Re: TWGS Upgrade Availability?

[Rubywand <rubywand@swbell.net>]

Ecitnerppa writes ...
> 
> I may be mistaken, but I thought that the 2B gal did the "slowdown" only on a
> specific 65186 instruction (or two), and that those instructions were of a
> "mode change" type, IE, not likely to be used often, or not used in
> heavy-computation loops.
> 
> I'm probably badly paraphrasing the specifics, but perhaps the gist of it
> survived. Anybody done some benchmark testing of the two chips?


     A check of the TWGS upgrade articles John Link did for Appleworks
Forum indicates that the GAL which took care of care of extending
instruction execution time is the E version of GAL 3.

     John does not mention any benchmarks comparing TWGS boards with or
without the 3E GAL-- probably because the overall affect is very small.
As you say, only a few relatively low frequency instructions were
involved. Anyway, he mentions no notable loss in speed when the 3E GAL
is used.

     One source of confusion about the need to swap in this or that GAL
is that some were just too slow when attempting speed upgrades to 10MHz
with the older 65C816's. The only GAL which John mentions as doing
instruction stretching is 3E.


     The good news in all this is that things changed when the new
65C816's became available. In an October, 1992 article, John Link notes
that his TWGS with the standard 32K Cache upgrade could be speeded up to
12.5MHz by swapping in the new 65C816 and a 50MHz clock oscillator
module.

     He says he obtained 12.5MHz without swapping in the 3E GAL or any
of the other hi-speed GALs mentioned in earlier articles. Also, he did
not need to increase board voltage above the standard 5V.

     For a step-by-step guide, download Scott G's TWGSupgrade.SHK.
HyperCard stack at ...

ftp://ground.ecn.uiowa.edu/apple2/apple16/Hypermedia/Hypercard/ .

     You can also find TWGS upgrade information in the FAQs Accelerator
file Csa2ACCEL.TXT at ...

ftp://ground.ecn.uiowa.edu/2/apple2/Faqs .


 
Rubywand