Building C64 custom chips
Fabrizio Gennari
Frequently Answered Question :)
Let's say I want to build a new 6510 chip, or 6526, or 6581 or whatever
custom chip produced by MOS/CSG. I'm not referring to the design of a new
chip, compatible with the old ones (like Jeri Ellsworth's brilliant
CommodoreOne or the re-implementation of a CIA in VHDL at
http://www.enteract.com/~watson/Watson6526/), but to getting the
original schematics and building chips from them.
After some investigations, there are some questions still open:
1. After Commodore's demise in 1994, Commodore assets were bought by Escom,
but some went to GMT Microelectronics, a company founded by some former CSG
employees. Did the C64 chip schematics go to Escom or GMT?
2. After Escom's demise, Escom property related to Comodore went to Gateway,
which probably sold it to Amiga Inc. Do Amiga Inc. own the right to the 6510
schematics?
3. GMT Microelectronics website is no more. Is the company still running?
And does anybody know how to contact its (ex?) employees?
From an old message in comp.sys.amiga.misc
(http://groups.google.it/groups?q=g:thl4266864368d&hl=it&selm=6hp0gj%24c4%40
vu-vlsi.ee.vill.edu) it seems that, for some Amiga chips, schematics in
eletronic form exist. Does this mean that they also exist for C64 chips?
Too many unanswered questions here. Let's hope somebody has at least some
answers.
Fabrizio Gennari
Frequently Answered Question :)
Let's say I want to build a new 6510 chip, or 6526, or 6581 or whatever
custom chip produced by MOS/CSG. I'm not referring to the design of a new
chip, compatible with the old ones (like Jeri Ellsworth's brilliant
CommodoreOne or the re-implementation of a CIA in VHDL), but to getting the
original schematics and building chips from them.
After some investigations, there are some questions still open:
1. After Commodore's demise in 1994, Commodore assets were bought by Escom,
but some went to GMT Microelectronics, a company founded by some former CSG
employees. Did the C64 chip schematics go to Escom or GMT?
2. After Escom's demise, Escom property related to Comodore went to Gateway,
which probably sold it to Amiga Inc. Do Amiga Inc. own the right to the 6510
schematics?
3. GMT Microelectronics website is no more. Is the company still running?
And does anybody know how to contact its (ex?) employees?
From an old message in comp.sys.amiga.misc
(http://groups.google.it/groups?q=g:thl4266864368d&hl=it&selm=6hp0gj%24c4%40
vu-vlsi.ee.vill.edu) it seems that, or some Amiga chips, schematics in
Axel
Fabrizio Gennari wrote:
>
> From an old message in comp.sys.amiga.misc
> (http://groups.google.it/groups?q=g:thl4266864368d&hl=it&selm=6hp0gj%24c4%40
> vu-vlsi.ee.vill.edu) it seems that, for some Amiga chips, schematics in
> eletronic form exist. Does this mean that they also exist for C64 chips?
>
> Too many unanswered questions here. Let's hope somebody has at least some
> answers.
They likely exist in some form, but you are going to need money and do a lot of
tracing.
As a start you would have to do patent search on most of the 8 bit CSG chips in
various countries to determine who may have held ownership. Then you would have
to trace through the various bankruptcies, receivership, etc documents to see
who purchased the rights, and physical items. Keep in mind that a physical item
likely was separated from the related intangible rights so you will have to
trace both.
These items are likely to be locked in some dark dungeon and abandoned, you will
need money to make it worthwile for the current owners to search and dig them
out.
Good Luck.
Riccardo Rubini
"Fabrizio Gennari" <fabri...@tiscalinet.noitspam> ha scritto nel
messaggio news:9sjk8q$4ce$1...@pegasus.tiscalinet.it...
> vu-vlsi.ee.vill.edu) it seems that, or some Amiga chips, schematics in
> eletronic form exist. Does this mean that they also exist for C64 chips?
>
Ciao Fabrizio,
puoi trovare un sacco di schemi presso questo sito :
http://highgate.comm.sfu.ca/~rcini/classiccmp/my_docs.htm Sono in formato
PDF e...Come vorresti fare a ricreare gli integrati da quegli schemi?
