[Coco] Any news on the so called CoCo4 or NextCoCoprojectthatBjork was heading?

[Little John]

Weighing in late or not... I still enjoy reading your posts and I have 
followed with interest your posts relating to your work on a replacement 
G.I.M.E.
The idea of multiple boards, I think, is a good one. Sort of like the old 
S-100 and SS-50 days, we could have a small motherboard that is mostly just 
the motherboard. The CPU and GIME cards could plug in to make a system, but 
better than the old S-100, etc. days, the plug in cards would remain 
completely reconfigurable. That would be a nice little set-up :)
-JT
----- Original Message ----- 
From: <jdaggett at gate.net>
To: "CoCoList for Color Computer Enthusiasts" <coco at maltedmedia.com>
Sent: Wednesday, November 17, 2010 9:47 AM
Subject: Re: [Coco] Any news on the so called CoCo4 or 
NextCoCoprojectthatBjork was heading?


>I maybe weighing on this a bit late due to to many other commitments and 
>irons in the fire.
>
>>From a hardware engineer's point of view I see a COCO4 or what ever it is 
>>to be called
>
> 1) FPGA route is pretty much the modern way to go. Saying that there 
> probably is no one
> board out there that is 100% everything needed to do a COCO4.
>
> 2) Software emmulation still requires a board of some type. Whether PC 
> base or FPGA
> based. Either case the cost will be upwards to $400+.
>
> 3) A dedicated FPGA board just for our needs may sound real nice but also 
> has it flaws. One
> is price. A second is a potential to lock into a design that may not be as 
> flexible in the near
> future. Therefore, bocomes to costly to upgrade or modify.
>
>
> So what to do?
>
> One suggestion that I would throw out that has already been suggested. 
> Instead of one board
> do all have several boards that plug into a main board. I would entertain 
> a suggestion that
> there be a main board about the size of the Arduino Mega (2.1 inch by 3.9 
> inch). The one
> constant that proliferates any design is the CPU, (John Kent's CPU09). 
> This can sit in one
> small FPGA with enough IO to interface to other boards. A GIME chip 
> stripped of the video
> section takes up 50 slices of a XC3S250E, 85 FFs and 81 4-LUTs. That is 
> coded but
> untested. It can somewhat be ported to Alterra if needed.
>
> John Kent's CPU09 utlizes 53% of the slices of a XC3S250E, 51% of the 
> 4-LUTs and 4% of
> the Fs. That leaves plenty of room for the GIME chip and some other 
> circuitry. To split
> functionality up into two or more FPGAs is not all that more in cost of 
> silicon. The major cost
> adders will be PC Boards and how many. I just processed a set of five 
> boards through Batch
> PCB for another project I am doing and the quality is as good as any I 
> have seen in my past
> jobs. Cost is reasonable, about $2.50 per square inch in single quantity 
> two layer with solder
> mask and top silk screen. All holes are plated. The five boards totalled 
> ~26square inches and
> cost me far less than $100 shipped. Process time is slow, 15 to 20 
> business days. There is
> also a means of distribution of any board(s) design through their market 
> place.
>
> Again I like the idea of the Arduino style with adaptor boards or shields 
> or wings or what ever
> you wish to call them. A design can be some what compact and a design that 
> is as flexible or
> more flexible that some development boards.
>
> just my thoughts.
> james
>
>
>
> On 20 Oct 2010 at 21:31, Paul Fitch wrote:
>
>> I think the FPGA route is the only realistic method available to do this
>> in hardware.  I'm just not that interested in a hardware project.  Doing
>> it in emulation (the Coco4) however, has had me wishing very hard that I
>> could program at that level.  I just don't see spending hundreds of
>> dollars on duplicating hardware that in most any matchup would be 
>> inferior
>> to the stuff found on every bargin basement Windows 7 starter computer
>> available today for under $400.00.  And that's just the brand new stuff.
>
>
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco