Contaced a Factory in China about producing a run of fully populated Mini-migs

[downix]

Crom00 wrote:

... but you'd never get any games.

That is why I think that for version 2.0 it is important to add a floppy drive connector (and maybe a usb connector for a cd-rom).

At the very least the next feature added should be UAE hard file support. If that could be added to the current board that would be ideal!!!!

Imagine a 2 gig MMC with a 1 gig hdrfile tuned for Mini-MIG?  Joygasm! Or better yet, a minimig on a 32" lcd? IN VGA no less... NICE.

[/quote]
MMC's max at 128, hence why they made SD.

And I have a design for a hard drive, but I'm not finished with the PCB for it.

[jdiffend]

little wrote:

are you mad.

Quite :-o

dvd-burner sata ddrive for £20 ($40)

AFAIK there are no cd-roms for SATA and  I can get an IDE CD-ROM for about $13 USD. When I talk cheap I mean it :-D

I OWN a SATA DVD burner so I know they exist.  

Oh wait... you wanted CD only.  Eeeewwww... who would bother?  Hardly anyone would want one.

[asian1]

SATA CD-ROM:

From LiteOn: LTN-52S1S-10

http://www.newegg.com/Product/Product.aspx?Item=N82E16827106038

My concern:
1. According to Freescale 68K is no longer available. All new products and design should avoid this product.

2. Possible lawsuits from Amiga Inc, Acer, Gateway.

Is it possible to create all-in-one Structured ASIC or Easic (e-beam) based on Minimig design?

Is this approach better for mass production (10 thousands units)?

[AJCopland]

asian1 wrote:

My concern:
1. According to Freescale 68K is no longer available. All new products and design should avoid this product.

2. Possible lawsuits from Amiga Inc, Acer, Gateway.

Is it possible to create all-in-one Structured ASIC or Easic (e-beam) based on Minimig design?

1. The 68k 3.3v is still in production and is still used in
a lot of products, just not so many desktop/consoles, see: Freescale 68000

2. I wish someone would actually dispell this idea that someone could be sued for the MiniMig hardware, there is NOTHING that is owned by anyone that the hardware could be sued for using. For the actual binaries it might just be possible but even then I don't actually see a legal mechanism since reverse engineering IS LEGAL.

You're idea of producing an ASIC based on the MiniMig at the moment is probably an easier legal target than we currently have as the MiniMig hardware at the moment is just a reprogrammable device.

Andy

[ajlwalker]

MMC's max at 128, hence why they made SD.

Er, MMCs max at 128Gb which happens to be the same for SD.

[alexh]

AJCopland wrote:
You're idea of producing an ASIC based on the MiniMig at the moment is probably an easier legal target than we currently have as the MiniMig hardware at the moment is just a reprogrammable device.

I'm not sure that is true. The Amiga OCS patents have expired and that was the protection of the ideas about making a chip with these capabilities. The original chips themselves are no longer protected http://en.wikipedia.org/wiki/Mask_work

Wikipedia wrote:
According to 17 U.S.C. § 904, rights in semiconductor mask work last only two years (if unregistered) or ten years (if registered).

This means that you could make an ASIC Amiga, say you copied the original mask work (honest guv!) and there is nothing that the original manufacturers can do as it is more than 10 years after they were made.

The copyright issue I previously brought up applies only to the translation of the HRM from English into Verilog. That is a copyright issue and copyrights last 70-95 years.

I think an ASIC (and maybe even an FPGA binary) would be beyond the scope of copyright. To be sure, some legal would have to read the 1996 WIPO treaty & the Berne convention on copyright.

[Hattig]

I'm in two minds about creating an ASIC...

For a mass produced "Amiga in a Joystick" device, yes, it will cut costs drastically, and I'm sure that the 68k would also be incorporated into the ASIC if this was done (I don't see it happening soon).

However the programmable nature of the current design appeals to me. I like the idea of eventually being able to emulate pretty much any 8-bit computer on the FPGA directly, and 68k based systems as well (although the Atari ST ran at 8MHz, so the MiniMig design would need to take that into account somehow because it is clocked slightly slower - that or clock it at 14.3MHz by default).

