Does anyone have a web site address for ordering the 20 MHz version?
Thanks,
- Mike
IIRC they're 14Mhz chips that have been tested and marked as reliable
at 20Mhz.
You might be able to get them from the current manufacturer of the CMD
SuperCPU, Maurice Randall Engineering:
http://www.cmdrkey.com/
Marc Walters
FYI - I'm thinking about getting a TWGS and upgrading it to ~20 MHz using
the new 65816-20. I guess that means I would also need to improve the TTL
chips on the TWGS to something with a < 5 usec response time. Does anyone
know the gate speeds for the various revisions of the TWGS?
I'll also need to find a source for 60-100 MHz crystal replacements.
Thanks,
- Mike
Hope this helps,
Jason Whorton
"Mike Pender" <mpe...@spamsucks.com> wrote in message
news:A_LTa.66772$kI5....@nwrddc02.gnilink.net...
A TWGS won't go that fast *yet* -- I realize that in addition to replacing
the 65816 it would be necessary to replace the gate arrays in order to reach
20 MHz. The FAQ assumes that it would be too difficult to read the contents
of a gate array to duplicate it with another chip at a higher speed--that's
an
awfully big assumption.
You shouldn't believe everything you read in a FAQ.
Hope this helps,
- Mike
You're right; Maurice said the same thing. They buy the 14 MHz part, check
it out to make sure it works to 20 MHz and just start using it. He also
told me that it is rare to get a 65816-14 that can't be clocked to 20 MHz,
but that they usually become unstable by 22 MHz.
Even so, a IIgs running at 20 MHz sounds pretty cool to me.
- Mike
Thank you,
Jason Whorton
"Mike Pender" <mpe...@spamsucks.com> wrote in message
news:%2WTa.69741$kI5....@nwrddc02.gnilink.net...
Thanks, it should be a nice project if I get it working. I think there
might even be a commercial market for a 20 MHz IIgs accelerator board.
The general plan is to study the approaches used by the Zip GSX and the
TranswarpGS and to simplify the design. Aesthetically, the difference
between the two is amazing. The Transwarp looks like a nightmare to
manufacture, with DIP switches and a daughterboard.
In comparison, the GSX is such a clean card that it is practically a work of
art. The circuit traces are laid out in nice organized patterns, it looks
like they minimized the path length for the circuit routing, and the
on-board memory layout is a model of how to conserve space.
As I understand the two design philosophies, the Zip plan was to recreate
the II on the card, whereas the Transwarp approach was to fetch instructions
and data for the cache, process on the card, and to use a type of glue logic
(implemented in gate arrays) to look for memory accesses or instruction
combinations that require the processor to slow down.
If anybody knows where I can find some of the engineering documents for
either of these cards, I would love to study them.
- Mike
M. Pender wrote:
> Thanks, it should be a nice project if I get it working. I think there
> might even be a commercial market for a 20 MHz IIgs accelerator board.
>
> The general plan is to study the approaches used by the Zip GSX and the
> TranswarpGS and to simplify the design. Aesthetically, the difference
> between the two is amazing. The Transwarp looks like a nightmare to
> manufacture, with DIP switches and a daughterboard.
>
> In comparison, the GSX is such a clean card that it is practically a work of
> art. The circuit traces are laid out in nice organized patterns, it looks
> like they minimized the path length for the circuit routing, and the
> on-board memory layout is a model of how to conserve space.
>
> As I understand the two design philosophies, the Zip plan was to recreate
> the II on the card, whereas the Transwarp approach was to fetch instructions
> and data for the cache, process on the card, and to use a type of glue logic
> (implemented in gate arrays) to look for memory accesses or instruction
> combinations that require the processor to slow down.
>
> If anybody knows where I can find some of the engineering documents for
> either of these cards, I would love to study them.
The ZipGS is a caching accelerator as the TransWarpGS is. The major
difference was that it used a custom chip to replace a lot of the
individual ones the TransWarpGS used. The story that circulated
when they stoped production was that another run of this chip was
$10K+ and there was no market for that many cards. At that time at
least the cards just were't selling. Someone bought a ZipGS and posted
that Alltech had said there was only one ZipGS left, it took two months
to sell.
There were a couple of II+,IIe accelerators that were almost computers
on a card, the Titan and even more so the TransWarp(not the
transWarpGS). The TransWarp doesn;t use the II, II+ or IIe processor
or motherboard RAM so you can run with those sockets empty.
For ZipGS or TransWarpGS knowledge Tony Diaz is the first person that
comes to mind. ///SHH Systeme offers ZipGS and TransWarpGS repair so
that's a possibility.
How many cards are in a production run? $10K for 100 cards isn't so great,
but $10K for 1000 cards is a bit more attractive.
> There were a couple of II+,IIe accelerators that were almost computers
> on a card, the Titan and even more so the TransWarp(not the
> transWarpGS). The TransWarp doesn;t use the II, II+ or IIe processor
> or motherboard RAM so you can run with those sockets empty.
I've been thinking about the computer-on-a-card approach since I posted that
note about the Zip GSX and TranswarpGS. At some point it does become less
expensive to incorporate the logic and functions of the motherboard into a
single chip.
It would probably cost less to design a new Apple II-type computer with an
FPGA or ASIC that incorporates the functionality of a 65816-20, two serial
ports, a display driver, an IDE disk driver and an ethernet adapter, than it
would cost to add comparable functionality to the IIgs.
It will be interesting to see how things play out with the new C-one (the
Commodore-64 uber-clone) which uses the same old 65816-14 chips, but
overclocks them to 20 MHz. IIRC the new motherboard design incorporates
some IBM-compatible features, like PS/2 keyboard and mouse connectors. I
haven't heard anything about USB ports, a hard drive or ethernet capability
though.
- Mike
Do you mean $10K just for a production run of the ASIC, not even for a
production run of completed Zip cards?
How many chips are in a production run? $10K for 100 chips is pretty bad,
but $10K for 1000 chips is a bit more attractive.
> There were a couple of II+,IIe accelerators that were almost computers
> on a card, the Titan and even more so the TransWarp(not the
> transWarpGS). The TransWarp doesn;t use the II, II+ or IIe processor
> or motherboard RAM so you can run with those sockets empty.
I've been thinking about the computer-on-a-card approach since I posted that
note about the Zip GSX and TranswarpGS. At some point it does become less
expensive to incorporate the logic and functions of the motherboard into a
single chip.
It would probably cost less to design a new Apple II-type computer with an
FPGA or ASIC that incorporates the functionality of a 65816-20, two serial
ports, a display driver, an IDE disk driver and an ethernet adapter, than it
would cost to add comparable functionality to the IIgs.
It will be interesting to see how things play out with the new C-one (the
Commodore-64 uber-clone) which uses the same old 65816-14 chips, but
overclocks them to 20 MHz. IIRC the new motherboard design incorporates
some IBM-compatible features, like PS/2 keyboard and mouse connectors. I
haven't heard anything about USB ports, a hard drive or ethernet capability
though.
Of course, the C-one has some kind of user-definable coprocessor, which may
make it a more attractive machine for a hobbyist who is willing to forgive
the
lack of amenities (no case, no power supply, no disk drives, etc.)
- Mike
M. Pender wrote:
> Wayne Stewart <way...@telus.dotnet> wrote in message
> news:kd6Ua.16321$46.30...@news2.telusplanet.net...
>
>>The ZipGS is a caching accelerator as the TransWarpGS is. The major
>>difference was that it used a custom chip to replace a lot of the
>>individual ones the TransWarpGS used. The story that circulated
>>when they stoped production was that another run of this chip was
>>$10K+ and there was no market for that many cards. At that time at
>>least the cards just were't selling. Someone bought a ZipGS and posted
>>that Alltech had said there was only one ZipGS left, it took two months
>>to sell.
>
>
> How many cards are in a production run? $10K for 100 cards isn't so great,
I think you missed something. He said $10K+ for one chip, not one card.
Roy
> but $10K for 1000 cards is a bit more attractive.
>
>
>>There were a couple of II+,IIe accelerators that were almost computers
>>on a card, the Titan and even more so the TransWarp(not the
>>transWarpGS). The TransWarp doesn;t use the II, II+ or IIe processor
>>or motherboard RAM so you can run with those sockets empty.
>
>
> I've been thinking about the computer-on-a-card approach since I posted that
> note about the Zip GSX and TranswarpGS. At some point it does become less
> expensive to incorporate the logic and functions of the motherboard into a
> single chip.
It has seemed to me that putting the CPU and RAM all on one card, and
perhaps adding itmes like ethernet/IDE/Video enhancements would be teh
best way to extend the //gs. Alas, I'm so ignorant that dreaming is all
I can achieve in this area.
>
> It would probably cost less to design a new Apple II-type computer with an
> FPGA or ASIC that incorporates the functionality of a 65816-20, two serial
> ports, a display driver, an IDE disk driver and an ethernet adapter, than it
> would cost to add comparable functionality to the IIgs.
Well, having the megachip and the IWM/SWIM wouldn't be cheap!
Roy
>
> It will be interesting to see how things play out with the new C-one (the
> Commodore-64 uber-clone) which uses the same old 65816-14 chips, but
> overclocks them to 20 MHz. IIRC the new motherboard design incorporates
> some IBM-compatible features, like PS/2 keyboard and mouse connectors. I
> haven't heard anything about USB ports, a hard drive or ethernet capability
> though.
>
> - Mike
>
>
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
M. Pender wrote:
> Do you mean $10K just for a production run of the ASIC, not even for a
> production run of completed Zip cards?
>
> How many chips are in a production run? $10K for 100 chips is pretty bad,
> but $10K for 1000 chips is a bit more attractive.
I mean just the ASIC. I recall a few postings on the subject either
here or on Delphi but I didn't pay much attention except the $10k
stuck. A google search might come up with something. Still what
people post isn't always correct and new technology sometimes
changes things dramatically. The cost of a production run might be
considerably less now.
> I've been thinking about the computer-on-a-card approach since I posted that
> note about the Zip GSX and TranswarpGS. At some point it does become less
> expensive to incorporate the logic and functions of the motherboard into a
> single chip.
>
> It would probably cost less to design a new Apple II-type computer with an
> FPGA or ASIC that incorporates the functionality of a 65816-20, two serial
> ports, a display driver, an IDE disk driver and an ethernet adapter, than it
> would cost to add comparable functionality to the IIgs.
>
> It will be interesting to see how things play out with the new C-one (the
> Commodore-64 uber-clone) which uses the same old 65816-14 chips, but
> overclocks them to 20 MHz. IIRC the new motherboard design incorporates
> some IBM-compatible features, like PS/2 keyboard and mouse connectors. I
> haven't heard anything about USB ports, a hard drive or ethernet capability
> though.
>
> Of course, the C-one has some kind of user-definable coprocessor, which may
> make it a more attractive machine for a hobbyist who is willing to forgive
> the
> lack of amenities (no case, no power supply, no disk drives, etc.)
>
> - Mike
A whole new IIgs built anong the C-One has some attractions but
Commodore and Apple II users are a bit different. Comodore users don't
really expect slots and the Apple II users that don't are already using
emulators.
A multifunction card would likely sell better and an acellerator would
be a good card to add other functions to since it only uses the slot
for DMA.
Upgraded video would likely be a worthwhile addition. I suspect in a
dozen or so years it'll be hard finding a working IIgs monitor and with
the move to LCDs there likely won't be too many shops left that can
repair CRTs. A lot more could be added but every time you add a function
the cost and complexity go up.
Wayne
Actually, he said $10K for one production run of the chip, not for one
chip.
There do exist chips that cost over $5K each in production, such as the
largest FPGAs (e.g., Xilinx XC2V6000), but this wasn't one of them.
Typo; I was asking if the production run was just for the ASIC or for
completed cards.
Eric Smith wrote:
> Exegete <mil...@noneofyourbusiness.com> writes:
>
>>I think you missed something. He said $10K+ for one chip, not one card.
>>
>>Roy
>
>
> Actually, he said $10K for one production run of the chip, not for one
> chip.
Since BOTH said "production run" I saw no need to repeat what was
understood. The misunderstanding was chip vs card.
Roy
>
> There do exist chips that cost over $5K each in production, such as the
> largest FPGAs (e.g., Xilinx XC2V6000), but this wasn't one of them.
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
The C-one has two slots, one of which could conceivably take a GS
compatibility card to support smartport and printer/modem ports. I think
the least painful approach would be to configure the FPGA to map GS I/O
addresses to hardware and make a separate card that adds generic USB
peripheral compatibility to the C-one.
That said, I haven't seen anyone offering IDE hard drives, CDROMS, ethernet
cards, or 1.44 meg 3.5" disk drives that support the C-one yet (emphasis on
the word "yet"). I expect that such things *must* be in the works, or the
C-one will have a very short market window.
- Mike
If power is the problem, an external power brick
could be added (like the never released Voodoo6
from the late 3dfx).
Now if you are talking add-on hardware for C-One,
I would much rather see a IIGS card for the PC
because today's software IIGS emulators just do
not cut it!
I think that's a great idea--although I think that some of the emulators for
standard Apple II machines are excellent, such as the Apple Oasis emulator.
Have you thought about the "essential features" list?
Or a PowerPC based card, since the best iigs emulator is already written
on a PowerPC, it should be easy to merely emulate the 65816 by a G3/4/5.
Roy
>Supertimer <super...@aol.com> wrote:
>>
>> Now if you are talking add-on hardware for C-One,
>> I would much rather see a IIGS card for the PC
>> because today's software IIGS emulators just do
>> not cut it!
>
>I think that's a great idea--although I think that some of the emulators for
>standard Apple II machines are excellent, such as the Apple Oasis emulator.
>Have you thought about the "essential features" list?
Ensoniq chip emulation.
Full IIGS graphics mode emulation including 3200 mode.
Ability to run all IIGS demos and software including the
offerings from FTA and HyperCard/HyperStudio GS.
Ability to run GS/OS without crashing or glitches like
XGS does all the time.
Ability to transparently run GS/OS software including
AppleWorks GS 1.1 with Pointless and a full library
of TrueType fonts and print using the LaserWriter
driver to any connected PC printer to the full printer
resolution (LaserWriter driver outputs PostScript).
Essentially, make a viable IIGS replacement which
emulators like XGS are not.
I know that there are close to viable replacements for
the 8-bit Apple IIs. In fact SimSystem IIe was quite
good for a IIe replacement. But IIGS replacement the
software emulators are not.
I still think that the best processor to use would be a TransMeta Crusoe. It uses
about 4 watts total power at maximum speed. It has an SDRam controller builtin as
well as a PCI interface for a possible 33Mhz secondary expansion bus. And it was
designed to emulate other (mainstream) processors. All I don't know at this point is
unit cost in small lots :/
Take a peek at their 5800 or 8800 processors :)
Later, tmyers
MP> It will be interesting to see how things play out with the new
MP> C-one (the Commodore-64 uber-clone) which uses the same old
MP> 65816-14 chips, but overclocks them to 20 MHz.
As I wrote earlier, Apple II fans should stay tuned.
MP> IIRC the new motherboard design incorporates some IBM-compatible
MP> features, like PS/2 keyboard and mouse connectors.
It has PS/2 keyboard and mouse, 2x atari joystick connectors, VGA
output, stereo sound, and a parallel port. Internally it has a floppy
connector, dual IDE connectors (with one CF slot), a PCI slot (with an
option of a 2nd one), a C64 cartridge slot, a CPU slot, and two A1200-
style clock ports.
MP> I haven't heard anything about USB ports, a hard drive or ethernet
MP> capability though.
Dual IDE is standard, a clock port ethernet adapter was released this
weekend, and a USB adapter is in development.
--
___ . . . . . + . . o
_|___|_ + . + . + . Per Olofsson, arkadspelare
o-o . . . o + Mage...@cling.gu.se
- + + . http://www.cling.gu.se/~cl3polof/
MP> That said, I haven't seen anyone offering IDE hard drives, CDROMS,
MP> ethernet cards, or 1.44 meg 3.5" disk drives that support the
MP> C-one yet (emphasis on the word "yet"). I expect that such things
MP> *must* be in the works, or the C-one will have a very short market
MP> window.
All of those are either standard components on the motherboard (IDE &
floppy) or available as an expansion card (RR-net, EUR 50 at
www.jschoenfeld.com). Booting from HD and CD-ROM works in the current
rom, floppy is in development. RR-net is supported in Contiki, and
should run out of the box in C64 mode.
Gee, that would be nice to run at on a 65C02 in a //e.
JaY