9 Mar
2011
9 Mar
'11
3:30 p.m.
On 9 Mar 2011 at 12:42, Eric Smith wrote:
People talk about "the entire chipset", but
there wasn't such a thing.
That would be similar to talking about a microcoded bitslice system
using "the entire 2900 chipset"; the concept is meaningless.
As I recall, we asked what the manufacturing cost of a multibus card
using the 432 would be, as our own bus was pretty close to Multibus.
After some hem-ing and haw-ing Intel estimated that $1K would be a
good ballpark. Mind you, this was before the iA432 was in
production, so Intel didn't have any clue on what the yield would be.
I don't think many of the PACE systems used
National's special
interface chips. It's not clear that those chips even went into
production.
Oh, they did, but marketed as parts for the National 8080 support
program. The MILE, for example, is the DP8301; the bidirectional
transceiver (I can't recall the acronym) was the DP8308. Note the
lack of focus on the part of National--were they interested in
selling the SC/MP, the PACE or the 8080?
National had a couple of their scouts over to do a presentation on
the NS32016 (or 16032, or whatever they called it initially). The
thoughtfulness that went into the design really blew me away--the
chip looked far and away better than any contender.
Of course, it all came down to "When can you start sampling?". A
look exchanged and a tentative "Well, we think we'll have silicon in
11 months". As I recall, it turned out to be about 2 years...
At least it wasn't boring. I did take my free sample PACE and put it
onto an S100 board for jollies, since nobody at work was interested
in pursuing it. The design issued pairs of reads and writes to
accommodate that 8 bit S100 bus, the thing still ran like molasses in
January. I lost interest pretty quickly.
--Chuck