[Coco] RS232 Pak needed

William Astle lost at l-w.ca
Fri Sep 20 13:10:50 EDT 2013


Mark can probably straighten out any misunderstanding on my part, but as 
I understand it, you can completely ignore the SCS signal and other 
miscellaneous dumbness in the design and just decode the address lines 
directly. I believe asserting SLENB disables the internal address 
decoding stuff in some manner so if you decode the address and determine 
it's yours, you assert SLENB and basically take control. Thus, no need 
to adjust the hardware inside the machine.

You have to be careful with PIA0, though. You can't re-purpose 
FF1C..FF1F - later versions of the Color Basic ROM actually use that 
ghost address during initialization.

On 2013-09-20 10:55, Gene Heskett wrote:
> On Friday 20 September 2013 12:23:54 William Astle did opine:
>
>> On 2013-09-20 06:08, Gene Heskett wrote:
>>> Assuming we do a 2 port version, and we could offer it for $60 by the
>>> time we get gold flashed boards, find out who made the cases Roger
>>> used, get some of those & modify that one for a pair of db9's, how
>>> many of you folks would buy it in the next year?
>>
>> I would certainly consider it as long as shipping outside the United
>> States is not going to be a problem. I might even consider more than one
>> if they have a selectable base address. I suspect I would be using them
>> in systems without a FDC so I could easily put a couple in the FF4x
>> range.
>
> Base addressing agility is a good point William.  In fact, that might be
> something we could do preparatory to this.  Unforch I suspect it would
> require trace cuts to do it on the coco's main board, and pasting the newly
> derived signals back to where they should go.  We have a lot of users who
> would not be at all comfortable with that, darnit.
>>
>> Even better if they can assert SLENB or whatever that signal is called
>> to open up the range of base addresses they can use even further. If I
>> have my facts straight, that might even allow for putting them in the
>> FF3x range.
>
> There are 7 usable 4 byte wide slots between the 2 PIA's on the main board,
> and another 7 such slots between the 2nd PIA and the disk controller. I'd
> like to isolate those 2 PIA's and ship the rest of it out to the MPI, but
> with the sloppy decoding in existing packs, we would likely break a few of
> those.  A tiny board, programmable to decode any one or two of those 4 byte
> wide addresses might be a project to consider though, basically fixing the
> charley foxtrots Ft Worth gave us.  But this falls backwards too because if
> we go hanging them here and there willy-nilly, using old ttl logic, we will
> soon be in need of a 6x09 to address bus buffer, at least on the 6809
> machines.  For the 63C09's, mine at least, looks to be capable of driving
> nearly any amount of such loads as it pulls what loads it has to within a
> very few millivolts of the supply rails, doing it nearly as fast as my dual
> trace 100mhz scope can display.  In modern 10 nanosecond CMOS, that would
> be a non-problem today.  Let me cogitate on that.  It could be a case of
> clipping off & removing the existing 76LS138 and using that board pad to
> insert the newer decoding.  If that kills SLENB, then we might be able to
> combine the leftovers and make a new SLENB.
>
> What if's, gotta love 'em.
>
>> --
>> Coco mailing list
>> Coco at maltedmedia.com
>> http://five.pairlist.net/mailman/listinfo/coco
>
>
> Cheers, Gene
>




More information about the Coco mailing list