Re: TWGS vs ZipGS (was Re: WOLF3D fot A2 GS)

[spec@vax2.concordia.ca (Mitchell Spector)]

In article supertimer@aol.com (Supertimer) writes...

>spec@vax2.concordia.ca (Mitchell Spector) wrote:
>>    I was just simply stating a fact. There ARE boards out there
>>with the original ROM (v1.4) and cannot be upgraded unless you
>>replace the daughterboard. There is no such thing as old Zip GS
>>that cannot be upgraded, so this definitely applies as a con.
>
>That's not true, Mitch....there were pre version 1.1 Zip boards
>that conflicted with RamFASTs and other cards.  Zip offered
>an upgrade for these earlier boards, but some are still out there.
>They are as rare as the 6Mhz ROM v1.4 TWGS you have.

    ZipGS 1.1 boards? The highest revision made was the 1.02 board.
I'm well aware of the early revisions that did not properly support
DMA--specifically hardware revisions 1.00 and 1.01--but that was
the only flaw. You could _still_ upgrade their cache and clock
speed and use them like any other ZipGS board.

    The TransWarp GS (all revisions) are more similar to the early
revisions of the ZipGS, in that they do not properly support DMA.
However, like the older ZipGSes, they too can have their cache and
clock speed upgraded (except when you have a ROM 1.4 daughterboard).

>>    My own TWGS has the original ROM, making my board virtually
>>worthless to me except for parts. The only way to get a new ROM
>>is to order a complete daughterboard replacement, and that costs
>>more than a complete TWGS. In fact, I discovered it is cheaper
>>to just purchase a new TWGS than ordering replacement parts and
>>that is what I'm about to do.
>
>The fact that you have a rare crippled version of the TWGS has
>biased you against the card.  Most TWGS cards are not like
>this just like most Zip GS cards are not the crippled v1.0 model.

    Well the fact is, if I had a ZipGS v1.00, I could be using
it right now and have the option of upgrading its cache/speed.
Mind you, I doubt there are many 1.00/1.01 boards out there,
as Zip Technologies/REMcorp offered a free upgrade to the 1.02
board if you mailed it in. That was probably valid up until
they discontinued it recently.

>TWGS works fine with my DMA Apple HS SCSI card and I got
>an email from a RamFAST user who says it can DMA to RamFAST
>too.  Must be the fact you have a rare crippled version of the
>RamFAST.  You know that ZipGS v1.0 did not properly support
>DMA either.

    Let's see, I have a rare crippled version of the RamFAST,
a rare crippled version of the TransWarp GS, and apparently,
a rare crippled version of AppleWorks GS. Heh, I must be a
very unluckly person. ;-) Seriously though, all the RamFAST
boards I own are rev.D with the 3.01e EPROM installed.

    There is an option that you must set to 'Yes' if you have
a TWGS installed (no such option is present for a ZipGS) or
it will not work at all. If you still have problems, it is
recommended you disable DMA. Incidentally, speaking of things
not working at all, if you don't have the -2B GAL in your TWGS,
it won't fucntion *at all* with the RamFAST or Apple HS SCSI.
That is yet another drawback to the card, especially since
there is no longer any place to order this GAL chip.

>You have a rare crippled version of the TWGS.  My TWGS
>works fine with ProTERM.  On the other hand, I distinctly
>recall reading about v1.0 Zip GS having problems running 8-bit
>software.  InCider/A+ wrote that the fix was to solder a wire
>bridging the two solder pads diagnally to the corners of the board.

    Well that's good to know, I was quite annoyed it wouldn't
function with ProTERM 3.1--the program that gets used the most
on the IIgs I plan to install the ROM version 1.5 TWGS in.

>>    Once again, I am just stating a fact. The ZipGS does not
>>need chips replaced depending upon the manufacturing date of
>>the 65C816 plugged in. All my 65C816's--with the exception
>>of the -14 rated one in my ZipGS--are all pre-1989 chips as
>>it happens.
>
>This is also probably the reason why the TWGS is a bit faster
>per clock speed....the ZipGS has logic to slow down the card
>at more occasions because it assumes the older 65C816.  The
>TWGS took a chance and maximized performance.  Those who
>wanted to go to higher clock speeds with existing CPUs could
>swap for GALs that would slow down the card on more occasions,
>similar to what the ZipGS does.  Now with the newer 65C816,
>the ZipGS is locked into the old logic.  With the TWGS, we
>just don't get the replacement "slowdown" GALs that were
>once needed....and enjoy slightly greater speed with the better
>CPU chips.  On the ZipGS, it does not matter that the CPU
>chips have gotten better....the Zip is stuck slowing down the
>CPU as though it were the old ones.....

    Actually I don't think this is the reason. It has more to
do with the way the TransWarp GS caches frequently used data.
It works in a more efficient way than the ZipGS in terms of
caching CPU instructions. I can't remember the exact specifics,
but it has little or nothing to do with slowing down the
65C816.

Mitchell Spector
spec@vax2.concordia.ca