ST506/412 failure modes? (in this case, an IBM 0665)

[tony duell]

> 
> This may be forgotten knowledge - or perhaps more likely, something that
> was never known in the first place - but are there any typical failure
> modes of ST506/412-type drives (beyond the obvious mechanical damage
> between heads and platters)?

For the genuine original ST506 ST412, etc there is a very annoying failure of the
hall effect device in the spindle motor. Annoying becuase the motor is on the
outside of the HDA, but it is impossible to get to the hall device to replace it. 

> In this particular instance, I've got an IBM 0665 30MB drive in a Compaq
> which spins up, bounces the heads around a little, then causes the machine
> to issue a fixed disk failure at boot time. This is an embedded servo drive
> with a voice coil, not a stepper type.  Oddly enough, it passes Compaq

I am sure you know this, but there were ST412-interfaced drives with a 
separate servo surface. Most of the time if there are an odd number of 
data heads then it has a separate servo surface, an even number of data
heads implies embedded servo (or a stepper)..

> diag's spare cylinder read/write tests, but fails the seek test. I've not
> tried a LLF yet because I was interested in trying to salvage whatever data

My first guess is that it is losing the servo signal at some point. Possibly
due to platter damage. Do you get any nasty noises as the heads fly to one 
side or the other and then re-seek to the right track?

> Yeah yeah, MFM emulators and whatnot... but I do want to keep noisy old
> boat anchors alive in my machines for as long as I can - while a modern
> replacement is the only long-term solution, I always think of modern tech
> as detracting from the experience of using old hardware.

Oh I agree with you there... A computer is a lot more than just a processor!

-tony