[Coco] [Color Computer] The evolution of the Coco..

j_e_daggett j_e_daggett at yahoo.com
Fri Apr 22 16:34:59 EDT 2005

--- In ColorComputer at yahoogroups.com, Mike Pepe <lamune at d...> wrote:
> I'd like to see a faster CoCo in the future. The stuff Cloud9 is doing 
> is impressive but I think it might be in the wrong direction.

Right now the easiest and cheapest way to get a faster Coco is to use
a HD63C09E. Using a FPGA and configuration prom is in the $40 to $60
range just for the silicon chips. Atmel and lattice do make flash
FPGAs. The problem is that the MC6809 cores that I have seen so far
fit in a Spartan 2 200K gate FPGA minimum. 
> The "C-64 on a chip" is probably the better direction to go.


That right now is a pipe dream unless you can find a million dollars
for the research and development cost to do the die. The whole MPU,
GIME chip, and the PIAs and the salt chips can all be done on one big
die. But be ready to pay for it. Even with 0.13 micron CMOS mixmode
process you are most likely looking at a die with about 100 to 200 mm
suare area. T hat is not going to be cheap. The better way to go is to
do three separate die. Then eutectic bump the die onto a thin film
flex circuit and the form one big multichip module that contains most
of the circuitry. This will still take between 2 and 5 miliion dollars
to get production up and running. To make the whole thing worth it
will need sales of at least 100,000 units minimum. The only way the
C64 came about as a product was that someone started it in an FPGA and
some company bought out the design and hired the engineer. 

> Emulating the CoCo on some CPLD devices releases us from reliance on
> old chips and hardware which- while not rare- are no longer in

Yes we can do most of the chips in either a CPLD or a small FPGA.
Besides through Rochester Semiconducter you can still purchase
"obsolete" Motorola chips. The problem is  you may have to buy a wafer
at a time and that can be as many as 300 parts minimum. Motorola sold
 all its "obsolete" designs to Rochester. As well as many other IC
makers have done with their obsolete ICs. 
> In addition all these plastic packaged chips in our CoCos are slowly 
> degrading. Not sure how long it takes but they're not hermetically 
> sealed and will eventually oxidize.


The normal life expectancy of a chip in everyday use is up to 20
years. Beisdes if the passivation layer over the die has not been
damaged during packaging then there is little to worry about moisture
and humidity. Injection mold are not hermetically seal packaging but
do resist humidty quite well. All die have a thin layer of nitride
pasivation to protect the metal runners. The main killer to teh metal
runners is salt water. Sodium Hydroxide will play packman on the
aluminum ru nners on the die. The failure rate for injection molded
plastic packaging due to salt water intrussion is extremely low. Again
if th enitride passivation layer is not cracked or damaged then it
becomes extrememly  rare. 

I know I  have done IC failure analysis on defective ICs that have had
the nitride passivation cracked and in high humidities. The ICs failed
in products that were used in Okinawa Japan. Out of 50,000 units there
I only saw 170 total failures. But you can bet that NTT,  Nippon
Telephone and Telegraph, was pissed that we had that many.


Brought to you by the 6809, the 6803 and their cousins! 
Yahoo! Groups Links

<*> To visit your group on the web, go to:

<*> To unsubscribe from this group, send an email to:
    ColorComputer-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:

More information about the Coco mailing list