(translation : you can find a lot of resources here :
http://highgate.comm.sfu.ca/~rcini/classiccmp/. Stuff is in PDF format. Do
you have any ideas for rebuilding them from those schematics? )
--
Riccardo Rubini
------
togliere xxx & yyy dall'indirizzo per rispondere
Please remove xxx & yyy from mail address to reply
Andrew Bosch
news:9sjk8q$4ce$1...@pegasus.tiscalinet.it...
> 1. After Commodore's demise in 1994, Commodore assets were bought by
Escom,
> but some went to GMT Microelectronics, a company founded by some former
CSG
> employees. Did the C64 chip schematics go to Escom or GMT?
Escom got the designs which were subsequently sold to Gateway and then to
Amino Corp (now called Amiga Inc.). You can read all the ugly details at
http://www.cucug.org/sr/.
According to http://amiga.emugaming.com/commodore.html, the Commodore name
and trademark is owned by Tulip Computers NV and recently licensed to
another hardware company called Computers National Inc.
(http://www.commodore2000.com) who is evidently planning (or has planned) to
release a low-end computer with that name.
> 2. After Escom's demise, Escom property related to Comodore went to
Gateway,
> which probably sold it to Amiga Inc. Do Amiga Inc. own the right to the
6510
> schematics?
No. The design of the 6502, which is the core of the derivatives like the
6510, is still owned by the Western Design Center
(http://www.westerndesigncenter.com). You can get the 6510 from Sierra
Circuit Design, http://www.sierracircuitdesign.com, though.
You can still find the datasheets for the custom chips at
http://www.6502.org or even in a good C64 programmer's guide.
> 3. GMT Microelectronics website is no more. Is the company still running?
> And does anybody know how to contact its (ex?) employees?
GMT Microelectronics is still alive and kicking, although they don't have a
domain name anymore. I found numerous references to them at Google. Their
postal address is:
GMT Microelectronics Inc
Valley Forge Corporate Center
950 Rittenhouse Road
Norristown PA 19403
Ph: 610-666-7950
Fax: 610-666-2500
> eletronic form exist. Does this mean that they also exist for C64 chips?
Since the sale of Commodore to (eventually) Amiga included all intellectual
property, this would include all chip designs, including those for the C64.
It does not guarantee that those designs have not been destroyed by Amiga to
make way for the new models.
Andrew
Burt
documents and the information?
To think that one could purchase new SID chips ah...
--
___
/ __|__ Burt /Terminator / /
/ / |_/ www.museum.c64.org / /
\ \__|_\ Adoring C= 64 / 128 and \ \/ /
\___| Getting into Amiga, too... \/\/
Rainer Buchty
Burt <c64du...@NOSPAMyahoo.com> writes:
|> this sounds so pessimistic? Isn't theres some positiveness in finding those
|> documents and the information?
|> To think that one could purchase new SID chips ah...
Wouldn't happen anyway. Why?
(1) The Commodore custom chips were designed having a (from today's standards)
absolutely crude technology (1um) in mind. You won't find a fab who is capable
of fabricating in that. Now you could say, ok, just do it in 0.35um (not to
think of 0.13um) but that will kill the original design since delay lines
will work faster than expected destroying/disturbing the internal timing.
(2) The Commodore custom chips were designed on resistor/transistor level
(RTL), so you are not able to directly map this design to any modern ASIC
technology or even cheap FPGAs. Reworking the schems to a high-level
description (VHDL, Verilog) would be probably more work than just doing a
hardware emulation.
Rainer
--
Rainer Buchty, LRR, Technical University of Munich
Phone: +49 89 289-28401, Fax +49 89 289-28232, Room S3240
Riccardo Rubini
> Reworking the schems to a high-level
> description (VHDL, Verilog) would be probably more work than just doing a
> hardware emulation.
>
> Rainer
And how would hardware emulation work ? For example, if you wanted to
emulate a CIA 6526, what should you need ?
Cameron Kaiser
>No. The design of the 6502, which is the core of the derivatives like the
>6510, is still owned by the Western Design Center
>(http://www.westerndesigncenter.com). You can get the 6510 from Sierra
>Circuit Design, http://www.sierracircuitdesign.com, though.
Interesting URL. They seem to have quite a few intriguing cores there. This
must be a new design, however, since it's based on the 65C02 and not the
original 6502 (which the real 6510/8500 is based on).
>GMT Microelectronics is still alive and kicking, although they don't have a
>domain name anymore. I found numerous references to them at Google. Their
>postal address is:
>GMT Microelectronics Inc
>Valley Forge Corporate Center
Are you sure? One of my sources reported that GMT has gone belly up.
