[Coco] Re: What it is

Gene Heskett gene.heskett at verizon.net
Tue Aug 29 05:38:00 EDT 2006


On Monday 28 August 2006 22:53, Fred D. Provoncha wrote:
>So I found the serial mouse drivers on RTSI. But I have a question: Are
>they compatible with the most-recent NitrOS9 Distro?

Since the one I used at the time is a patch for (IIRC) either cc3io or 
ioman, probably not.  However, since we now have the srcs for all of 
nitros9, putting them back in is probably relatively trivial.  It involves 
registering another src of interrupts in the IRQ polling mechanism, and a 
short routine to grab the 3 or 5 byte packet from the mouse and decode it, 
3 button mice sending a 5 byte packet when they are moved or a button is 
pressed and the decoding logic IIRC needs to know which mouse its dealing 
with.  IIRC the hardware address of the piggy-backed chipset in the rs-232 
was hard-coded into the module at assembly time, something I always 
considered a hack because it was changeable only by dEding that module in 
the bootfile to adjust it, and re-verifying the module.  It ought to be in 
a config file someplace IMO.

Suggestion:  Make a seperate module out of that patch so it can be used in 
any system without the hassles of patching other system modules.  It, if 
properly done, shouldn't add more than 20 bytes to the bootfile due to its 
module header, over the length of a bootfile with the patches applied.

That would greatly simplicate the addition of a serial mouse to the system, 
isolating its code from other stuff it doesn't need to touch at the end of 
the day.

>Specifically what 
>driver files should I use? Do I also need to replace the Clock module in
>order to get a serial mouse to work as the instructions with the drivers
>seem to indicate? Or are the Clock modules that come with NitrOS just
>fine?

They should be. I know of no connection between the seriel mouse patches 
and the clock module(s) as long as the clock modules are ed9 or later.

Because you'll be adding another src of IRQ's to the system, and the shack 
scattered IRQ pullup resistors all over the system like a pepper shacker, 
RELIABLE response to these IRQ's will require that all the pin 8's in the 
mpi's row of sockets be bussed together, and since each pin 8 had its own 
pullup resistor, 3 of the 4 of them should be removed in order to reduce 
the fanout drive required of each source of IRQ's.  Ideally, any pullups 
in the cartridges being plugged in should be removed too, but I only 
removed those that got in my way.

This is one of the mpi fixes thats a given or it will miss interrupts with 
monotonous, agravating regularity.

The problem is that the mpi improperly gated the IRQ's to the system such 
that the mpi had to have that slot enabled before the IRQ could get 
through it.  Big design mistake IMNSHO, the IRQ service routine itself can 
find the src of the IRQ just fine once its allowed thru at any time.  My 
absolute maximum IRQ latency is about 140 u-seconds, with typical response 
times under 20 u-seconds.  NO characters are lost, even at 9600 baud where 
the xon/xoff or rts/cts gets used heavily when a faster src is feeding the 
coco.

>Finally, how do you hook up a serial mouse to the RS232 Pack? 

I actually added a second piggy backed uart chip to my rs-232 pack, and 
brought that out to a db9 connector on the side of the pack, then put that 
same connectors mate on a now extinct logitek DexXa mouse.  It is a great 
mouse when compared to the miss-firing & jittery normal mouse.  I also 
changed the mouses cord exit to the back of the mouse so it wasn't taking 
a beating running it into the mpi all the time.

There are instructions for adding that piggyback 2nd seriel port to the 
rs-232 pack on rtsi.  I used the left button as the usual mouse event, the 
right button for some multiview menu stuff, and the middle button for a 
palette cycler in iconedit.

Since my rs-232 pack's will never operate alone, but always in the mpi, I 
also removed the very hot running little Aztec tin can power convertor 
from it, and ran jumpers up from the card edge pattern to supply it with 
plus and minus 12 volts from the mpi.  I had one of them fail and almost 
start a fire, the pcb under it was charred.  The rs-232 pack will operate 
about 75F cooler without it, so both of mine have now had that mod applied 
and run with no noticeable heat rise.

The other one I have has a different crystal and was used as a midi output 
port with ultimuse, at the same time as the rear panel bit banger serial 
port was being used so I gained additional polyphony by sending different 
midi instruments to different midi keyboards.  Obviously both of my 
keyboards are cheap dynosaurs, a bit like me & nowhere near General Midi 
spec stuff. :(

I hope this doesn't discourage you because once you've used a serial mouse 
on a coco, you will never, ever go back to the agravation of the shacks 
mouse, a horrible hardware hack if there ever was one.

>Fred Provoncha
>Stansbury Park, UT
>
>
>--- http://USFamily.Net/dialup.html - $8.25/mo! --
> http://www.usfamily.net/dsl.html - $19.99/mo! ---

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2006 by Maurice Eugene Heskett, all rights reserved.



More information about the Coco mailing list