A16 Bit

[droege@snapmail.us]

The last several weeks I have been fighting a strange bug in the linux
program.

I think I have now found the problem.  It is the kind of bug that only
shows up in testing, never in production.  But one tests before
production, so one has to fight such bugs.  The fearless who go straigt
from the drawing board to production would never see this bug.

The problem is that the Mark IV is designed to run two scan programs. 
This might be used, for example to run a fast scan of the center of the
image for focus or any fast scan application.  Which program is used is
determined by the A16 bit which shares a position on the register that
selects a test read channel.

Apparently the A16 bit was not initialized.  So it would run correctly or
not depending on the chance turn on state of this bit.  However, doing any
ADC read, like reading the CCD temperature, would automatically give it
some value.  Since the system worked, the some value given was probably
the right one. In testing one does a lot of power on an off.  So the
problem would come and go.  If the test did not include a temperature read
or some such, then the bad state would persist until the next power up
chance position.

Thus while testing we would see these garbage images.  I finally tracked
this down by switching cables back and fourth between a Linux and Windoz
machine to run the telescope.

A great relief.  Tonight it is clear and I can try to get some images.

Tom Droege