[Coco] Multi-pak interface and CART*

[Gene Heskett]

On Saturday 05 July 2008, Chuck Youse wrote:
>Does CART* only work for the selected slot?  I.e., if in a Coco 3, I
>wanted to be able to generate an interrupt (either IRQ* or FIRQ*,
>depending upon the GIME settings), is it possible?  Are there any
>limitations (e.g., the slot must be selected)?

In a stock MPI, yes.  One of its 2 major design flaws IMO.

>I don't have an MPI, but I'd like for this thing to work for folks who
>have 'em, which is probably most ..

Generally speaking Chuck, if the usual fix for lost IRQ's has been done to the 
MPI, which consists of jumpering all 4 card sockets pin 8's together, and 
removing 3 of the 4 IRQ pullup resistors along the front of the pcb, then the 
CART* problem is a total non-issue, at least for os9/nitros9.  By adding the 
jumpering, an IRQ from any slot gets through regardless of the state of $FF7F 
or the slide switch.

You were concerned with IRQ response times in an earlier msg.  While I was 
working on rzsz a decade+ back, I had rigged my scope up to see what those 
times were, and generally, with a jumpered MPI, any interrupt generated by the 
6551 was at least serviced with a data read to clear the IRQ in around 16 
microseconds, with an occasionally long lag (I think it was writing to disk at 
the time, or possibly scrolling a screen) of about 150 microseconds, so if you 
can handle a 320 u-s delay, I'd say you were pretty much home free.  I never 
saw a response any longer than that.

-- 
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)