Sticking HDD heads

[Brian Todd]

As I have posted elsewhere, I have a Zenith SupersPort laptop with 20MB JVC drive. The heads have a horrible habit of sticking to the platter at seemingly random times. Due to the fact the controller is JVC and not IDE, I have not been able to source a viable replacement drive. The heads stick so much that a whack on the side of the drive does not help. The only solution is to dismantle the drive and carefully move the platter until the drive spins freely again. Dust or no dust, I have done this at least 10 times and the drive boots, reads and writes perfectly still. The issue is that it is just a matter of time until the heads stick again, and as the arm moves in tangiental direction there is no park possibility.

I am pretty much at the point of wanting to dismantle the drive so that I can clean the heads properly. My issues are that I do not know how to remove the platter without damaging the heads as they do are always on the platter. The other issue is that I have no clue what I could do to clean the heads so that they do not stick again. I know that there is a high risk in destroying the drive, but the way I see it is that its pretty much scrap anyway so nothing to lose! Any advice would therefore be very welcome!

 

[GeoffB17]

My understanding always was that it was a small amount of magnetism that caused the sticking. The head has got slightly magnetic, not enough to interfere with the disks normal operation, but enough to cause the head to 'stick'. Just the act of moving the head will free it, for a while, hence the (gentle ?) bash usually helps.

Or do others know better?

Maybe the head needs a tiny amount of demagnetising?

Geoff

 

[Timo W.]

The head has got slightly magnetic, not enough to interfere with the disks normal operation, but enough to cause the head to 'stick'.

That's not possible. Apart from the fact that even 10-times the magnetic force used to write data to the platter would not be enough to make the heads stick, they just can't become magnetic without current running thru them. Also, the platters are made of aluminium or glas - a magnet won't stick on them. Think about it. A hard disk would be impossible if your theory would work, as every write attempt would end in a head crash.

If the spindle motor is old and has gotten weak, the adhesion between the sleek surfaces of the platter and the heads can be enough already to prevent the disk from spinning up.

 

[Brian Todd]

If the spindle motor is old and has gotten weak, the adhesion between the sleek surfaces of the platter and the heads can be enough already to prevent the disk from spinning up.

The heads physically leave material on the platter and I have to free them up very carefully, so it actually seems to be physical adhesion.

 

[mbbrutman]

My understanding always was that it was a small amount of magnetism that caused the sticking. The head has got slightly magnetic, not enough to interfere with the disks normal operation, but enough to cause the head to 'stick'. Just the act of moving the head will free it, for a while, hence the (gentle ?) bash usually helps.

Or do others know better?

Maybe the head needs a tiny amount of demagnetising?

Geoff

This is absolutely wrong.

Sticking heads is called "stiction", and it's usually caused by the thin layer of lubrication on the platters. If the heads land on the platters (as they do when they park for most older drives) and sit for too long, the heads can get stuck in place. This was much more of a problem in the 80s and early 90s; the chemical composition of the lubricant has been improved and mechanical changes, such as mildly dimpling the landing area make it much less likely to happen.

Side note: IBM designed drives were famous for this. They moved to a design later that removed the heads from the platters entirely, making stiction impossible after that. Designs derived from these drives (HGST and WD) probably are still using the same "ramp unload" mechanism today.

 

[Brian Todd]

This is absolutely wrong.

Sticking heads is called "stiction", and it's usually caused by the thin layer of lubrication on the platters. If the heads land on the platters (as they do when they park for most older drives) and sit for too long, the heads can get stuck in place. This was much more of a problem in the 80s and early 90s; the chemical composition of the lubricant has been improved and mechanical changes, such as mildly dimpling the landing area make it much less likely to happen.

Side note: IBM designed drives were famous for this. They moved to a design later that removed the heads from the platters entirely, making stiction impossible after that. Designs derived from these drives (HGST and WD) probably are still using the same "ramp unload" mechanism today.

So is there any way to remove the "stiction" by alternate means of lubrication?