--
Cameron Kaiser * cka...@stockholm.ptloma.edu * posting with a Commodore 128
personal page: http://www.armory.com/%7Espectre/
** Computer Workshops: games, productivity software and more for C64/128! **
** http://www.armory.com/%7Espectre/cwi/ **
Adam Dunkels
> No. The design of the 6502, which is the core of the derivatives like the
> 6510, is still owned by the Western Design Center
> (http://www.westerndesigncenter.com). You can get the 6510 from Sierra
> Circuit Design, http://www.sierracircuitdesign.com, though.
There is also a free VHDL 6502 core avaliable at
http://www.free-ip.com/6502/.
/adam
--
Adam Dunkels <ad...@dunkels.net> (Spambait)
http://dunkels.com/adam/
asdfd
it seems totally possible to emulate a 6502 in software in real time (emulated
1MHz) in a 10Mhz Atmel AVR microcontroller part, even with exact cycle times.
You've got 20 cycles to do the quickest instruction, like LDA #, and each AVR
instruction takes just 1 cycle, it should be plenty of time. Hmm, there's a
bit of overhead in indexing the instruction decode table, but no I think it
could be done.
But why not go further- I was thinking that you could emulate a 6502 and SID in
realtime plus mix and play the waveforms all in one small chip with no external
parts! Then you could have a portable sidmusic walkman. You might need
something more powerful like a MSP430 though.
Anyhow, to replace chips you could emulate the hardware completely in software
and just use I/O pins, this could be on a small circuit board with a 40pin dip
pins underneath. Plus you could add new instructions as you want; simulate the
rockwell version, a 65816, or any cpu you want just be uploading new firmware.
You could even make every instruction run single cycle and speed up programs
almost 4x that way even at 1Mhz. It could also have a built in DMA controller
and other peripherals. Wow, what a concept..
In fact I'm right in the middle of programming one of these chips and have a
working board here, today even I could write some code to simulate 6502 and
just see how well it works..
Even better - many bad chips could be replaced just by plugging in a
cartridge. An emulated SID could respond to memory accesses by watching the
cartridge port, so you would need to simulate all the pins just the registers,
which is a far better solution. And if one of you CIA's blew, you could
replace some of it.. umm.. I forget what each contorls, isn't one the disk
drive plus vic bank and the other joystick plus keyboard? You could plug the
disk drive into the cartridge and be 100% compatible but you can't fix the
screen or keyboard.
But, you could emulate the swift232 cartridge, a dual sid, an REU, maybe a few
other little things, cpu coprocessor, maybe a few other things.
At the very edge of ability is a vic emulator. It is possible to create a
video display purely in software with just 2 reistors and two i/o pins. It
would be like making a scroller using only border colors, if you know what I
mean. But that's pretty crazy, you probably want to add some more chips here.
But, wouldn't it be nice to have an emulated vic which outputs a nice clean
signal to a VGA monitor?
Except some things would look worse because pictures look better when blended
together a bit.
Rainer Buchty
rru...@galactica.it (Riccardo Rubini) writes:
|> And how would hardware emulation work ? For example, if you wanted to
|> emulate a CIA 6526, what should you need ?
A good documentation of the CIA, a text editor and an FPGA design tool.
Or in terms of the CPU or the VIC any decent document on Funet like the
VIC-II document or the 6510 description.
@host.sk CreaMD/Dmagic
quality is still around, somewhere. Just to find it.
CreaMD
Rainer Buchty
"CreaMD/Dmagic" <creamd @host.sk> writes:
|> Rainer I'm pretty sure technology to produce those chips in the same "low"
|> quality is still around, somewhere. Just to find it.
I severely doubt it. 1um is far too long gone that anybody would still keep
such "stone-age" fabs running just for the fact that someone might come along
with a 25 year old design which he wants to have reproduced.
But even *if* there would be some fab left, you possibly have to pay prices
way beyond your weirdest nightmares (just check the rates of companies who
are specialized in "geriatronics" offering services for those who just need
to stick with their technology). Never forget: Making chips isn't a "I need
10 of this and 5 of that" business but more like "how many 1000s do you
want - and how many $1000s do you wanna spend?".
