jbww@ukc.ac.uk (J.B.W.Webber) (03/16/89)
**** DON'T PANIC ****
O.K., I think I have resolved the problem with using ADC/DAC cartridges
in newer 1040 STs.
------- THE ATARI STs ARE NOT AT FAULT --------
The scenario was : New model 1040 STs, running Fast Basic, including
some assembler code, reading an ADC/DAC chip plugged into the
cartridge port. The `read' just consisted of a straightforward
`read byte at location', with the device bus enabled with not S3.
(Dead simple, nothing to go wrong, ideal for short student projects,
nothing to go wrong ... Ha. )
Students reported that software and hardware that ran fine on an
older ST failed to run with the new ST. Other Basic software ran,
but attemting to read the cartridge failed. I could see nothing
out of the ordinary on the FastBasic disk.
As a news report of similar behaviour on Mega 2s had appeared the
the previous day, it seemed to be a manifestation of the same.
We had just ordered a Mega 2, to be used with our ST to Transputer
Link cartridges, so I was rather eager to nail this problem ....
So -
Students were back in the Project Lab. today, so I was able to get them to
try an idea that come to me over the weekend - the students here have become
a bit worried over a virus that they have found on a few of the boot disks
for general use, and usually boot with a virus checker, a mouse accellerator
- AND TURBODOS !
When I checked with them, they said yes, they probably had ....
anyway, *** booting a clean system has resolved the problem ***
Quite clearly trying to patch a different (and at least partially
fixed OS) is NOT the right thing to do. I have since heard that
some 1040 STs are getting the 1.2 ROMS from the Mega STs. (Thanks to
Ken Badertscher, Roy Good for very prompt response on the later.)
Sorry to frighten so many people (as well as my self), but with
these cartridge add-ons the Atari STs are becoming rather useful in
the Lab. (although I would like to make more use of their DMA port).
P.S. the write technique I use is a common one that has been round a while,
both on the news net and in magazines (I couple of days ago I just happened
to come across a Byte article that mentioned a variant of the method -
Byte, June 1987, P161 ). Simply, any read of a certain range of addresses
causes the lower bits of the address to be latched and presented as data
to the peripheral - take care A8-A1, or A7-A1,UDS. Anyway, the above
non-problem was for just standard reads.
J.B. Webber jbww@ukc.ac.uk
P.P.S. Any information anyone has on the DMA port would be of great use.
I have some, but it does not always agree with itself or with
what a logic analyser tells me.