Jump to content

X8/X16/Xwhatever "Pure" FPGA Unique Sales Proposition


Recommended Posts

A question original posed elsewhere by@Janne Sirén, relocated here to avoid the appearance of me derailing another thread:

Quote

I hope this is a suitable thread to answer in, coming from the X8 poll thread. Moving onwards from any of that debate, I guess my question/point is how/what do you perceive the unique sales proposition being for an FPGA-only X8 or X16? There are certainly FPGA-only computers already out, but usually they have some other USP such as a case/compability/legacy with ZX Spectrum Next or the MEGA65, or they are lean-mean FPGA monsters like the MiSTer family, the ZX-One family or training boards such as Nexys. Others, like the Foenix computers, have the unique proposition of offering several discrete chips such as "real" processors and several hardware sound chips in addition to FPGAs.

What would be the unique sales proposition of a simple FPGA board like X8/16 that could easily run on any existing FPGA platform that are already being produced? I am genuinely curious about your answer and I hope my question also illustrates the area where I think it would be beneficial for the X16 project to continue having a clear message and a clear point.

While the initial question was posed to @Scott Robison I certainly think everyone is welcome to pose their own thoughts.

Link to comment
Share on other sites

Im not @Scott Robison but I think one perhaps unintended side effect of the X8 is the possibility that it can be programmed to run other 8bit/6502 Machines in FPGA alongside the VERA in FPGA  for Video output. Granted that would probably require "erasing" the X8 ROM and assuming the ICE40UP5k could be programmed multiple times means that the X8 platform could be the basis for a truly inexpensive 8bit AIO FPGA machine. 

  • Like 3
Link to comment
Share on other sites

Thank you, Scott. Indeed I am very interested in your thoughts. When X16 used to be a 65816 processor and a pair of discrete video and audio chips, that was something unique. Today the Foenix project of course offers almost that in a range of several professionally manufactured computers (the video is FPGA, but otherwise).

If the X16 is turning into more of a simple FPGA board, what is unique about it that isn't already out there? For example, here is the SiDi FPGA from the ZX Spectrum community:

https://manuferhi.com/p/sidi

89 € ($100), professionally produced and finished/cased, available today. Or 69 € ($80) professionally produced without the case. This is one of maybe 20 variants of simple FPGA computers made for running Spectrum, Commodore and similar cores that are readily available at lucrative prices. These are fairly powerful little machines with a plethora of connectors as well.

Edited by Janne Sirén
  • Like 1
Link to comment
Share on other sites

On 10/15/2021 at 8:03 PM, Mtemal said:

Im not @Scott Robison but I think one perhaps unintended side effect of the X8 is the possibility that it can be programmed to run other 8bit/6502 Machines in FPGA alongside the VERA in FPGA  for Video output. Granted that would probably require "erasing" the X8 ROM and assuming the ICE40UP5k could be programmed multiple times means that the X8 platform could be the basis for a truly inexpensive 8bit AIO FPGA machine. 

If it is that, there is already an entire range of MiSTer and ZX-Uno based FPGA boards that probably do such things even better and cheaper.

  • Like 1
Link to comment
Share on other sites

On 10/15/2021 at 8:16 PM, Mtemal said:

Wow, that SIDI is one heck of a machine.

And I mean, there are 10-20 more like it. They have been making really great FPGA computers in the ZX Spectrum/C64 and ST/Amiga communities for the past decade+. They've nailed it down to an art by now, very cheap, very powerful too.

Edited by Janne Sirén
  • Like 1
Link to comment
Share on other sites

From a purely legal standpoint, in a pure FPGA solution, the ROM is what makes the X-whatever unique. It is based on intellectual property from Cloanto (the kernal, as the current successor to Commodore) and Microsoft (BASIC, though I don't know how the initial license fee Commodore paid transfers to Cloanto but it seems it does indeed transfer). As such, these are not free to copy to other platforms. So having a retro-inspired computer with legally licensed ROM that is derived from the Commodore heritage is one USP. While it would be possible to port the HDL that makes the hardware to other platforms, it would be "useless" to most people without the ROM, and that is encumbered in its current state.