What quite amuses me is the fact that, let's say quarterly, someone stands
up and asks for replacement chips. But despite the fact that a lot of people
did severe reverse engineering and documented their results (which in turn
led to more than authentic emulations), nobody took the step to do a freely
available VHDL or Verilog description of the chips which would allow everybody
with basic electronic skills to create their own replacement chips. Instead,
people are on the hunt for mystic schematic diagrams which are pretty much
useless for a today's reproduction of those chips - as if they would lead
to different results from what has already been discovered and implemented
into the emulators.
Alex Eisenhut
> (2) The Commodore custom chips were designed on resistor/transistor level
> (RTL), so you are not able to directly map this design to any modern ASIC
>
RTL means 'register transfer level'. What is this 'resistor/transistor level' you
speak of? Are you claiming the C64 uses resistor-transistor logic?
Andreas Boose
On Sun, 11 Nov 2001, asdfd wrote:
> I've been thinking about this,
> it seems totally possible to emulate a 6502 in software in real time (emulated
> 1MHz) in a 10Mhz Atmel AVR microcontroller part, even with exact cycle times.
Hm, I have only experience with emulating just one 6502 core in software
and only program Atmel AVR for a living, so I am probably not as
competent, but it DOES NOT seem to be possible to me.
> You've got 20 cycles to do the quickest instruction, like LDA #, and each AVR
> instruction takes just 1 cycle, it should be plenty of time. Hmm, there's a
And now tell me how to squeeze ADC/SBC into 20 cycles.
And this "1 cycle" thing is a myth, ADIW, SBIW, RJMP, IJMP, CPSE, SB*,
BR*, LD, ST, PUSH, and POP may take 2, RCALL, ICALL and LPM take 3 and RET
and RETI take 4 cycles.
> bit of overhead in indexing the instruction decode table, but no I think it
> could be done.
And do not forget to check for IRQ, NMI, READY and to generate PHI2.
> But why not go further- I was thinking that you could emulate a 6502 and SID in
> realtime plus mix and play the waveforms all in one small chip with no external
> parts! Then you could have a portable sidmusic walkman. You might need
> something more powerful like a MSP430 though.
Is it you Joe?
MfG Andreas
Rainer Buchty
Of course you are right... Don't know where I had my mind when associating
resistor/transistor logic to RTL.
|> Are you claiming the C64 uses resistor-transistor logic?
Oh, I'm not only claiming it... During my Vantis/AMD (long ago as you
probably know that Vantis belongs to Lattice for years now) times I had access
to the 8520 schematics in my hand for a customer driven evaluation about how
hard it would be to realize clones in modern CPLDs or FPGAs.
Unfortunately, for *that* effort the schems were completely useless, since
they indeed were on the lowest possible level, resistors and (FET, IIRC)
transistors (not to be mixed up with RTL as opposed to TTL or HCMOS). No
gates nor flip-flops on the schematics, which made them a bit uncomfortable
to read.
@host.sk CreaMD/Dmagic
maybe coordinate some efforts for this VHDL technical specification or some
chip tests. Maybe it's really possible to reproduce SID chip with all it's
plusses and flaws, it would be great to have more of those around as they
die very fast. (I've broken 3 already ;-(. since 1995)
CreaMD
--
_STUDIO STYLE______________
e-mail : rchl...@studiostyle.sk
phone : +421 37 6515 425
___________ www.studiostyle.sk _
___________ www.c64.sk _______
Rainer Buchty
"CreaMD/Dmagic" <creamd @host.sk> writes:
|> Rainer, you seem to have a great insight to this problematics, you could
|> maybe coordinate some efforts for this VHDL technical specification or some
|> chip tests.
I'd love to see people working together on chip clone projects instead of
starting their very own clone and finally dropping the project after some
time due to, well, time reasons.
|> Maybe it's really possible to reproduce SID chip with all it's
|> plusses and flaws, it would be great to have more of those around as they
|> die very fast. (I've broken 3 already ;-(. since 1995)
I'd say the SID is described even more detailed than the VIC (see the Yannes
interview by Andreas Varga plus the schematics included in the ReSID sources).
What I'm quite curious about is how many people already started some 6502,
VIA, CIA, SID or VIC (or TED) re-design and dropped it cause other things
in life got more important (unbelievable, I know :) ... It would be just great
to throw all these designs together to see if we couldn't together turn them
into working chips.
