Re: AVG chip replacement

>All hail Clay! All hail Clay! The hacker supreme! (with some well
>deserved credit to the birdie as well)

The "birdie" gets *most* of the credit. I just did the grunt-work on it. :-)

>>The only gotcha is that there are a lot of chips in the design, and thus
>a
>>lot of PCB real-estate, and thus a higher cost per board. :-(
>
>Just how big is the daughterboard that would plug into the socket?

I'm still tinkering around with it, but it looks like around 3.5x3.5".
It's not *huge*, but I need to look at all the vector games to see what the
best arrangement/size is to fit on all the original boards. Basically it's
just a PCB with a 40 pin DIP header that plugs in where the AVG chip goes.
Picture the little "daughtercard" from a Pac-Man (the 285 board I think?)
if you've seen one. About like that.

>I would be more interested in the bare boards as I recently stripped a
>mess of 74LS TTL chips from some junk boards but I concede Clay deserves
>something for cooking up the solution.

A couple of people have said that. Sounds good to me-- less work. :-)
It'll shave a few bucks too. (Probably $4-$5)

>BTW How much would a PLD solution run for the same circuit? I remember
>you saying the PLD would be more than the TTL method but how much?

I looked at the PLD thing and there are a few issues:

        1) The circuit is mostly 4 bit counters. Like 5 of 'em. Takes a lot of
           PLD. (We're getting into the ~$10 chip price range for a single part
           and...

        2) The other big chunk of the circuit is RAM. Not fun in PLD's. I
           don't know how much a 4-word 16-bit hardware stack would occupy...

        3) There's quite a bit of I/O, so you're looking at fairly big PLD's
           just for the pin count. You could probably find a PLD that's kinda
           cheap and has enough logic to do the job, but they tend not to have
           enough pins...

I tried splitting this up, (like PLD's for everything but the register
files), but TTL is so cheap (<$.33 a chip) you can buy *all* the chips for
the price of one 22V10 or something. The PLD design would have the
advantage of board size, which saves $$$, but that's at the expense of
"time to market". Most of the price of the boards is in setup charges and
whatnot anyway right now, square inch-wise it isn't too bad.

I'm inclined to do the TTL thing, let someone else whip it up in VHDL, or
ABEL or something and then target it to a cheap PLD. It's definately
possible to do cheaper/better, but I didn't want to distract myself long
enough to try to pull it off. I want to finish the G-80 multi-game ASAP...

>That and he probably has a TTL supply that makes my junkbox look like the
>IC selection at Radio Shack.
>
>Virtu-Al

Clayton N. Cowgill Engineering Manager
_______________________________________________________________________
/\ Diamond Multimedia Systems, Inc. clay@supra.com
\/ Communications Division http://www.supra.com/
Received on Sun Jul 13 12:06:57 1997