Having a compact FPGA based design that is not subject to some of the downsides of software emulation is another USP. Software emulation is great, and I quite enjoy my The C64, but it isn't ideal. Especially if you think of a Raspberry Pi form factor, an FPGA solution that behaves as a real hardware solution is going to be superior to porting the emulator to an existing SBC platform.

While not directly related to the FPGA nature of the original question, a PETSCII inspired keyboard with symbols printed on the keys is a third USP for the project as we understand it at this time.

Given some time and less of a time crunch (I have to get ready to go teach some middle schoolers how to write Python) I may be able to come up with more, but I think those three are significant. Whether they are enough to entice people to buy into the platform is another question.

Link to comment
Share on other sites

On 10/15/2021 at 11:13 AM, Janne Sirén said:

If it is that, there is already an entire range of MiSTer and ZX-Uno based FPGA boards that probably do such things even better and cheaper.

I'm not familiar with the ZX-Uno (I mean, I've heard of it but literally know nothing else). The MiSTer is nice, though it currently costs a lot more than we expect an FPGA X16 to cost (though we're guessing on that based on past information that is incomplete).

Link to comment
Share on other sites

On 10/15/2021 at 8:23 PM, Scott Robison said:

I'm not familiar with the ZX-Uno (I mean, I've heard of it but literally know nothing else). The MiSTer is nice, though it currently costs a lot more than we expect an FPGA X16 to cost (though we're guessing on that based on past information that is incomplete).

ZX-Uno is basically a whole family of FPGA computers that run various cores. One variant available here for 55 € ($65):

https://www.antoniovillena.es/store/product/zxunoplus/

As for MiSTer, SiDi is MIST compatible (all but its largest cores) and costs 65 € ($80):

https://manuferhi.com/p/sidi-fpga-sin-carcasa-without-case

I believe both ZX-Uno and SiDi are way more powerful than X16/X8.

  • Like 2
Link to comment
Share on other sites

Oh, a fourth USP: Building on from the Commodore heritage. By this I mean more than just the ROM, I mean the system as a whole. 17M+ C64 computers, 1M+ VIC 20, 5M+ C128. I know there are a lot more of other computers out there at this point, but the Commodore and 6502 lineage is going to appeal to a completely different audience than the Sinclair and Z-80 lineage. That doesn't make one better and the other worse, but I think this is the first time someone has tried to put out something like this (FPGA based actual hardware vs emulation on an ARM or other alien hardware platform) at something like this price point. The fact that the Sinclair inspired models have done as well as they have I think indicates the market might really enjoy something like this. Similar space, similar objectives, achieved differently.

Link to comment
Share on other sites

On 10/15/2021 at 8:21 PM, Scott Robison said:

From a purely legal standpoint, in a pure FPGA solution, the ROM is what makes the X-whatever unique. It is based on intellectual property from Cloanto (the kernal, as the current successor to Commodore) and Microsoft (BASIC, though I don't know how the initial license fee Commodore paid transfers to Cloanto but it seems it does indeed transfer). As such, these are not free to copy to other platforms. So having a retro-inspired computer with legally licensed ROM that is derived from the Commodore heritage is one USP. While it would be possible to port the HDL that makes the hardware to other platforms, it would be "useless" to most people without the ROM, and that is encumbered in its current state.

Having a compact FPGA based design that is not subject to some of the downsides of software emulation is another USP. Software emulation is great, and I quite enjoy my The C64, but it isn't ideal. Especially if you think of a Raspberry Pi form factor, an FPGA solution that behaves as a real hardware solution is going to be superior to porting the emulator to an existing SBC platform.

While not directly related to the FPGA nature of the original question, a PETSCII inspired keyboard with symbols printed on the keys is a third USP for the project as we understand it at this time.

Given some time and less of a time crunch (I have to get ready to go teach some middle schoolers how to write Python) I may be able to come up with more, but I think those three are significant. Whether they are enough to entice people to buy into the platform is another question.

