of platters would lead to no data. But we know from Lukes forum posts that is not the case for SOME drives. When most of us do platter swaps we always try to keep the platters in the same orientation, but sometimes that is not possible if you are hunting for missing sliders or cleaning contamination.

 

Tim Homer - Lead Engineer

Desert Data Recovery

t...@desertdatarecovery.com

www.desertdatarecovery.com

 

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/CAJEn65ouW%2B9h8%2BMW17_fGhVXC3SitC0SnkuYREaBAOvDTo1X-w%40mail.gmail.com.

[wayne horner]

adding... to zones

the zone allocation tables describe the density in those cylinders

the outer cylinders are bigger and have a higher sectors per track compared to the inner cylinders

also

the service area is usually written in a low spt to make them more reliable

Alandata Data Recovery -  (949)287-3282  
"Cleanroom Data Recovery of RAID, VMware, Network Attached Storage, Linux, Tape, Disk, Forensics"

www.AlanData.com   www.AlanDataRecovery.com

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/049501d69ce2%24f0837fa0%24d18a7ee0%24%40desertdatarecovery.com.

[Sling Shot]

Yes I've only ever removed platters when I am in charge, so I mark them and realign once re-installed. Not had any issues like that. But this case has brought a new thing to my attention, what happens when someone else removes them, and doesn't put them back in alignment of any kind.

I know from experience of problematic 'outlook.pst' files spanning multiple platters that some drives must write them like that, but I assumed the drives that don't need alignment only write to a platter at a time. This now seems like fuzzy logic and its more to do with servo labels as to whether alignment is necessary or not. It certainly makes things bit easier thinking about it this way and having a list of drives that don't care about alignment is really helpful, especially in this 'DIY Data Recovery' era that we are all experiencing. 

[t...@desertdatarecovery.com]

I know from experience of problematic 'outlook.pst' files spanning multiple platters that some drives must write them like that, but I assumed the drives that don't need alignment only write to a platter at a time.

 

No. The drives that don’t need so much alignment use the servo markers one each platter surface to tell where the data is. So misalignment is not an issue (to an extent). But perhaps other drives only have servo markers on the SA platters and ‘expect’ all other platters to be aligned. So when the drive need to find LBA xxx on platter 2 it finds LBA yyy, it will give an error.

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/7e9db1e5-0f46-46b5-aefc-fbbf8af2b661n%40googlegroups.com.

[wayne horner]

in the old days...

maxtor drives only had servo on 1 head the other heads tracked along using that servo

this made headswaps difficult because you needed heads that were near enough to read

I think embedded servo mixed the servo with all the tracks

microjogs are something else

they relate to how much each head is twisted

Alandata Data Recovery -  (949)287-3282  
"Cleanroom Data Recovery of RAID, VMware, Network Attached Storage, Linux, Tape, Disk, Forensics"

www.AlanData.com   www.AlanDataRecovery.com

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/7e9db1e5-0f46-46b5-aefc-fbbf8af2b661n%40googlegroups.com.

[Sling Shot]

Glad I started this thread now as its been a great insight to in-depth information about a little known part of the industry. Many thanks for all your input guys :) 

[pbzcbf...@gmail.com]

What if the headstack is sitting on track X of head A and is then commanded to read a sector on track Y of head B on a different platter? Surely it doesn't need to seek to track Y of head A before switching to head B??? Also, I would think that head switching would involve micro-jogging to account for differing TPI on each platter.

[Alandata Recovery]

I think it in general works like this:

It initiates the seeks to cylinder y.

while its twiddling its thumbs waiting for the seek to complete it would have plenty of time to

electronically switch heads.

On Wed, Oct 7, 2020 at 6:27 PM pbzcbf...@gmail.com <pbzcbf...@gmail.com> wrote:

What if the headstack is sitting on track X of head A and is then commanded to read a sector on track Y of head B on a different platter? Surely it doesn't need to seek to track Y of head A before switching to head B??? Also, I would think that head switching would involve micro-jogging to account for differing TPI on each platter.

--

You received this message because you are subscribed to the Google Groups "DataRecoveryCertification" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datarecoverycertif...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/e964e0ee-b843-4fcd-af27-8b510f02dba1n%40googlegroups.com.

--

Alandata Data Recovery -  (949)287-3282  

"Cleanroom Data Recovery of RAID, VMware, NAS, Linux, Tape, Disk, Forensics"

www.alandata.com   www.alandataRecovery.com

[pbzcbf...@gmail.com]

The point I was making is that you cannot simply electronically switch heads and expect that the other head will be on-track. That's because each surface has a different number of tracks. After switching electronically, the next head would need to reposition itself slightly ("microjog").

[LarrySabo]

On Wednesday, October 7, 2020 at 10:13:25 PM UTC-4, pbzcbf...@gmail.com wrote:

....  because each surface has a different number of tracks. After switching electronically, the next head would need to reposition itself slightly ("microjog").

Really?  That's news to me.  I assumed the tracks were concentric, one above the other, just on different surfaces.  References, please?

[compos mentis]

HDD inside: Tracks and Zones:

https://hddscan.com/doc/HDD_Tracks_and_Zones.html

"... some heads can write with larger TPI but some won't and modern hard drives may use different TPI for different heads in different zones, such feature is called Variable Tracks Per Inch or VTPI."

You can see the VTPI in Seagate drives via terminal. I've seen some massive variation between heads.

[Alandata Recovery]

so when the build a drive

they get it all assembled

the platters are blank - no servo

the then connect it to a massive machine to isolate it from  vibration of like cars driving by...

a robot arm reaches into the drive and grabs the arm

they spin it up and start writing servo

they mechanically  step the heads cylinder by cylinder

when the done all of the heads are in perfect alignment with their tracks - because they wrote the track.

at least thats how they used to do it...

--

You received this message because you are subscribed to the Google Groups "DataRecoveryCertification" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datarecoverycertif...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/datarecoverycertification/18fe085b-c296-42bf-a3bc-01e36464202co%40googlegroups.com.

[LarrySabo]

Thanks for that, Franc

[Sling Shot]

Forgot to reply to this, got a great result in the end, so can confirm FBlite's don't care for alignment.