Riccardo Rubini
"Rainer Buchty" <buc...@nina.hotswap> ha scritto nel messaggio
news:9smccc$1ie$1...@sunsystem5.informatik.tu-muenchen.de...
> In article <76b8501b.01111...@posting.google.com>,
> rru...@galactica.it (Riccardo Rubini) writes:
> |> And how would hardware emulation work ? For example, if you wanted to
> |> emulate a CIA 6526, what should you need ?
>
> A good documentation of the CIA, a text editor and an FPGA design tool.
>
> Or in terms of the CPU or the VIC any decent document on Funet like the
> VIC-II document or the 6510 description.
>
> Rainer
Rainer, in few words and superficially...How hardware emulation of a chip,
let's say the 6526, works ? For example, what Jeri used in her CommodoreOne
prototype for substituting the custom MOS chips?
Riccardo
asdfd
Andreas Boose wrote:
> > You've got 20 cycles to do the quickest instruction, like LDA #, and each AVR
> > instruction takes just 1 cycle, it should be plenty of time. Hmm, there's a
>
> And now tell me how to squeeze ADC/SBC into 20 cycles.
>
What's so hard about that?
>
> And this "1 cycle" thing is a myth, ADIW, SBIW, RJMP, IJMP, CPSE, SB*,
> BR*, LD, ST, PUSH, and POP may take 2, RCALL, ICALL and LPM take 3 and RET
> and RETI take 4 cycles.
>
I know, don't have to use them often anyhow.
>
> > bit of overhead in indexing the instruction decode table, but no I think it
> > could be done.
>
> And do not forget to check for IRQ, NMI, READY and to generate PHI2.
>
Are you talking about full hardware emulation? OH, I agree with you 10Mhz is not
enough. However, I am talking about software emulation of just enough instructions
to play a sid music.
>
> > But why not go further- I was thinking that you could emulate a 6502 and SID in
> > realtime plus mix and play the waveforms all in one small chip with no external
> > parts! Then you could have a portable sidmusic walkman. You might need
> > something more powerful like a MSP430 though.
>
> Is it you Joe?
>
Huhn? Who's that?
>
> MfG Andreas
John Zielinsky
Burt wrote:
> this sounds so pessimistic? Isn't theres some positiveness in finding those
> documents and the information?
> To think that one could purchase new SID chips ah...
>
I agree. If you shake enough trees and rattle enough cages, you can get the
information out of the various companies. I've had to sweet talk a secretary at
Motorola for some information on long defunct chips.
As for freshly manufactured SID chips, why aim that low?
With today's advanced ADC's and DACs, why not design a super-sid?
64 voice polyphony, with stereo panning, 8 bit ADSR values, a 24 bit SPDIF
digital output, etc, etc,etc!
It'd just add a few registers... although with the added outputs, it wouldn't be
pin-compatible with the 6581... but it'd rock!
-Just dreaming,
John
Andrew Bosch
news:3beea119$0$57967$45be...@newscene.com...
> "Andrew Bosch" <bosc...@rcnchicago.com> writes:
> >No. The design of the 6502, which is the core of the derivatives like the
> >6510, is still owned by the Western Design Center
This
> must be a new design, however, since it's based on the 65C02 and not the
> original 6502 (which the real 6510/8500 is based on).
I guess that's what happens when you need to replace a 20-year-old chip.
Considering all of the soft cores and PGA's out there, you'd think that
someone would have burned a few equivalents, already.
> >GMT Microelectronics is still alive and kicking, although they don't have
a
> Are you sure? One of my sources reported that GMT has gone belly up.
It must be really recent, then, since the major search engines and most
engineering resources still list the old web site. Even EE Times had a part
review (voltage regulators) from GMT.
Andrew
Rainer Buchty
"Riccardo Rubini" <riccardo....@yyytin.it> writes:
|> Rainer, in few words and superficially...How hardware emulation of a chip,
|> let's say the 6526, works ? For example, what Jeri used in her CommodoreOne
|> prototype for substituting the custom MOS chips?
She took a good description of the 6526, figured out how this could be mapped
to logic circuits (e.g. we need these registers, that bus connection, those
bidirectional drivers) and formulated it in VHDL. Alternatively, one could draw
schematics and compile those into a programmable logic chip, but the high
level description is usually the more human readable way to do things.
What you finally get is a chip which behaves like the original but which is
not necessarily sticking to the original design techniques.
