Buster chip replacement project

[midwan]

Hi people,
I've been thinking about an idea I had recently:
Would it be technically possible to produce a Buster replacement chip for the A4000 (maybe A3000 too), which would keep the compatibility but improve the performance?

So I talked to Dave Haynie the other day and he verified that yes, that could be done. He also mentioned that he was thinking of something similar himself. :-)

My idea was that if enough people are interested in this, we could organize a Kickstarter (or other crowdfunding) campaign to fund it. If Dave would be interested in designing this, it would be great opportunity to "do things right". We could have the Zorro busboard and any cards on it to perform at their maximum potential.

Of course, if not enough people are interested this would probably not work.

So, I posted a relevant question on the "Commodore Amiga" facebook group and I've started posting this on a few forums as well, just to see what kind of reactions we'll get.

Note: I am not a hardware guy myself (more of a software one) and have no previous experience with Kickstarter as an organizer. If people with more experience in these fields would like to help, that would be greatly appreciated.

So, what do you think?

[source]

This would be great

[spudje]

Given the struggle faced by the new A1200 cases kickstarter I 'm prettuy sure a Buster chip kickstarter will not succeed.

[alexh]

You'll never raise the money to make a 5v tolerant ASIC. We would be taking hundreds of thousands of dollars.

You'd have to use a CPLD/FPGA. But you may as well recreate the whole Amiga.

Add to that, too few Amiga users have an Amiga which uses a buster chip. A3000 / A4000.

This would have to be a hobby project, labour of love.

[Yulquen74]

I think you should build a board that holds a CPLD/FPGA solution that can be plugged into the original buster socket, a lot cheaper than creating a new chip.

[midwan]

Thanks for all the feedback.

I think the idea is to go for a CPLD/FPGA solution anyway, though it's not determined in what form yet.

There's an interesting discussion going on over at A1k.org's forum as well, where Jens has some interesting insights to offer.

Here's the thread there if you're interested: http://www.a1k.org/forum/showthread....233#post855233

[NorthWay]

Not all of us have a socketed Buster (even with a replaced 11 I don't).
Any chance of a "vampire" style clamp-on solution that you press onto the original?

[alexh]

Most have socketted busters. I thought only A4000D-CR has soldered buster?

[NorthWay]

The repair shop elected to solder mine in place.

[Jope]

Originally Posted by alexh

Most have socketted busters. I thought only A4000D-CR has soldered buster?

Many old A4000Ds with Rev9 busters have it soldered on.

[hooverphonique]

Originally Posted by Jope

Many old A4000Ds with Rev9 busters have it soldered on.

exactly.. mine now has a rev11, but it's also soldered on..

[edd_jedi]

Unfortunately people have a habit of over-estimating the Amiga market, which is why most new products and kickstarters never take off. The people behind the projects have the technical know-how, but not the economic skills to establish if it is feasible.

The A1200 housing project is a good example - 125,000 euro target for a 79 euro product requires over 1,500 sales. Yes of course a few people will pay more for the 'special editions' but most people just want the product. Simple maths like this show that it's not realistic, there are not 1,500 Amiga 1200 users that want a new case. I'd be surprised if there are more than a couple of thousand active Amiga hardware users in total across all systems.

[matthey]

Originally Posted by midwan

Would it be technically possible to produce a Buster replacement chip for the A4000 (maybe A3000 too), which would keep the compatibility but improve the performance?

So I talked to Dave Haynie the other day and he verified that yes, that could be done. He also mentioned that he was thinking of something similar himself. :-)

Thanks for talking to Dave and getting him interested in retro hardware projects. Amiga legends like him could also provide valuable marketing for a Kickstarter project (needs planning and prototyping first). There are several ways to go to avoid Buster bugs and bottlenecks.

1) Simple Buster replacement using FPGA
+ fully programmable and customizable
+ simplest and cheapest solution
- voltage regulator probably needed
- socket and surface mount Buster variations increase production costs
- possible connection problems with a clip on for surface mount Busters
- minimal speed gains while maintaining Zorro compatibility

2) Buster replacement using FPGA with PCI slots added
+ fully programmable and customizable
+ PCI gives major speed increase and cheap commodity PC hardware
+ good compatibility is possible as speedup is with PCI instead of Zorro
- voltage regulator probably needed
- socket and surface mount Buster variations increase production costs
- possible connection problems with a clip on for surface mount Busters
- added complexity and cost with PCI
- best PCI slot locations vary by Amiga model and may need cabling

3) FPGA accelerator card with PCI slots added
+ fully programmable and customizable
+ huge speedup is possible for CPU and I/O
+ PCI gives major speed increase and cheap commodity PC hardware
+ good compatibility is possible as speedup is with PCI instead of Zorro
+ socket or surface mount Buster doesn't matter
- much more complexity and added cost everywhere
- best PCI slot locations vary by Amiga model and may need cabling

4) Stand alone FPGA board with PCI slots
+ fully programmable and customizable
+ huge speedup is possible everywhere
+ PCI gives major speed increase and cheap commodity PC hardware
+ socket or surface mount Buster doesn't matter
+ standard power supplies and cases could be used
+ everything fits and a small Amiga is possible
+ could be used to simulate other retro hardware
- much more complexity and added cost everywhere
- some compatibility would likely be traded for cheaper production cost

PCIe (and SATA as they both need FPGA high speed serial transceivers) could be added in place of PCI for additional cost. PCI still gives the cheapest hardware and some Amiga drivers already exist while providing adequate speed (no more busboard bottleneck) for the CPU power.

