User avatar
grshaw

Posted Sat Oct 19, 2019 6:28 am

I have been thinking about the long term for Amiga recently. Beloved as our old machines are, they will not last forever.

For the long term health Amiga hardware we need to be able to build new Amigas from scratch ourselves. I am in the early stages of building an A2000 myself. You can get replica PCBs ( thanks Acill ). For most simple components, such as resistors and capacitors, you can still get these from electronics suppliers. What about the Amiga Custom Chips? You can buy these from Individual Computers NOS. However their stock will run out eventually. You can also take them out of a compatible donor machine. However, this is not a sustainable option because there is now one less Amiga.

So, what to do about Amiga Custom Chips?

Seems to me that one potential solution available to us hobbyists would be to use FPGAs to implement the custom chips. To be clear, I am not thinking along the lines of The Vampire here. They are using one FPGA to implement much of the Amiga. I’m thinking of a pin for pin compatible replacement for each custom chip. One for Paula, one for Denise, probably more than one for Agnus depending upon required waist measurements. Also, potentially replacements for Amber and Buster etc.

What would this look like? I am thinking that physically, it could look similar to what the FPGASid looks like. A small PCB with an FPGA mounted on it. Pins on the bottom to allow it to plug into the relevant socket.

How feasible is this? Unknown at this stage. I would certainly be interested to hear everyone’s thoughts.

FPGASid for the C64 provides a good example of how such a project could work.

Possible problems? Yes. There are. I’ve tried googling this and there are concerns about voltage compatibility. I may not have this correct, but I think that the Amiga uses +5V to represent high. This is too much for modern FPGAs, which require less voltage, such as 3.3V? There may be ways round this. If it’s a heat issue, perhaps you could run the FPGA at 5V but add extra cooling. Perhaps, running at the low ( by modern standards ) clock frequently of the Amiga would mean that 5V would actually be okay?

However, there is a bigger problem. How would you actually implement a custom chip as an FPGA? Strikes that this would potentially be a Herculean task. There are some potential shortcuts though. For example, there is the Open Source minimig project, which implements a whole A500 as an FPGA. How much of this could be re-used? I have read that the minimig is more like an FPGA implementation of an Amiga emulator rather than a pin for pin reimplementation. Would still be a good place to start looking I guess. Also , the Vampire project itself. If they have a modular design, then they may already have FPGA code for the custom chips. Vampire is not open source though, so they may be unwilling to share.

Now, if we still had the original design for the custom chips from Commodore, then that would help massively. Anyone know what happened to them?

Next steps? Once you have an FPGA implementation of a custom chip, then it’s a short step to actually creating new silicon again. That’s why FPGA’s were developed originally after all. Burning new silicon would not really be necessary. I’m sure that FPGA’s would be more than a match to the task. However it would be damn cool to build real Amiga chipsets in the 21st Century wouldn’t it? It would come down to economics I think. There is a minimum production run, under which it would not be economically viable. But the 21st Century already has a solution to such problems. Crowdfunding! I can imagine the Kickstarter or Indiegogo campaigns even now.

I don’t have solutions to any of the problems above. Just a brain dump of what I am thinking on the subject really. I’d be really interested to hear anyone else’s thoughts.

What do we do when the NOS chips run out?

User avatar
McTrinsic

Posted Sat Oct 19, 2019 6:54 am

Hi grshaw,

This has recently been discussed on a1k.org when Matze previewed his upcoming FPGA-based accelerator. This accelerator effectively replaces the 68k CPU with an FPGA replacement (and some more things).

Hope the stuff is readable via translators.

Bottom line: currently not possible and/or extremely expensive.

The one exception we have so far is Denise in the Indivision ECS, both MKI & II.

And with iComp leaving the Amiga it might get worse, as the initial cost and time (=efforts) seem to be significant.

To be honest, I don't know what to think about. I always assumed the Amigas will live longer than I will, and I don't think our kids will really care. Why should they. We don't really care about wired phones either don't we? Also, the interest in old vinyl based juke boxes or audio in general is rather lingering around.

As Frank Zappa said, "Jazz is not dead. It just smells bad..."

And frankly, I don't care about Jazz, Charleston or other stuff from the past.

Except Amiga :D

Cheers,
McT

User avatar
nonarkitten

Posted Sat Oct 19, 2019 1:13 pm

People who talk about ASIC have no clue how difficult and expensive it's become. Sadly, the days of 130-micron cheap fabs are behind us and designs need to work sub 90-nanometre. If you don't have a million dollars to spare and aren't targeting 100K volume or better, do NOT waste your time thinking about it.