Sure, I get that if you can build an ecosystem around the X16 core, that can be something lead to something unique on the "soft" front so to speak. The keyboard is a nice bonus as well. I already have the premium one at home. Maybe David's YouTube following will help make all that happen and then the generic nature of the hardware might not even matter. Perhaps that will be the way it goes.

I guess my point is, there is already a plethora of cheap, powerful FPGA retro boards available in reasonable volumes on one hand, as well as unique cased retro platforms like the ZX Spectrum Next, MEGA65 (licensed too both) and perhaps Foenix to a degree, it will be important for the X16 to finds its niche. The previous plan with a desktop case, discrete chips etc. certainly was one way. If that were not to be, IMO it would be useful to define some other niche that sets it apart from these, but it is unclear to me what that might be. Just my two cents.

  • Like 1
Link to comment
Share on other sites

On 10/15/2021 at 11:35 AM, Janne Sirén said:

Sure, I get that if you can build an ecosystem around the X16 core, that can be something lead to something unique on the "soft" front so to speak. The keyboard is a nice bonus as well. I already have the premium one at home. Maybe David's YouTube following will help make all that happen and then the generic nature of the hardware might not even matter. Perhaps that will be the way it goes.

I guess my point is, there is already a plethora of cheap, powerful FPGA retro boards available in reasonable volumes on one hand, as well as unique cased retro platforms like the ZX Spectrum Next, MEGA65 (licensed too both) and perhaps Foenix to a degree, it will be important for the X16 to finds its niche. The previous plan with a desktop case, discrete chips etc. certainly was one way. If that were not to be, IMO it would be useful to define some other niche that sets it apart from these, but it is unclear to me what that might be. Just my two cents.

I think your point is certainly reasonable. I can come up with reasons why X16FPGA has some USP, but whether other people will agree is another issue. I think X16 has the potential to break out beyond those who are into FPGA recreations of classic computers at a much more affordable price point than Foenix (sp?) is making available, for example. But that supposes that people are willing to buy into a niche product. Only time will tell on that one.

I know I certainly couldn't do it with my YouTube audience (which is 0.0074% the size of David's). Maybe his dream will inspire others, and maybe it won't.

Link to comment
Share on other sites

On 10/15/2021 at 8:29 PM, Scott Robison said:

Oh, a fourth USP: Building on from the Commodore heritage. By this I mean more than just the ROM, I mean the system as a whole. 17M+ C64 computers, 1M+ VIC 20, 5M+ C128. I know there are a lot more of other computers out there at this point, but the Commodore and 6502 lineage is going to appeal to a completely different audience than the Sinclair and Z-80 lineage. That doesn't make one better and the other worse, but I think this is the first time someone has tried to put out something like this (FPGA based actual hardware vs emulation on an ARM or other alien hardware platform) at something like this price point. The fact that the Sinclair inspired models have done as well as they have I think indicates the market might really enjoy something like this. Similar space, similar objectives, achieved differently.

Perhaps, but then again SiDi and MiSTer and related products already serve the Commodore market as well. And there is the C-One, Minimig, Turbo Chameleon legacy alive still as well, and the Vampire direction. Both communities have been into FPGAs since the early naughts and there are great cores for running most Commodores. And the Ultimate-64 is selling by the thousands (I believe soon 4000 units solds), an FPGA C64 motherboard not too far above the intial X16 price projections. And of course there is the MEGA65 now taking preorders as you of course know.

I get what you are pointing at, though, if the Commander X16 is packaged well (with software, licenses, peripherals, the lot), it can certainly be one way to carve out a niche for it.

  • Like 1
Link to comment
Share on other sites

Another angle to look this from is that the Foenix actually turned out to be closer to the original dream computer specs than the X16 is headed towards. With its 65816 processor (and beyond) as well as discrete sound chip(s) as well as professional manufacturing for the boards, it is certainly quite close to that goal in all ways expect the price (and its video is FPGA). Even then Foenix U's $300 is not that terrible considering its vast features and the expected price of the initial X16 non-kit version which seems to be headed towards that region as well. Of course just having the hardware does not a platform make, but just painting a picture of the landscape where I think it is important for the X16 to find its niche.