Cameron Kaiser
>>>GMT Microelectronics is still alive and kicking,
>>Are you sure? One of my sources reported that GMT has gone belly up.
>It must be really recent, then, since the major search engines and most
>engineering resources still list the old web site. Even EE Times had a part
>review (voltage regulators) from GMT.
One other possibility is that they filed for reorganisation and what we
see now is the 'new' GMT. I'll have to investigate.
MagerValp
JZ> As for freshly manufactured SID chips, why aim that low? With
JZ> today's advanced ADC's and DACs, why not design a super-sid? 64
JZ> voice polyphony, with stereo panning, 8 bit ADSR values, a 24 bit
JZ> SPDIF digital output, etc, etc,etc!
JZ> It'd just add a few registers... although with the added outputs,
JZ> it wouldn't be pin-compatible with the 6581... but it'd rock!
So you want a SoundBlaster Live card? They're sold over the counter in
every computer store.
--
___ . . . . . + . . o
_|___|_ + . + . + . . Per Olofsson, konstnär
o-o . . . o + Mage...@cling.gu.se
- + + . http://www.cling.gu.se/~cl3polof/
Matthew W. Miller
>There is also a free VHDL 6502 core avaliable at
>http://www.free-ip.com/6502/
Ah, but look at the Opcodes list, and compare the clock times of the
original CPU to those of the VHDL implementation-- it's not ready to be
wedged into a Cloneodore VIC20 just yet, I suspect.
--
Matthew W. Miller -- mwmi...@columbus.rr.com
Matthew Montchalin
How many extra pins would be needed? You could always bend a
couple pins over, and wire-wrap them to some unused lines on
the cassette port...
John Zielinsky
MagerValp wrote:
> >>>>> "JZ" == John Zielinsky <JZiel...@cfl.rr.com> writes:
>
> JZ> As for freshly manufactured SID chips, why aim that low? With
> JZ> today's advanced ADC's and DACs, why not design a super-sid? 64
> JZ> voice polyphony, with stereo panning, 8 bit ADSR values, a 24 bit
> JZ> SPDIF digital output, etc, etc,etc!
>
> JZ> It'd just add a few registers... although with the added outputs,
> JZ> it wouldn't be pin-compatible with the 6581... but it'd rock!
>
> So you want a SoundBlaster Live card? They're sold over the counter in
> every computer store.
Good point, althoug the SB doesn't have the analog filtering or ring
modulation, etc.
as a Digital/Analog hybrid, the SID just plain kicks butt!
John
Jo'ogn (J. Oppermann)
Rainer Buchty wrote:
>
> I'd love to see people working together on chip clone projects instead of
> starting their very own clone and finally dropping the project after some
> time due to, well, time reasons.
> What I'm quite curious about is how many people already started some 6502,
> VIA, CIA, SID or VIC (or TED) re-design and dropped it cause other things
> in life got more important (unbelievable, I know :) ... It would be just great
> to throw all these designs together to see if we couldn't together turn them
> into working chips.
having read the KEYS 12/2001 article about SID-Cult, Sample CDs
and the quadra-sid VST-PLug-in (there's even a link to my site ;)
i really wonder if all theses naive and melancholic
"ohhhh why not rebuild C64 chips" is a bit useless considering
all theses emulators. which are way less hassle, i guess.
even kept in mind that most ppl here involved and - there's no insult
meant, but truth - are not having the lightest experiences w/
electronics or chip designs. and those who have usually lack
time and motivation to finish sth like that.
and especially this SID-stuff wont work at all considering
the most important part: the filter which rely on rainers mentioned
old HCMOS design-flaws and dont even have proper resonance.
unless, this is all a "just for fun" project ...
to which even i lost interest over the "decades".
--
* Jo'ogn
* http://www.joogn.de
* "Das Weiß bestimmt niemand."
Rainer Buchty
"Jo'ogn (J. Oppermann)" <j...@joogn.de> writes:
|> and especially this SID-stuff wont work at all considering
|> the most important part: the filter which rely on rainers mentioned
|> old HCMOS design-flaws and dont even have proper resonance.
There is no reason why it shouldn't work... Just have a look at the PPG 2.V
VST plugin which emulates true SSM2044 behaviour in software. Not using the
very same fabrication process doesn't necessarily mean it can't be done
another way.
