Dear Ladies and Gentlemen,
   
   This weekend I have had an almost disastrous experience with
CDC/Imprimis/Seagate Wren 6 drives, and since I would imagine these are
widespread in the VAX world (the SCSI version was used by DEC itself under
the name RZ56), I have decided to share it with you. In the summer of 1997
I have got hold of three such drives (the ESDI version) from Skybridge, the
old Smith Computer Lab (SCL) server (used to be a VAXserver 3600, but once
its RA82 was replaced by these drives, it was no longer a 3600 in the
strict sense). For a couple of months, I have been using two of them in my
IBM PC AT-compatible at home. It has a big case with 10 5.25" half-height
bays, or 5 5.25" full-height bays, depending on how you look at it. The
bays are stacked vertically. I have had the two Winchesters mounted in
adjacent bays (the two bottommost ones). There was perhaps only a
millimeter or two between them.
   
   And so the machine was running fine until lately I had started getting
problems with the Winchesters. The top one started occasionally indicating
write faults over the ESDI interface. Everything was reading fine, so the
drives seemed OK, but I wasn't able to write anything. Then this problem
started occurring more and more often. Then I have noticed an interesting
detail: when the machine was being first powered on, it was running fine,
but after a few hours the problems were starting to occur. Furthermore,
when the drive did write a sector without a fault, reading that sector
afterwards was sometimes producing data errors.
   
   The problem had reached its climax when the latter effect has blown out
a directory sector with references to subdirectories full of important
files. This was an emergency. Trying to come up with an idea of what to do,
I have suddenly realized where the problem was. The heat from the bottom
drive was frying the read/write board on the bottom of the top drive and
blowing out the write tract! I have tried taking the drives out of the case
and putting them bare on the floor, and guess what, the problem has
disappeared! Both drives are now running just fine on the floor of my room.
My theory was right. Then it was only a matter of restoring the nuked
directory sector by finding the subdirectories it was referring to, and
here I have been saved by Norton Disk Doctor.
   
   Moral: NEVER STACK TWO OR MORE WREN 6 DRIVES RIGHT ON TOP OF EACH OTHER!
Each drive dissipates only about 20 W of heat, but as I have found out the
hard way, this enough for a drive to fry the read/write board of the one on
top of it.
   
   Apparently the people who have designed this PeeCee case haven't thought
about this. I wasn't able to find a satisfactory way to put two Wren 6
drives in my case. Oh well, I have figured that all my stuff should fit on
one when I finally write the program to use the cylinders beyond 1024.
   
   Sincerely,
   Michael Sokolov
   Phone: 440-449-0299
   ARPA Internet SMTP mail: sokolov@alpha.ces.cwru.edu