 

[GeoffB17]

The dates certainly tie up. I got seriously involved with PCs very late 80s, and we certainly had problems (occasional) with heads sticking, but I've never heard of any problems more recently.

I think I've heard mention of 'sticktion' before.

I was going to ask if, therefore, there was any relevance to the 'cleaning' that OP was asking about, but I see that he's just come back to - in effect - raise the same point. I assume that it would NOT be good to remove the existing lubrication?

Geoff

 

[mbbrutman]

So is there any way to remove the "stiction" by alternate means of lubrication?

If you have a drive that has stiction problems you can sometimes free the heads by giving the drive a vigorous spin on it's axis of rotation. The idea is to use the weight of the platters; as you provide that vigorous spin you are generating a lot of torque against the heads, which are relatively fixed in position. The spindle motor normally does this too, but the force required might be too high which is why the drive isn't spinning up. The "spin on the axis" trick might generate more torque.

There is also a tapping trick along the same lines. We called it "percussive maintenance."

Putting the drive in a older place for a few hours and then spinning it up sometimes helps too.

Note that the first two techniques are dangerous, as they can just rip the heads off of the arm and leave them on the platter, destroying the drive. If you are desperate try the spin technique first. Hold the drive the entire time so that it doesn't go flying and impact something. Give it a vigorous shake on the axis of rotation for the platters, and that might free things up.

Do not attempt to clean the platters; the lubricant is there for a reason. Unfortunately, it didn't age well.

To prevent stiction you can try exercising the drive, with the intent of moving the head around to all of the surfaces of the platters. A build-up of lubrication makes it more likely that you'll have stiction problems when the drive cools off. Also, always park the heads. Dimpled landing areas greatly cut down on stiction problems.

 

[Brian Todd]

If you have a drive that has stiction problems you can sometimes free the heads by giving the drive a vigorous spin on it's axis of rotation. The idea is to use the weight of the platters; as you provide that vigorous spin you are generating a lot of torque against the heads, which are relatively fixed in position. The spindle motor normally does this too, but the force required might be too high which is why the drive isn't spinning up. The "spin on the axis" trick might generate more torque.

There is also a tapping trick along the same lines. We called it "percussive maintenance."

Putting the drive in a older place for a few hours and then spinning it up sometimes helps too.

Note that the first two techniques are dangerous, as they can just rip the heads off of the arm and leave them on the platter, destroying the drive. If you are desperate try the spin technique first. Hold the drive the entire time so that it doesn't go flying and impact something. Give it a vigorous shake on the axis of rotation for the platters, and that might free things up.

Do not attempt to clean the platters; the lubricant is there for a reason. Unfortunately, it didn't age well.

To prevent stiction you can try exercising the drive, with the intent of moving the head around to all of the surfaces of the platters. A build-up of lubrication makes it more likely that you'll have stiction problems when the drive cools off. Also, always park the heads. Dimpled landing areas greatly cut down on stiction problems.

Thanks MB. My drive is beyond "percussive maintenance" as I physically have to disassemble it and rotate the platter by hand As the arm moves in tangential direction there is no possibility to park the heads (this is a late 80s drive) so that doesn't make things easier. I have tried exercising the drive many times, but when it cools down ... yep you guessed it. Is the bottom line that the drive is scrap?

 

[Gary C]

The word 'stiction' is a portmanteau for static friction rather than actual adhesion.

 

[mark0x01]

I recall back in the day that cold made the problem worse, so would advise users to leave the computers running in winter.
Seagates commonly had the issue, and percussive maintenance was often the cure.
Warming it up may also help - gentle to around 30C, not cooking it
One Seagate version of the lubricant thickened over a reasonably short time and make the problem more frequent for a while.

 

[mbbrutman]

mark0x01 - Warm would seem to be better in this case. I was probably crossing wires with failing electronics when I suggested cold.

Brian - If you have already taken it apart it's not going to be a reliable drive anymore. When I get to that point with a drive I use it as a demonstrator:

-Mike