• Please review our updated Terms and Rules here

Hard drive preservation

thenzero

thenzero

Experienced Member
Joined
Feb 19, 2021
Messages
430
Location
Perry, GA
Thinking about this today as I’m about to replace yet another fully working hard drive in one of my machines with a cf card.

What steps do y’all take to preserve your old hard drives? Do you spin them up once a year (or more frequently?) in an external enclosure to keep them alive? Something else? How do you store them? I feel like it’s important to preserve the original drives but I’m not sure how to go about it.
 
I wouldn't waste much time thinking about this. They will all fail eventually, since the servo-tracking data pre-recorded on the platter will degrade over time. There's nothing you can do about it. Funny enough that the very old hard disks using a stepper for head positioning will survive all the more modern ones.

Those drives I had in storage so far and used years later worked without any flaws. I did not check or power them on while they were in storage.
 
Timo W. said:
They will all fail eventually, since the servo-tracking data pre-recorded on the platter will degrade over time.
Click to expand...

That is a very odd failure mode. I've never seen data on a platter 'degrade' (not counting the case of a head crash) but have seen mechanical failures in heads/wires, and the head positioner along with component failures in the servo electrical components.
 
Probably the number one most important thing is just to keep them stored in cool, dry, clean places. Don't store them in a hot attic, damp basement, or dirty garage.

Periodic powering up might help prevent tantalum capacitors from going boom, and might keep moving parts lubricated enough that they don't freeze when you need them. But that is not usually a big deal. With enough drives, if you later do some general test you will find something has just died for no good reason anyway.

BTW, if you do have some data on the drive that you might want to get at later, with ST-506/ST-412 interface hard drives, you might want to leave a note with them specifying which model controller they were formatted with.
 
Servo data comes in two basic forms: an entire surface devoted to servo data or embedded servo data. The latter method has been preferred for many years now. (If you have a drive with an odd number of heads, that might be a hint that it is using one surface for servo data.)

In either case, the servo data is written at the factory. There is no way to rewrite it, unless you have the original firmware that was used at the factory which is basically part of the manufacturing process. (That code never gets shipped.)

Drives with embedded servo data are more likely to have the servo data degrade. When writing, the magnetic signal is not perfectly accurate and depending on the shape of the specific head, it is possible to cause small amounts of damage to adjacent tracks. Modern drives will do background scans to detect when data is becoming degraded and they will rewrite tracks, but they don't rewrite the servo data. So the more writes you do, the more likely it is that you are slowly damaging the servo data.

Enterprise class drives tend to have more space dedicated to servo data for performance and error recovery reasons, so all other things being equal these should last longer under heavy write loads.

For preservation purposes, I spin my drives up once in a while and give them some mild exercise. It helps to spread the lubricate on the platters and to keep capacitors or other things from going bad.
 
This servo tracking data issue recorded on the disk drive, at the factory, with "secret software" is very interesting.

For example, does anybody have a list of Seagate drives, where the servo racking data scheme was introduced, or what model did it first appear ?

For example would a Seagate ST-251 (which is a low level format-able drive) have it ?

What about other more modern drive models ?

If we are fault finding defective drives, we need to know which models are not retrievable due to this issue, and which ones are.
 
All drives after the open-loop stepper models have some servo information on them. The ones with dedicated surfaces have an odd number of user heads.
I'm pretty sure the first embedded servo Seagate drives were the Atasi copies (st40xx) circa 1987 with voice coil actuators
All of their drives from the acquisition of MPI (Wrens, etc.) are closed-loop.
I have a schematic for the 251, it appears to be an open-loop stepper

There was a time when you could find 5" and 3" servo writers in surplus places here in the valley, but they and the stores are long gone.
 
Last edited:
Personally, I wish there were a way to format embedded-servo Drivetec floppies. That was the weakness in their scheme. Factory-preformatted only if you wanted the extended density. Same for later UHD floppies.
 
Al Kossow said:
All drives after the open-loop stepper models have some servo information on them.
I have a schematic for the 251, it appears to be an open-loop stepper
Click to expand...

I had a few ST-251 drives that I tried for my IBM5155. I got one working successfully. Others reported Track 0 errors. Presumably though, these drives then don't rely on pre-recorded servo information are not "Doomed" if there was a way around that track 0 issue, like slight head repositioning ?
 
You must log in or register to reply here.
Back
Top