Chances of Recovery
Start RecoveryHitachi Head Swap: Clean Room Data Recovery and Imaging | HTS541010A9E680 2.5″ Hard Drive with Bad Heads and Platter Damage Hitachi Head Swap | January 19th, 2017
Our customer noted on his mail-in form that his drive was inaccessible and emitted “many loud click and beep noises.” So before anything else, we inspected the hard drive in our clean room. We noticed two of the 4 heads inside were visibly broken. Also, the dust filter was out-of-place (actually on top of the head stack) and there was some dust on the platters.
This is what the heads looked like (but after being safely removed from the drive).
In this case, a partial recovery (by disabling the bad heads in the drive’s firmware) was impossible. This is often possible on Hitachi hard drives, but not when the heads are physically broken. Therefore, we gave our customer the option of paying for a donor drive so we could attempt a head swap. We located a donor online, confirmed our customer wanted to proceed, and ordered the donor drive. It arrived a few days later the Hitachi donor drive arrives and we started the head swap process in our clean room.
After the head swap was finished, we connected the drive to one of our data recovery rigs (PC-3000) to test.
The drive was spinning up without clicking, but had no ID (it wasn’t showing the correct model number/serial number/size). But, the “open” modules were “ok.”
After enabling “Tech Key” detection…
…we can see the drive’s full ID!
Since we now have access to the hard drive’s firmware, we take the precautionary step of disabling “Autoreassign” of the bad sectors.
We test the drive by reading the first and last sectors to confirm the user data section of the drive is okay. Both sectors are readable so we continue to Data Extractor to start cloning the drive.
First, we need to setup our “clone” drive. It needs to be the same size (Set Max LBA in this case) and we also like to wipe a few sectors from the beginning of the drive as well.
Then this drive is connected to the PC-3000/Data Extractor PC.
We make a new task in Data Extractor with all options unchecked (because the drive may not ID properly still).
And select the drive we will clone to…
Since we unchecked “Read Drive ID” we manually enter in the Max LBA value.
First thing we need to do is map the heads, so we know what sectors belong to what heads (in case some heads are weak or the drive has platter damage).
We began cloning the drive. We noticed heads 0 & 1 read without major issues, but heads 2 & 3 are not reading smoothly. Since we have a head map of the sectors, we first clone the sectors from the two “good” heads. This is successful and so far 50% of the sectors have been recovered. Now onto heads 2/3… we need to adjust settings (like timeout, jump size, read commands) and add some model-specific tweaks (like sending tech key and RAM translator initialization) to get these heads to read better.
The result is good and both heads are reading smoother, but not perfectly (there is some platter damage present on the drive). We image the drive in reverse on just head 3 and get back the majority of the data from this head.
Then we continue with head 2 in reverse. After 2 days of cloning, more than 85% of the drive has been successfully imaged so far. We’re confident that 95%+ of the sectors will be recovered once the cloning stage is finished in 1-3 days.
UPDATE: After several more days of imaging, we got back around 98% of the sectors from the User files. Due to platter damage throughout the drive, there are many bad files (although the majority were 100% recovered). We sent our customer a file listing detailing exactly which files were fully recovered and which contain at least one bad sector (and, therefore, may be corrupt). Once our customer decides to pay our $300 fee for the recovered data, or decline our recovery for our labor fee ($100), we’ll update this post. We are really hoping they approve our recovery since this one was not easy. 🙂
UPDATE 2: Customer approved our recovery and all possible files have been recovered.