IMO, idea #3 and #4 above make the most sense. Idea #3/#4 could be a 2 piece design with the accelerator board fitting into old big box Amigas and mating to a new I/O board (which could fit in Amiga 1200s for example). In any case, I hope there would be some cooperation for the 3rd generation of Amiga FPGA hardware (the FPGA Arcade and Mist are 2nd gen). Getting C= legends involved is brilliant both for their innovation and marketing inspiration. Bil Herd has also shown interest in doing some retro Amiga hardware (not so much an Amiga guy historically but very talented). There is much young FPGA talent around also. There is also more money available than what people would be willing to spend on new hardware if a well planned project was created. I know some people who could be valuable in planning, hardware creation and financing so contact me for further discussion if you and Dave are serious about doing an advanced retro Amiga project.

[Jope]

My eyes started glazing over when I read PCI slots.. How about just making a bugfixed buster that runs Zorro3 as fast as it can be run?

[alexh]

Pretty sure Zorro III already runs "as fast as it can" with a Buster 11. They can't run faster cycles and work with existing Zorro III cards

They are talking about Zorro II speed improvements, bug fixes and arbitration improvements.

[Locutus]

Who is going to design any hypothetical cards that benefit from faster bus driver? ;-)

[matthey]

Originally Posted by Jope

My eyes started glazing over when I read PCI slots. How about just making a bugfixed buster that runs Zorro3 as fast as it can be run?

Jens on a1k.org and alexh here are saying there isn't much "compatible" speed to be gained over a Buster 11. A new Zorro standard could be created which could negotiate faster speeds if the existing hardware could handle it but there would still be few new boards using it. Dave Haynie has said several times that he would use good existing standards like PCI which he has specifically mentioned. PCI or PCIe would open up a whole world of cheap PC hardware. I bought a PCI ethernet card for my Mediator for $5. PCI (32 bit 66MHz) has a bandwidth of 266MB/s where ZIII realistically has a bandwidth of about 15MB/s (ZIII was designed to do 150MB/s but it would take some magic in the old hardware to make that happen). Even if Dave could pull out a miracle and double the current ZIII speed on some old hardware which could handle it, it would still be a fraction of the speed of PCI. A bug fixed Buster would be neat in theory but even fixing bugs can cause incompatibilities in some cases.

A Buster card is probably not as easy and simple as many people think. You need the FPGA, voltage regulator (FPGAs run at <5V), flash memory to load the FPGA code (FPGAs turn on with no code loaded) and a port to update the FPGA code for bug fixes and enhancements. Then you have to figure out how to make it work reliably with socket and surface mount Busters. I would expect the price to be not more than 50% over a partially bug fixed Buster 11 to sell well. If Dave wants to do this project for fun and can work some miracles then maybe he could come close to this target. IMO, his name and expertise would be wasted on a project with a very small benefit to less than half of existing Amiga users (only big box Amiga users). Keeping the classic Amiga alive requires more modernization. Adding PCI is probably one the cheapest ways to enhance existing classic Amigas. Some kind of PCIe+SATA stand alone board might appeal to people outside the Amiga community (bringing in new Amiga users) if the price was low enough. Besides faster I/O and custom chipset (with bug fixes), a faster 68k CPU is also needed to improve and modernize the Amiga experience. I own a CSMK3 68060@75MHz with Mediator providing Voodoo 4+ethernet+USB. This hardware setup makes a great semi-modern Amiga but new software is lacking because there isn't enough of this hardware. I can see the potential as more speed, less bottlenecks and better integration of hardware is possible with today's technology. A Buster only upgrade doesn't do much to help, IMO.

[alexh]

Originally Posted by matthey

A Buster card is probably not as easy and simple as many people think. You need the FPGA, voltage regulator (FPGAs run at <5V), flash memory to load the FPGA code (FPGAs turn on with no code loaded)

Some FPGA's (especially the smaller CPLD's) have their flash memory integrated.

Originally Posted by matthey

a port to update the FPGA code for bug fixes and enhancements.

The newer Xilinx parts will allow configuration from the design. (i.e. you could in theory change the flash contents from the Amiga)

Originally Posted by matthey

Keeping the classic Amiga alive requires more modernization.

