RD53 failure modes

[Carlos E Murillo-Sanchez via cctalk]

Hello folks;

I've recently had some free time and decided to look at some hardware 
failures in my small collection.  I fixed a couple of analog boards in 
the compact Mac department, and a failed scsi disk in one of them 
prompted me to test the small stash of such drives that I have.  Turns 
out, about half of those marked as "working" in 2015 have failed now in 
many different ways.  Oh well... sucks, but it was to be expected...

I decided to image some of the still working ones, both in the stash and 
in working machines.   The one in the vaxstation 2000 (a 1GB scsi with 
5.5-2) turned out to have file system corruption (probably from  AC 
power failures), which has taken a while to fix.  And then, I remembered 
that said system originally came with an RD53 that has sat elsewhere for 
25 years.  It passed a read test in 2005.  So I tried to see if I could 
read it and maybe image it now, but no go. The disk spins up, initiates 
a seek (the arm is not stuck in a gooey stop pad; I've read that this is 
a common failure mechanism for Micropolis 1325's; I opened it and saw 
the arm move) but then the arm returns violently (clacking) to the rest 
position; it does this a number of times (two to four, usually) and then 
it spins down. Applied voltages and currents remain ok as this happens.

I've seen plenty of internet content about solving the stuck arm 
problem, but not this.

Advice, please?

Carlos.