Date : Sat, 06 Jul 1985 17:55:00 PDT
From : Alan Bomberger <ACB.TYM@office-2.ARPA>
Subject: More on AMPRO Little board BIOS
Some problems never seem to get fixed.
My wife has been using an Ampro Little Board for over a year now for a writing
project. I put together a system including TEAC 55B and 55F drives for her.
One of the chronic problems with the Little Board and the TEAC drives has been
intermittant write errors. The latest version (received a few weeks ago) of
their BIOS shows that in the past the problem may even have been solved. It
ain't today. The problem:
How long do you wait for the heads to settle. The controller has a really good
idea that 1 second is good if the motor was not running (and the heads are not
loaded) but Ampro has tried several attempts at optimization which are all
failures.
Katie started having trouble with bad writes after file saves from Magic Wand.
MW always writes to a brand new file which it opens before it allows any
editing. There are no disk reads involved as output to a new file only
requires allocation of blocks. So this problem is the same one observed with
MODEM transfers where the writes are after pauses long enough for the motors to
stop.
Well Katie had been running on an old version of the BIOS which clearly had the
problem as there was almost no wait time for the motor to start. The latest
release "fixes" the problem by waiting for several revolutions (requires
several good reads of sector id of the same id ) after starting the motor.
Still broke!
Apparently older TEAC drives slow down (longer head load times) as his started
to become a burden just this week. I got out my trusty AM radio and tuned into
a good buzz and started poking around with DU. When writing to the TOP side
(the side with the loading head) nothing starts to happen until the head slams
to the surface. In this case the current 3 revolutions for the heads to settle
is quite adaquate. On the BOTTOM side the ID reads happen before the flux in
the head solenoid had even started to build up. The write operation starts
even before the head is loaded and the head slams to the surface in the middle
of the write. The head hitting the disk rattles the disk and completely
destroys any synchronization that may have been achieved before the write
began, result: BDOS BAD SECTOR. Two weeks ago the head came down either before
the write started or AFTER the write ended!.
The only thing that works ALL the time is the technique used in a very early
BIOS version. Turn on the bit in the Disk Controller commands that causes a 1
second pause whenever the motor is started. If the motor is already running,
the pause is not executed. 1 second is not a very long time to wait to insure
that your data gets written on the disk! The current BIOS just isn't good
enough! So maybe my drives are a little slow! Big deal!
The patch is really very easy. The MOTOR ON routine has a few READ ID commands
in it for Ampro's "optimized" timing. Simply patch on the Motor Pause bit for
these READ ID operations and all your troubles are over. Peace !! I hope
that someone is saved an aggravating day by reading this. Incidentally I
bought a TEAC 53B drive and like it a lot. No head loading and very quiet
seeks!