  • Like 1
Link to comment
Share on other sites

On 10/15/2021 at 11:45 AM, Janne Sirén said:

Perhaps, but then again SiDi and MiSTer and related products already serve the Commodore market as well. And there is the C-One, Minimig, Turbo Chameleon legacy alive still as well, and the Vampire direction. Both communities have been into FPGAs since the early naughts and there are great cores for running most Commodores. And the Ultimate-64 is selling by the thousands (I believe soon 4000 units solds), an FPGA C64 motherboard not too far above the intial X16 price projections. And of course there is the MEGA65 now taking preorders as you of course know.

I get what you are pointing at, though, if the Commander X16 is packaged well (with software, licenses, peripherals, the lot), it can certainly be one way to carve out a niche for it.

The primary downside to those other products is that in order to use them, you have to violate copyright law in most cases. They don't have to provide ROM images (maybe they do, but it's not a prerequisite) but someone has to acquire the ROM images. I think a good case can be made that the copyright system is broken and that it shouldn't be an issue, but it is.

Actually, I guess it is possible to license the ROM images as an individual by purchasing C64 Forever or whatever it is, but I suspect the number who actually do is infinitesimal.

Link to comment
Share on other sites

On 10/15/2021 at 9:09 PM, Scott Robison said:

The primary downside to those other products is that in order to use them, you have to violate copyright law in most cases. They don't have to provide ROM images (maybe they do, but it's not a prerequisite) but someone has to acquire the ROM images. I think a good case can be made that the copyright system is broken and that it shouldn't be an issue, but it is.

Actually, I guess it is possible to license the ROM images as an individual by purchasing C64 Forever or whatever it is, but I suspect the number who actually do is infinitesimal.

In the case of Spectrum, emulation use of the ROMs has been allowed by the license owner and of course ZX Spectrum Next has even the Sinclair name license. In Commodore's case they are owned by Cloanto, who has licensed them to various parties like Turbo Chameleon, MEGA65, THEC64 etc. So it is not quite black and white in that regard. For the Ultimate-64 they can be installed from a legal source as well, so it is not a pirated board. And I mean, all of these certainly have a tighter connection with their respective "parent" platforms than X16, which is very loosely connected to anything Commodore.

Personally I think the best chance X16 has is probably as a sort of virtual platform (indeed one that already exists in emulation), an FPGA platform defined by the team loosely based on the dream computer idea, then popularized by the YouTube channel and distributed on a somewhat generic piece of hardware. I think that could certainly work. Probably the hardware won't be very unique or different and that's kind of too bad, but certainly a platform could form this way.

If anyone has any thoughts on how to make the hardware proposition unique, when so many cheap retro FPGA boards already exist, I think hashing those out might benefit the project and community here.

  • Like 1
Link to comment
Share on other sites

On 10/15/2021 at 10:21 AM, Scott Robison said:

From a purely legal standpoint, in a pure FPGA solution, the ROM is what makes the X-whatever unique. It is based on intellectual property from Cloanto (the kernal, as the current successor to Commodore) and Microsoft (BASIC, though I don't know how the initial license fee Commodore paid transfers to Cloanto but it seems it does indeed transfer). As such, these are not free to copy to other platforms. So having a retro-inspired computer with legally licensed ROM that is derived from the Commodore heritage is one USP. While it would be possible to port the HDL that makes the hardware to other platforms, it would be "useless" to most people without the ROM, and that is encumbered in its current state.

Having a compact FPGA based design that is not subject to some of the downsides of software emulation is another USP. Software emulation is great, and I quite enjoy my The C64, but it isn't ideal. Especially if you think of a Raspberry Pi form factor, an FPGA solution that behaves as a real hardware solution is going to be superior to porting the emulator to an existing SBC platform.

While not directly related to the FPGA nature of the original question, a PETSCII inspired keyboard with symbols printed on the keys is a third USP for the project as we understand it at this time.

Given some time and less of a time crunch (I have to get ready to go teach some middle schoolers how to write Python) I may be able to come up with more, but I think those three are significant. Whether they are enough to entice people to buy into the platform is another question.

