Byron A Jeff wrote:
>On Thu, Jun 08, 2000 at 09:19:21PM +0300, Peter L. Peres wrote:
>> Hi,
>>
>> which and how many medium size PALs replace a $0.6 [(CMOS)|(74C)]4040
>> address counter, do not require a programmer, and are second sourced ?
>

>Actually that isn't a fair comparison. The real objective of the PAL/GAL is
>to create a multibyte programmable counter. That way with a 13/14 bit interface
>You can interface with up to 16 MB of RAM, with a 3 instruction sequential
>cycle read time and up to 8 instruction or so random access time. With the 4040
>while the sequential time isn't bad, there's simply no effective way to do
>random access because it would require clearing the counter then clocking the
>counter to the desired address.
>

Yes, this was the point of my starting this thread. The '4040 provides
sequential addressing only, cannot be used for random access or for
addressing multiple data buffers, and can't provide anywhere near 20Mhz
accesses, since it is a ripple counter. With the 4 '161 counter approach
that I am wishing to streamline, all 16 counter bits change state
simultaneously.
=============

..........
>Personally I think that such an interface could be done with a PIC16F877
>using the PSP port. Essentially using a second PIC as the RAM address
>interface. However I'm not sure if the speed of the interface would be
>accptable though.
>

I am now developing another app where a single 16C74 takes care of 16
address and 8 data lines. It acts as a slave to another uC. The '74 has
just enough pins to do it all, inc a few for I/F to the host. My [very
rough] estimate is max read/write rate is only about 500Khz. I should
know more accurately in a few days.

regards,
- Dan Michaels
Oricom Technologies
http://www.sni.net/~oricom
==========================