[little]

Is it possible to create all-in-one Structured ASIC or Easic (e-beam) based on Minimig design?

Yes it would, but IMO it would be a bad idea:

1. The FPGA allows for bug-fixing. ATM the minimig OCS is not 100% compatible with the one in the amiga 500, as more people experienced in verilog get their minimigs this situation will improve, gradually.

2. The FPGA allows upgrading. Chances are that people will also work in an ECS and AGA implementation (both are 16 bits chipsets) and maybe even an AGA+.

3. The FPGA allows versatility. The minimig design allows for the emulation of other 68k based computers, consoles (this would require more ram) and arcade games (this would require even more ram).

[little]

although the Atari ST ran at 8MHz, so the MiniMig design would need to take that into account somehow because it is clocked slightly slower

If I undestood correctly from other posts in this board, the 68000 used in the minimig is the 20 mhz version, so atm it is underclocked to 7 mhz but the clock can be changed to suit diferent systems and even overclocked to 28mhz.

[freqmax]

The DCM can be used to accomplish 8 MHz or whatever other computers need without any hardware modifications. Just take a look at clock_dcm.v
AGA is 32-bit asfaik.

[alexh]

Hattig wrote:
I'm in two minds about creating an ASIC...

Dont worry you wont be doing it, at between £30,000 to £1,000,000 NRE per ASIC only a large company will be looking at this.

[asian1]

> NRE 1 Million GBP

What about easic:

From Design-reuse.com:
"eASIC Corporation, a provider of configurable Structured ASIC technology, today introduced a maskless customization approach aimed at eliminating NRE (non recurring engineering) cost for ASIC. The NRE and mask-set cost are removed since eASIC employs direct-write e-Beam approach rather than conventional mask lithography for IC customization. Due to the company’s innovative Via-customization technique, eASIC’s fabric yields about 10 times higher throughput from Direct-write e-Beam machines, compared to metal customization. This is made possible as Vias occupy about 1% area of the customization layer, while metal occupies at least 30%, thus reducing the time e-Beam machine needs to spend writing the customization. Moreover, only a single Via-layer is required for Structured eASIC customization, which further shortens the turnaround time and eventually cuts the cost."

[AJCopland]

alexh wrote:

AJCopland wrote:
You're idea of producing an ASIC based on the MiniMig at the moment is probably an easier legal target than we currently have as the MiniMig hardware at the moment is just a reprogrammable device.

I'm not sure that is true. The Amiga OCS patents have expired and that was the protection of the ideas about making a chip with these capabilities. The original chips themselves are no longer protected Wikipedia:Mask_work

Cheers AlexH I hadn't realised most of that. Now is anyone on here one of those legal types? :-D

Andy

[Hattig]

alexh wrote:
Dont worry you wont be doing it, at between £30,000 to £1,000,000 NRE per ASIC only a large company will be looking at this.

Of course, that's why I think it is a long way off, if ever. Plenty of time for debugging, adding ECS, maybe even AGA.

[little]

AGA is 32-bit asfaik.

Apart from the graphics data fetches, AGA still operated on 16-bit data only, meaning that a lot of bandwidth was wasted during register accesses and copper and blitter


Of course only AGA games/applications that used the 68EC020 as a 68000 would have any chances of working at all, at least that is how I undestand it.

[ChrisH]

@Crom00
If you want to avoid wasting a LOT of money, with only a load of NON-working MiniMigs to show for it, then I think you need to be REALLY careful:

You need to double-check EVERY piece of info that you give to the manufacturer, because one wrong bit will mean nothing will work.  More specifically, I think that you need the help of someone who has already built a working MiniMig themselves, and use the BOM that they used.  Or else make your own BOM, but then get someone to use that to make one MiniMig first (as a test).

Even if you do that, you still need to provide an IDIOT-PROOF means of testing whether the boards are working, or else you may get a nasty surprise when you receive them.  You have to remember that the factory guys will have *zero* idea what the MiniMig is (or what it is supposed to do), and probably they won't care either (beyond a step-by-step list of instructions).

And even if you do all that, *if* you haven't dealt with these sorts of manufacturers before, then I'd rate your chances of getting some working MiniMigs as being pretty low.  :-(  But I wish you good luck - you're going to need it!