We've already talked about the ROM in other threads. Short version, there's no need to license Commodore's ROM. There are several 6502 BASICs out there, and there's more than enough brains in this community to crowdsource a kernel - even without relying on something like the open KERNAL started by the MEGA 65 crowd. 

 

Link to comment
Share on other sites

On 10/15/2021 at 1:56 PM, TomXP411 said:

We've already talked about the ROM in other threads. Short version, there's no need to license Commodore's ROM. There are several 6502 BASICs out there, and there's more than enough brains in this community to crowdsource a kernel - even without relying on something like the open KERNAL started by the MEGA 65 crowd. 

This is all true. I was more addressing the USP for what X16 is today, but you're correct, it could be done without those dependencies.

Link to comment
Share on other sites

Answering what is the unique sale proposition (USP) of an FPGA only X16.

 

The others, good and bad points:

 

The C256 Foenix products, especially the A2560K, look and seem amazing on paper!

Downsides: Cost, user base and current available software.

 

Mega65. Stunning, if released early/mid 90’s.

Downsides: Cost. Tied to legacy hardware and backwards compatibility.

 

ZX Spectrum Next. If you are nostalgic for the Speccy, great! Possibly the best way of playing old Spectrum games, plus some new ones taking advantage of the new features.

Downsides: Cost OK, but a bit high. Graphics enhancements seem held back by backwards compatibility?

 

Colour Maximite 2. Very powerful. Cheap. Fast enough to run BASIC decently.

What about the user base, and software?

Main code and graphics are done in software on a CPU. There is a reason why graphics are done by separate chips/boards on most games consoles and PC’s.

 

Commander X16 FPGA USP’s:

Cost (potentially).

Not tied to backwards compatibility.

VERA – unique graphics chip.

Direct to the metal programming.

User base (hopefully!).

Already has some programs, games, software available, and its just an emulator at the moment!

Commander X16 FPGA could be an amazing product.

Link to comment
Share on other sites

From the thread where the question was posed:

On 10/15/2021 at 3:17 PM, BruceMcF said:

The thing is, the "unique value proposition" of the two are different.

The "unique value proposition" of the X16e is that it is hardware that lets you run X16 software out of the box, if you like, with a complete system costing somewhere on the order of half as much as the X16c. SBCs that can run the X16 emulator are already available in the $50-$100 price range, but if people really press the limits of the hardware, there could well still be some cool demos or games that don't run correctly or at full 60fps frame rate on the emulator.

The "unique value proposition" of the X8 is a small, inexpensive board that a single person can understand down "to the hardware specification", a la the C64 or Atari 800, of particular interest to those who already understand 6502 Assembly Language (and optionally also Forth) programming.

Relative to the FPGA Speccy+, the X8 is substantially cheaper, the X16e is a 6502 core while the Speccy is a Z80 core. So the X16e vs the Speccy+ (watchamacallit) is more like horizontal product differentiation, the X8 vs the Speccy+ is more like vertical differentiation.

As far as what you can DO with the X8 ... well, I spilled likely over 1,000 words on the simple proposition that if it had a spare select line for using the existing SPI interface to access a "hat", and brought select, SCLK, MOSI and MISO out to the existing debug interface pinout, to my mind it would be able to do a lot more interesting things.

Also, the value proposition of the X16e depends much more directly on following up the X16p and X16c, which are the anchors for the development of the "interesting X16 software" that drives the X16e value proposition.

Link to comment
Share on other sites

On 10/15/2021 at 1:08 PM, Janne Sirén said:

... If the X16 is turning into more of a simple FPGA board, what is unique about it that isn't already out there? For example, here is the SiDi FPGA from the ZX Spectrum community: ...

Note that there is a false premise here. The X16 is not "turning into more of a simple FPGA board" ... first, the FPGA board "X16e" version was on the "development plan" long before the X8 question was raised in The Megathread, and second, it was quite explicit before and it was made quite explicit in the "Megathread" that the "simple FPGA board" system is not taking the place of the X16p.

 

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
 Share

×
×
  • Create New...

Important Information

Please review our Terms of Use