But even when realizing the filters w/ OTAs personally I don't really count
on emulating a muffled, thin sound (no offense, but the SID filters were truly
the weakest point of the whole thing) - but would be glad to get a richer,
more voluminous & crisp clear one virtually for free...
@host.sk CreaMD/Dmagic
CreaMD
Daniele Gratteri
> Oh, I'm not only claiming it... During my Vantis/AMD (long ago as you
> probably know that Vantis belongs to Lattice for years now) times I had
access
> to the 8520 schematics in my hand for a customer driven evaluation about
how
> hard it would be to realize clones in modern CPLDs or FPGAs.
8520? Isn't that the CIA used in the Amiga computers?
--
____
/ ___|____ Daniele Gratteri, Italian Commodore and Amiga user / /
| | |___/ E-MAIL: daniele_...@inwind.it / /
| |___|___\ URL: http://utenti.tripod.it/danielegratteri \ \/ /
\____| IRC: FIAT1100D - ICQ: 53943994 \/\/
Burt
>
> CreaMD
How did you manage that? (out of my own curiosity and to prevent this to
happening to my stuff & others... I know plugging in stuff when c64/128 is on
is a no-no. Other cases?)
--
___
/ __|__ Burt /Terminator / /
/ / |_/ www.museum.c64.org / /
\ \__|_\ Adoring C= 64 / 128 and \ \/ /
\___| Getting into Amiga, too... \/\/
Burt
> With today's advanced ADC's and DACs, why not design a super-sid?
> 64 voice polyphony, with stereo panning, 8 bit ADSR values, a 24 bit SPDIF
> digital output, etc, etc,etc!
YEAH, now you really got me excited... that would be sooooo cool...
Then getting a Hardsid card (maybe remade in PCI form) would then be truly awesome
with easy replacement chips...
A kick ass card for PC... Not only this it would be a really good replacement for
C64 (assuming it would work) or maybe addition to JERI's Commodore "Uno" (once and
if it becomes available)?
@host.sk CreaMD/Dmagic
Switching on amplifier (sid kicked the bucket)
using amplifier + big old TV and c64 together (???)
my friend noplaying with my setup (cables)
just switching on the c64 and realising that sid is brumming... ;-(
creamd
MagerValp
DG> 8520? Isn't that the CIA used in the Amiga computers?
Yes, and the 1581 and a couple of other. It's almost identical to the
6526, it's just lacking the TOD clocks.
Andreas Boose
On Sun, 11 Nov 2001, asdfd wrote:
> > And now tell me how to squeeze ADC/SBC into 20 cycles.
>
> What's so hard about that?
Come on, just write it down. And don't forget decimal mode.
> > And this "1 cycle" thing is a myth, ADIW, SBIW, RJMP, IJMP, CPSE, SB*,
> > BR*, LD, ST, PUSH, and POP may take 2, RCALL, ICALL and LPM take 3 and RET
> > and RETI take 4 cycles.
> >
>
> I know, don't have to use them often anyhow.
Hm, we can analyse the example you may have written down for ADC/SBC how
much branches and skips you have actually used.
> Are you talking about full hardware emulation? OH, I agree with you 10Mhz is not
> enough. However, I am talking about software emulation of just enough instructions
> to play a sid music.
But even then I cannot understand what the "AVR" has to do here. If it
does not have enough horsepower to do "hardware emulation", it also won't
have enough power for "software emulation", especially if it has to
emulated other parts of the C64 system as well then.
Just take look at the CPU consumption of modern (and accurate) SID
implementations like ReSID.
> > Is it you Joe?
>
> Huhn? Who's that?
Someone who "planned" to do nearly everything you could imagine (C64
hardware / software stuff), but acutally he has never got anything done
(or even started, if I remember right).
MfG Andreas
Raymond Carlsen
> Yes, and the 1581 and a couple of other. It's almost identical to the
> 6526, it's just lacking the TOD clocks.
The 8521 must be similar. I had a C64C (white case but older board) with
two different VIA chips soldered in at the factory... a 6526A for U1 and an
8521RO for U2.
Ray
Riccardo Rubini
"Daniele Gratteri" <viva...@libero.it> ha scritto nel messaggio
news:9ss2nn$15j2lu$1...@ID-23382.news.dfncis.de...
> 8520? Isn't that the CIA used in the Amiga computers?