Way off topic but no-one is ever going to use a Classic Amiga as their main computer no matter what modernisation you add.

[Amiga1992]

People need to lay off the fan fiction with Amiga.
How is the "classic Amiga" an "Amiga" if you add a lot of shit that isn't "Amiga"?

None of the proponents of "modern Amiga" are ABLE to define what the fuck Amiga is or would be today. Seems to me like "Amiga" would just be a generic computer platform running a variety of unsupported OSes, based on already obsolete technology needed to keep compatibility with said archaic operating systems.

Now, CLASSIC AMIGA, that we can define very easily. It would be the classic computers, 500s,600s,1200s,300s,400s, etc., running on the 68k architecture. Keeping THAT alive means a lot of research and development into hardware that expands its capabilities somehow but keeps as near as 100% compatibility as possible with legacy applications. This research is usually not affordable because there just ISN'T a user base big and strong enough to make things happen (unlike the C64 one).

Why would any of the proposed expansions, be it SATA or PCI or whatever, bring ANYONE to the platform, if you have absolutely fuck all to use the computer with that you can't ALREADY DO with myriad of other platforms available in the market more easily and even cheaper??

I totally can understand people wanting to get into an Amiga platform and push it to the limit and use it today as their main platform, people with tinkering/hacker attitude, however painful that sounds to me, but you aren't going to miraculously attract anyone to an obsolete platform that doesn't have much going on for it and a platform in which you have to pay through the nose for everything, starting with the obsolete computer to base it all on, and adding up the obsolete and modern peripherals you might need to put it on par with everyday computing technology of the year TWO THOUSAND AND FOUR.

[matthey]

Originally Posted by alexh

Some FPGA's (especially the smaller CPLD's) have their flash memory integrated.

It is in the interest of FPGA manufacturers to make it easy to load the FPGA code so, yes, some FPGAs have help with this.

Originally Posted by alexh

The newer Xilinx parts will allow configuration from the design. (i.e. you could in theory change the flash contents from the Amiga)

A nice feature that could take some time to program, if possible for the Buster. At least the prototype would likely use a port.

Originally Posted by alexh

Way off topic but no-one is ever going to use a Classic Amiga as their main computer no matter what modernisation you add.

There is a lack of software because there is a lack of high enough spec hardware to run more modern software. A Buster upgrade is fairly complex but does not help much. An FPGA accelerator with faster I/O options would help some and isn't too much more complex. A standalone FPGA board would be more complex yet but could help a lot if the price was reasonable. Off topic when comparing multiple hardware options which all have an FPGA and I/O enhancements?

Originally Posted by Akira

People need to lay off the fan fiction with Amiga.
How is the "classic Amiga" an "Amiga" if you add a lot of shit that isn't "Amiga"?

None of the proponents of "modern Amiga" are ABLE to define what the fuck Amiga is or would be today. Seems to me like "Amiga" would just be a generic computer platform running a variety of unsupported OSes, based on already obsolete technology needed to keep compatibility with said archaic operating systems.

The most important quality of an Amiga, to me, is that it runs most Amiga software which requires an OS like AmigaOS. This infers executing 68k code.

Originally Posted by Akira

Why would any of the proposed expansions, be it SATA or PCI or whatever, bring ANYONE to the platform, if you have absolutely fuck all to use the computer with that you can't ALREADY DO with myriad of other platforms available in the market more easily and even cheaper??

Why would anyone buy a Raspberry Pi when they can buy a 10 year old PC for twice the price that has several times the CPU power, better graphics and way more expansion? Can we say it was just the low price 5-6 million Pi users later because it isn't value. Maybe some people want something different.

Originally Posted by Akira

I totally can understand people wanting to get into an Amiga platform and push it to the limit and use it today as their main platform, people with tinkering/hacker attitude, however painful that sounds to me, but you aren't going to miraculously attract anyone to an obsolete platform that doesn't have much going on for it and a platform in which you have to pay through the nose for everything, starting with the obsolete computer to base it all on, and adding up the obsolete and modern peripherals you might need to put it on par with everyday computing technology of the year TWO THOUSAND AND FOUR.

Only the standalone board(s) option could appeal to the outside world. The first design may not be high enough spec for most of the outside world and would be aimed at existing Amiga users and ex-Amiga users while development improves. The CPU would be limited in performance while it remains in FPGA but memory could be expanded to a more usable size for modern programs. In FPGA, it could approach the Raspberry Pi single core performance which might surprise some people using an efficient OS like the AmigaOS. My nephew in his late teens, who owns an XBox, described the graphics of QuakeGL on my Amiga as very good and never complained about the speed (~24 fps). I am blessed to have what I have but we could have even faster, more modern and better integrated Amigas using modern FPGA technology. Is it an Amiga? Who cares if it is more fun, more usable and most Amiga software runs on it.