[Coco] Multi-Pak card edge connectors.

[Gene Heskett]

On Thursday 14 January 2010, Brian Blake wrote:
>Hi Phil,
>
>The link is below. I'm in the process of laying out a PCB in Express PCB
> (too lazy to wire-wrap...). I'll be putting it up on my blog and coco3.com
> when I'm done for everyone to review. I'm on my 3rd revision right now, so
> I can only imagine what the experienced guys here are going to say, lol...
>
>http://www.coco3.com/community/2009/10/color-computer-expansion-interface-c
>ustom-multi-pak-interface/

The first thing I note is that Tim assumes the U20 inputs will adequately 
pull all the cart# signals high in the event there is no cartridge plugged 
into a given slot.  This is hanging out a huge billboard sign inviting noise 
problems in.

The 2nd thing is that for use with os9, he didn't bridge all the cartridge 
connectors together at pin 8, effectively bypassing the logic that can block 
the rs-232 packs IRQ assertions, which in turn leads to lost characters in 
receiving data.  This particular fix is one of many that purports to fix 
this, but if you analyze the logic, its the only method that makes sense.  An 
IRQ assertion must be able to get through to the cpu, and be serviced in a 
timely manner if lost characters are to be avoided.  When this is done, it is 
common practice & recommended that 3 of the 4 pullup resistors the multipak 
has in it, pulling each cart line up by itself, be removed, one pullup is 
more than sufficient to achieve the goals of a solid logic one when no 
cartridge is asserting its IRQ line.  IN the real MPI, leaving all 4 of them 
in place when all 45 slots are bridged, also makes for a heavy job for the 
individual cartridges IRQ driver when it does pull the IRQ down due to the 4x 
as much loading when the pin  8's are all bridged together.

For all of the above reasons, there should be a 10k resistor connected 
between the 5 volt line and one of the cart lines, and all 4 cart lines 
should be connected together.  Very easily done if you know which end of a 
soldering iron gets hot..

Under os9, the IRQ service routiine has no problem finding the real source of 
an IRQ, usually in less than 15 microseconds, so there is no obvious 
advantage to copying the MPI's original treatment, and several disadvantages 
if it is duplicated.

I have not noted any problems from doing this mod when running rsdos in the 
limited time I have run it.  That doesn't mean there might not be any because 
I don't normally run DECB, I boot directly into Nitros9 here.

Also, using the S family of chips makes it very power hungry, and I don't 
believe there is anything in this whole thing that cannot also be done with 
the HCT chip family (<1% the power needed compared to S) unless that 
particular function is not available in HCT, in which case I'd use the next 
lowest power, probably the LS stuff.

>afvg
>
>
>----- Original Message ----
>From: Phill Harvey-Smith <afra at aurigae.demon.co.uk>
>To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com>
>Sent: Thu, January 14, 2010 6:27:12 PM
>Subject: Re: [Coco] Multi-Pak card edge connectors.
>
>Tim Fadden wrote:
>> On 1/14/2010 10:52 AM, Brian Blake wrote:
>>> As I'm looking at building Tom Gunnison's MPI clone, I was looking at
>
>You don't happen to have a link for that do you ?
>
>Cheers.
>
>Phill.
>
>-- Phill Harvey-Smith, Programmer, Hardware hacker, and general eccentric !
>
>"You can twist perceptions, but reality won't budge" -- Rush.
>
>--
>Coco mailing list
>Coco at maltedmedia.com
>http://five.pairlist.net/mailman/listinfo/coco
>
>
>
>
>
>--
>Coco mailing list
>Coco at maltedmedia.com
>http://five.pairlist.net/mailman/listinfo/coco
>


-- 
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)

If I wanted Windows, I'd live in a greenhouse!