>
Ciao Daniele,
yeah, it's also used in 1581. I just bought one 8520 today for...14 euro, I
wonder if it works as a substitute for the 6526. Anyone tried it ? I heard
that the 6526 on Amiga creates some problems. What about the 8520 into a C64
?
--
Riccardo Rubini
------
togliere xxx & yyy dall'indirizzo per rispondere
Please remove xxx & yyy from mail address to reply
Nicolas Welte
I also have one or two machines with one 8521 in them, but IIRC it is a C128.
I ran some tests on the chip, and the 8521 seems 100% software compatible
with the old 6526, including the TOD clock and the interrupt timing. The new
6526 on the other hand, which was manufactured after 1987, has some changes
in the timing which makes some software act strangely or even crash. There
must've been some reason, though, why Commodore used only one of the 8521
chips and why they dropped it in favour of the new 6526.
Also, the 8520 doesn't lack the TOD clock, it just behaves differently (and I
still think Commodore should've used another name than TOD in their docs for
the 8520). The registers are not laid out as a BCD coded time register, but
as a straight 24bit binary counter. The 50/60Hz bit is missing, because the
corresponding divider stage by 5 or 6 isn't needed. The clock input can also
be fed with up to 2MHz (in the 1581), and I don't think this is written in
the docs.
Nicolas
Riccardo Rubini
> still think Commodore should've used another name than TOD in their docs for
> the 8520). The registers are not laid out as a BCD coded time register, but
> as a straight 24bit binary counter. The 50/60Hz bit is missing, because the
> corresponding divider stage by 5 or 6 isn't needed. The clock input can also
> be fed with up to 2MHz (in the 1581), and I don't think this is written in
> the docs.
>
> Nicolas
Nicolas, what does happens if you use an 8520 in place of a 6526 in a
Commodore 64 or in a Commodore 128 ?
Also, what does happens if you substitute the 8520 into the 1581 with
a 6526 ? Ever tried this ?
Is there any risk of ruining the chips swapping them?
Riccardo
Matthew Montchalin
|Also, the 8520 doesn't lack the TOD clock, it just behaves differently
|(and I still think Commodore should've used another name than TOD in
|their docs for the 8520).
Interesting. What would you have suggested they call it?
Nicolas Welte
> Nicolas, what does happens if you use an 8520 in place of a 6526 in a
> Commodore 64 or in a Commodore 128 ?
I don't know, I never tried it. Very probably everything will work,
except programs that try to use the TOD clock. Those will act strangely,
because the chip behaves differently than expected.
> Also, what does happens if you substitute the 8520 into the 1581 with
> a 6526 ? Ever tried this ?
No, I didn't try it. Since the 1581 uses the TOD clock to detect if the
jumper J1 is opened or closed, this detection might fail. It's also not
worth thinking further about this, because 8520 chips are much cheaper
and easier available than 6526 chips.
> Is there any risk of ruining the chips swapping them?
No, except for the usual risk when handling those chips nothing should
happen.
Nicolas
Nicolas Welte
CNT, maybe? At least a better name for a counter than TOD.
Nicolas
Matthew Montchalin
If we need a 3 character combination, DAY sounds like a good alternative.
CNT is also pretty good.
Kelli Halliburton
news:9sm80j$1jni$1...@not.sics.se...
> There is also a free VHDL 6502 core avaliable at
> http://www.free-ip.com/6502/.
AHA! With this, it should be fairly simple to add the #0000 / #0001 I/O
port, and have oneself a VHDL 6510.
...shouldn't it?
Dave R.
$aZ2.17...@newssvr17.news.prodigy.com>:
>AHA! With this, it should be fairly simple to add the #0000 / #0001 I/O
>port, and have oneself a VHDL 6510.
>
>...shouldn't it?
Last I checked it lacked BCD mode and some of the opcode timings were
different, but more or less that's all you'd have to do.
mike
<JZiel...@cfl.rr.com> writes:
>
>As for freshly manufactured SID chips, why aim that low?
>With today's advanced ADC's and DACs, why not design a super-sid?
>64 voice polyphony, with stereo panning, 8 bit ADSR values, a 24 bit SPDIF
>digital output, etc, etc,etc!
But then it wouldn't be a SID chip and that would defeat the purpose of
emulating a SID chip, i.e. to replace broken SID chips.
Mike
Matthew Montchalin
If the thing fit into the SID socket, and had 28 pins, and was 90%
compatible, it would be good enough for me.