That absurdity aside, putting each Amiga chip (Agnus/Denise/Palua) into an FPGA would not be very hard, and in-fact, most of the MiniMig files seem to keep each chip logic fairly separate. The whole chipset fit into a single 400,000 gate Spartan, so we're not talking about big chips here -- around 150,000 gates each?

Which is still absurdly huge -- the original ECS chipset was supposed to be around 27,000 gates, so we're 5 times the size with the MiniMig reimplementation. I believe when Jeri looked at the design of MiniMig she commented on all the pointless buffering and how complex it was versus the real chips. Sadly, without a real reference we'll never know.

I would recommend looking at instant-on FPGA's such as the ProASIC3 or IGLOO from Microsemi. The normal delay of programming the FPGA at power-on might take too long. There's a 5x5mm CSP with 81 IO called the IGLOO nano which might spare enough board space to pack on all the buffers for level shifting to 5V. However, even the 14x14mm VQPF packages should fit in a DIP48 form-factor and still have room for the shifters.

This FPGA's use Flash fabric instead of SRAM and aren't based on Microsemi's older 'antifuse' technology. They are every bit as fast as SRAM devices when running and have the benefit of being completely non-volatile. There are already several FPGA's on PLCC out there and some VERY tiny ProASIC3 boards on DIP boards.

Image

This isn't pin-compatible, but shows how everything could fit.

User avatar
grshaw

Posted Sun Oct 20, 2019 10:37 am

Thanks guys for your thoughts.
McTrinsic: this has recently been discussed on a1k.org
I will definitately take a look at this. Google translate is good enough to more or less understand what is being said. :D
McTrinsic: I always assumed the Amigas will live longer than I will, and I dont think out kids will really care. Why should they. We dont really care about wired phones either dont we? Also, the interest in old vinyl based juke boxes or audio in general is rather lingering around.
Yeah - you could well be right about that. I guess time will tell.
nonarkitten: I would recommend looking at instant-on FPGA's such as the ProASIC3 or IGLOO from Microsemi. The normal delay of programming the FPGA at power-on might take too long. There's a 5x5mm CSP with 81 IO called the IGLOO nano which might spare enough board space to pack on all the buffers for level shifting to 5V. However, even the 14x14mm VQPF packages should fit in a DIP48 form-factor and still have room for the shifters.

This FPGA's use Flash fabric instead of SRAM and aren't based on Microsemi's older 'antifuse' technology. They are every bit as fast as SRAM devices when running and have the benefit of being completely non-volatile. There are already several FPGA's on PLCC out there and some VERY tiny ProASIC3 boards on DIP boards.
Wow. You certainly seem to know your stuff on FPGAs. These sound like great recommendations. If I ever get round to attempting something like this myself, then I will certainly look into this suggestion. These FPGAs look ideal.

Interesting also that the minimig core does in fact seems to have separate code for each chip. This would seem to be the best place to start indeed.

At this time, I am spending all my free time on my A2000 build project. But once this is done, I think I can see another project on the horizon. :D

User avatar
McTrinsic

Posted Sun Oct 20, 2019 11:02 am

Maybe one point to consider is how to make an FPGA implementation of the chips into products with added functions so they make it worth selling it.

This is my conclusion from thinking about the Indy ECS.

The V2 is an FPGA implementation of Denise. The added functionality here is the Flickerfixer/Scandoubler use.

For Agnus, it could be a 2MB Chip expansion. One that could be switchable from 512kb/1M/2MB. And/or NTSC/PAL. At least for ECS/OCS Amigas.

For Gary, it could be a KickROM-Flash replacement.

I am not sure about an added functionality for the buster.

Paula could be a challenge. Added speed for HD Floppy disks? Added audio capabilities?

And I have no idea about Buster and Akiko.

User avatar
obitus1990
New Orleans, LA, USA

Posted Thu Oct 24, 2019 7:25 pm

Gary-Eingebaut2-K.jpg
Matze released his CPLD code and design for Gary yesterday. If you're on A1k.org, go grab it. Photo is above.


Here's the file, for those who do not have access to A1K:
Gary-Repalcement.zip
(114.2 KiB)

User avatar
grshaw

Posted Fri Oct 25, 2019 4:59 am

Awesome! I will take a look.

When I started this thread it was really just a brain dump of my thoughts. I had no idea that people were already working on exactly the same thing that I had been imagining. This Amiga community never fails to surprise me. We are so lucky to have so many talented and enthusiatic people. :D

User avatar
Crazyeights

Posted Thu Nov 07, 2019 8:18 pm

I am pretty sure that Jens (Icomp) already did this years ago. Try searching for the Clone-A project.

User avatar
dalek

Posted Thu Nov 07, 2019 8:34 pm

Jeri Ellsworth did all this years ago. Watch this YouTube Video

Shame it hasn't been released. But as you can see, the original chip designs still exist.





Return to “Hardware”