Peter,

I have made a small change to the QLROMEXT logic so that it will disable the
interface if an access to address zero (the location holding the 68xxx reset
vector) is made. This kludge means that if the QL is reset then the
interface is guaranteed to be disabled - this actually solved my White
Screen of Death issue. It's not perfect, of course, because if some code
accesses address zero for other reasons (perhaps a debugger or a monitor
program)  then the interface will get disabled but I think it's a change
worth making because that reset button gets a lot of use on some QLs ;)

As for the clash with the ROM I have thought about that some more. I have
absolutely no explanation for the Double Banner message with my Ser-USB ROM.
A Toolkit 2 ROM does not do this and the Ser-USB ROM doesn't do anything
different. I can't see how the ROM could get enabled on other addresses
(which would cause the double banner if it appeared in several places in the
memory map) as the QLROMEXT does not affect the address or romoe lines that
go to the ROM socket, so I'm wondering if the ROM initialisation routine is
screwing up in some way. Any suggestions would be very welcome because I am
genuinely stumped.

Not being able to read registers when the ROM cartridge is plugged in ...
now that I've thought some more about it, I can see that you're right: even
if the Ser-USB ROM has a value of $FF at address $FFF5 (the Read MISO
address) that doesn't mean that the CPLD will be able to drive the data bus
to its "correct" value. The ROM in the socket is enabled, therefore its
outputs are _driving_ the bus high. I was mistakenly thinking open collector
with pull-ups.  Actually, this could be an issue, because people want the
ability to plug in a ROM cartridge while the QLROMEXT is in place ...
especially if they have more than one ROM cartridge.

OK, so that leaves the Trump Card and, again, any thoughts you have would be
most welcome. As far as I am aware the TC does not map anything to
$C000-$FFFF so it shouldn't be interfering with the QLROMEXT. Do you agree?

Regards,



Adrian
www.memorylanecomputing.com




_______________________________________________
QL-Users Mailing List
http://www.q-v-d.demon.co.uk/smsqe.htm

Reply via email to