pes@ux63.bath.ac.uk (Smee) (05/09/88)
(I have no way of knowing if this is of international or only UK
interest. Still, wider distribution is probably better than narrow
distribution, both since Cumana are available worldwide, and because
other makers may produce similar drives.)
I bought a Cumana 1-Meg disk drive (from a company called LAN) at the
recent Atari show in London. Reasons for choosing this were: (a) the
price was right; (b) the Cumana drive was fitted with its own internal
power supply (unlike many current 3rd party drives which parasite off
the joystick port for power); (c) I have a Cumana 1-Meg disk drive at
work, and it works fine; and (d) LAN said the drive was ST-compatible.
(Indeed, once you've gotten into the box and found the instructions,
Cumana also say it is ST-compatible.) It isn't -- though it can be
made to be.
The problem: I took the thing home, hooked it up, and began some
simple tests. It did not seem to recognize media change -- this was
tested by the simple method of opening a window on the drive, swapping
(or not swapping) disks, and pressing escape to force a redisplay.
After some fiddling, I'd narrowed it down, and it was related to the
write-protect state of the disks. If you removed a disk and put in a
WRITE-PROTECTED disk, it would detect the swap; if you took out a
WRITE-ENABLED disk and swapped in a different WRITE-ENABLED disk, it
would not notice. (In full accordance with Murphy's Law, the absolute
worst case. I didn't try WP->WE, by the way.)
Not being shy, I opened the beast up. It contains, right next to the
WriteProtect sensor, a second ('DiskInserted') sensor. A bit of
experimenting confirmed that it does NOT report WriteProtected if there
is no disk inserted. (The ST, of course, relies on the disk drive to
report WriteProtected when there is no disk in, as this is how it
detects MediaChange. {Question for Atari, of course, is, why didn't
they use drives with 'DiskInserted' sensors, and use that to detect
MediaChange?})
A look at the bottom of the PC board showed that both sensors went into
the same three-contact plastic box (a double switch). Markings on this
(tiny) switch box indicated that both switches were normally-open
types, and that the centre terminal was a common terminal. So, I
soldered a small bit of wire in, shorting the 'DiskInserted' sensor so
that the drive always believes there is a disk in. (All this action
occurs beneath the leads to the 'Active' LED, so is a bit tricky.)
Works fine now; though I've probably voided my warranty. The point of
all this, though, is that this is another possible cause of the
'directory window mixed up' problem. If you are sure that your problem
is not identical disk serial numbers, you might want to check this.
If you pull a disk out of the drive and put it back, and then hit Esc
to force a redisplay, then EVEN IF YOU HAVE PUT IN THE SAME DISK, the
drive should spin (and select light come on) as the ST has to read the
header block to check the ID. If the drive does NOT spin, then the ST
has not noticed the disk being removed -- which means the problem
described above may apply.
If you find your drive has this problem, you have two choices, I'd say:
either modify the drive (or have it done by a repair shop) as suggested
above; or try to return it for refund or replacement on grounds that it
is NOT, in fact, ST compatible.
NOTES:
(1) This worked for me. If you decide to open your drive up and fiddle
it yourself, I WILL NOT ACCEPT ANY RESPONSIBILITY if you muck it up.
Make sure you are VERY gentle, and use a SMALL, LOW-LEAKAGE, soldering
iron. Also note that you will almost certainly be voiding any
guarantee if you do so.
AND, if you do decide to mod it, MAKE SURE that you nobble the correct
sensor switch. If you get the WriteProtect sense switch instead, by
mistake, you will end up with a drive which (a) doesn't recognize disk
swaps, and (b) thinks that all disks are write enabled.
(2) (As noted) this DOES NOT apply to ALL Cumana drives. The one I use
at work (bought about a year ago) worked fine right out of the box.
The one I have at home (bought about 2 weeks ago) did not. Cumana
drives, in our experience, are very reliable beasts -- but it is not
clear that they (or someone doing the repackaging, perhaps) know what
'ST compatibility' is. Also, it MAY apply to some other makers' drives.
A Brand A disk drive frequently really is a Brand B drive mech, with
Brand C electronics, packaged in a Brand A box. The point is, if at
all possible, make sure to see the drive working before you buy it --
testing is simple, as described above. Failing that, make sure that
whoever you buy it from gives you a 'statement' that the drive 'is ST
compatible' in whatever terms will allow you to use local consumer
protection laws to return it if it has this problem.
Cheers, Paulpes@ux63.bath.ac.uk (Smee) (05/19/88)
I've just had a reply to my letter to Cumana, in which I explained to them
the error of their ways, and told them how I'd fixed my drive. Aside from
rather profuse thanks, the letter included several bits of information I
thought I might pass on --
1) The fix I outlined is apparently appropriate ('elegant in its simplicity')
2) They will be contacting the people who *they* sold this series drives to,
telling them how to fix it themselves or to return it for modification. I
suspect that this means the *distributors*. Cumana themselves don't seem
to require you to send back a registration card, so are unlikely to know
who the final owners are. So, unless you bought directly from Cumana, you
may have to look up your dealer.
3) The problem relates to a specific variety of Chinon drive mechanism. I
infer from this that it may well affect drives 'made' by other makers than
Cumana, if they have used the same series Chinon drive. (The 'problem',
of course, being that the mech in question is too sophisticated for the
ST's way of working -- creeping featurism -- and so needs one of its
features nobbled.)
4) They had noticed the problem by the time my letter arrived, but had not
yet solved it. They have modified their production procedures so that the
problem will not occur in Cumana drives made since they noticed the problem.
(They had suspended production pending working out a solution.) Chinon
now know about the problem and the solution, so I assume that other makers
using Chinon mechs will also be getting their act together, but don't know
for sure.