> [...] He had a different view of the 4052. To improve performance,
the
designers use
an AMD bit-slice processor. It had to emulate the 6800
so the existing software base could be used. Larry said this product
was a disaster, it didn't run much faster. Sales were poor so he had
to cancel the product.
Interesting because all press reports are to the contrary. I haven't
done any
benchmarking between 4051 and 4052, but the 4052 is a
significantly
more
complex piece of equipment with the bitslice
implementation of the
6800.
The 4052 was in fact significantly faster than the 4051. I have both a
4051 and a 4052, and the 4052 is much faster than the 4051 in more ways
than just raw CPU power.
I argue what Mayhew stated as to his reasons for cancelling the series
of machines. Perhaps he was speaking from a sales revenue "performance"
perspective (typical of many upper-management types). Revenues for
the 4051 and the later 4052/4054 fell consistently after the 4052/4054
were introduced more because of the outdated display technology used.
The death of the 405x series came from its use of DVST (Direct View
Storage Tube) display technology rather than performance. Raster
graphics systems were gaining ground very quickly in the latter part of
the 1970's, rendering DVST rather pointless in all but some very fringe
environments where very high resolution was required. Even stodgy
Tektronix realized this, and began designing a series of raster graphics
terminals to replace its DVST terminals. Unfortunately, no such effort
was made to take the 4051/4052/4054 concept and implement it using a
nice raster display subsystem. Such a product could have been a major
hit.
The 4052/4054 used a bit-slice (AMD 2901) 16-bit implementation of the
Motorola 6800 instruction set, and added microcoded floating point math
instructions (utilizing unimplemented opcodes in the 6800 instruction
set), along with major improvements in the code that implemented the
BASIC programming environment on the machine to improve performance with
math, graphics, and I/O. The 4054 added another graphics processor
and a 19" DVST tube to allow even faster graphics, as well as some
limited refreshed vector graphic capability (though the refreshed
vectors were rather dim to prevent the storage tube from storing the
image).
Tek made quite a number of blunders in numerous attempts to get into
the computer marketplace, all mostly a result of rather clueless
management that didn't know how to market anything by Test & Measurement
equipment...and when they did have good computer products, they tried to
shoehorn them into the T&M business by turning them into instrumentation
controllers, eliminating any chance of them succeeding in the computer
marketplace.
No bitterness here...all companies make mistakes. Working for Tek was
one of the best employment experiences of my lifetime. I worked for Tek
between 1977 and 1990, and saw many really great computer products come
and go, ranging from the amazing Magnolia Smalltalk Workstation from
1976-ish, to the 405x "personal computers", to some pretty powerful
desktop Unix graphic workstations in the mid-to-late 1980's.
Rick Bensene