Tms9900 based moog: the SL8

[Elia Spallanzani fdt]

"One of the really cool attributes of the TMS 9900 MCU was it had a considerably faster 3.3 MHz clock than the 8051/8031 MCU used in the Roland and the 8085 in the Korg units at the time. It ran rings around a Z80 because of the 16 bit data path and the unique ability of the 9900 to store all user registers in RAM and update the "RAM Registers" in a single instruction cycle. If you called a subroutine or a NMI call was true, you could update the user register directly. This was how all of the parameters were stored/updated once a program or configuration change was made. It also made complex multi tasked threading for the housekeeping, DCO timing, triggering, MIDI, and controller interaction much simpler to code.

 

http://moogarchives.com/forum/viewtopic.php?t=9

Edited March 10, 2022 by Elia Spallanzani fdt

[+mizapf]

Which brings me to the question - did the author actually measure this performance bonus of the 9900, or just believe it?

Comparing clocks between different architectures does not make sense, and our glorious 16-bit data path brought us read-before-write.

[Elia Spallanzani fdt]

Bit of a mistery here:

"The circuit guts and operating system source code disappeared at some point before Moog's liquidation, and no recordings of the SL-8 are known to survive"