[Coco] Embedded coco
jdaggett at gate.net
jdaggett at gate.net
Thu Dec 18 14:27:52 EST 2003
On 17 Dec 2003 at 22:40, jimcox at miba51.com wrote:
> James:
>
> Regarding the 68K - 6809 mode/emulation stuff. I just
> glanced over the emails and it appears I missed a lot of
> stuff.
>
> Regarding the HC12, couldn't a small FPGA be created to
> augment HC12 and provide resources that it would need to
> make porting NitrOS-9 to it?
>
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
One has to be very carefull as to redoing the HC12. There are patent infringement
issues that Motorola will not be to happy about. Now the CPU11 core used in the
HC11 line has a few opensource cores out there now. I think this is possible as the
HC11 CPU core now has passed the patent expiration point. Not sure about that.
Putting another register in the bank of register is easy. In ASIC that is real easy and
does not take up much logic or die space. Microcode/Control logic is another thing.
Now what could be done and may not violate any patents is to strip out some of the
microcode for the fuzzy logic and replace it with other instructions and add more
registers may be different enough to be outside the patents.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> I'm going to check with some of the ASIC engineers at work
> and see if they know of any open source tools. If there
> is I'll post them to the list, since I'm just getting back
> into things and need to ramp up.
>
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
There are tools that are free. Xilinx and Lattice and Altera all have limited version
tools to work witheir own FPGAs. Also there are a few free HDL editors and
simulators. Synthesis is still requires the FPGA librariesa and tools. Same for any
ASIC.
James
> Jim
>
> On Wed, 17 Dec 2003 22:45:15 -0500
> jdaggett at gate.net wrote:
> >jim
> >
> >I found the 68B09E on Ebay. Three tubes for, 27 parts
> >all 1999 date codes for $35
> >shipped. The 68B21's also on Ebay a tube of 9 of them for
> >$15 plus shipping. The
> >20v10 and the 16V8's I also got on Ebay at slightly under
> >half of what a new part
> >costs.
> >
> >As for a 6809 in a FPGA there are two opensource versions
> >out there. One is a
> >strict behavioral model and does not synthesize very well
> >into a Xilinx Part. It takes
> >up 95% of a SpartanII/IIe 200K gate part. There is
> >another core that has other part
> >to it as a SOC that syntesizes into a Spartan II/IIe 300K
> >gate. If you strip out just the
> >6809 cpu core that John Kent has done it synthesizes into
> >Xilinx 200K gate Spartan
> >II/IIe and occupies 90% of that.
> >
> >I have looked ot the VHDL code of both models and and
> >there is work that can be
> >done on both to make it synthesize more efficiently. The
> >John Kent code runs only
> >at about 10MHz clock. The other I think might run at
> >15MHz. I have not simulated
> >that or tried any of the testbenches to see if it works
> >and at what speed.
> >
> >There is a commercial 6809 software compatible core that
> >company's name elludes
> >me now. I have no idea of what they want for the core
> >code. They offer the core
> >code or they will synthesize it to what ever FPGA or ASIC
> >you want. They clain 25
> >MHz speed with 25 nS ram.
> >
> >With modern silicon it would not be impossilbe to get 40
> >to 80 Mhz speed in ASIC.
> >
> >As for porting OS9 to the HC12 is not that to far fetched
> >an idea. What quickly
> >comes to mind is that the HC12 does not have the U
> >pointer register, access to the
> >DP register as the 6809, and does not do the pre/post
> >double decrement/increment
> >on the pointer registers. There is the LEAx instruction
> >and a bunch of other
> >instructions that are nice.
> >
> >Just looking at the sources in an ide written for the
> >HC11/12 it the 6809 code is not
> >highlighted and therefore easy to see pnuemonics that are
> >not the same.
> >
> >The main problem in using XLATE09, which is a Motorola
> >freeware program to
> >convert 6809 code to 68K code, is that it is 95%
> >efficient at best in converting code.
> >You will have to do a lot to cleanup. It to is doable.
> >There is no 6809 emmulation
> >mode in the 68K or CPU32 core. One can write software to
> > emualte a 6809 on the
> >68K processors.
> >
> >james
> >
> >
> >On 17 Dec 2003 at 16:40, jimcox at miba51.com wrote:
> >
> >> James:
> >>
> >> Given the age of a lot of these parts, how much trouble
> >> have you had finding them? Maybe a FPGA version would
> >>be
> >> more of what is needed or take another chip such as the
> >> HC12 and port NitrOS-9 over to it. Someone was talking
> >>of
> >> porting it over to the 68000 or at least running a 68000
> >> in 6809 mode (never heard of that before the recent
> >> conversation)
> >>
> >> Jim
> >>
> >> On Wed, 17 Dec 2003 19:36:18 -0500
> >> jdaggett at gate.net wrote:
> >> >Gee whiz
> >> >
> >> >Sounds like something I was thinking of for a project.
> >>I
> >> >was going to start with a
> >> >CoCo3 and then maybe into a striped down version. Maybe
> >> >still able to run a
> >> >stripped down version of OS9 in flash and a small to
> >> >medium LCD display.
> >> >
> >> >A couple of questions:
> >> >
> >> >1) what kind of ram and how much?
> >> >2) Which PCB layout tool you plan to use?
> >> >3) Will you include the current expansion port of the
> >> >CoCo?
> >> >4) Total Ram and Rom on board?
> >> >5) Any provisions for SRAM in surface mount package
> >> >6) Resistors and caps in Surface mount or thru hole?
> >> >
> >> >I have been accumalating parts now for about three
> >>months
> >> >and looking at some
> >> >software.
> >> >
> >> >james
> >> >
> >> >On 17 Dec 2003 at 15:06, peak at mail.polarcomm.com wrote:
> >> >
> >> >> About using a coco mother board as an embedded
> >> >>controller!
> >> >> How many people would be interested in a mini coco?
> >> >> It could/would be on a small ckt board and would
> >>have:
> >> >> cpu(6809)
> >> >> rom
> >> >> ram
> >> >> a PIA or two (address decoded to the exact same
> >> >>locations as
> >> >> a real coco.
> >> >> DAC
> >> >> Small PAL/PLD used as an address decoder. Both the
> >>CTS,
> >> >>and
> >> >> the SCS lines would also be duplicated.
> >> >>
> >> >> WhY?
> >> >> Because by using the same hardware memory map for the
> >> >>PIA's
> >> >> and DAC would allow a much smaller controller . also
> >>a
> >> >>real
> >> >> cocoI or II would not have to be sacrificed to use as
> >> >>the
> >> >> controller. But a real coco could be used to test and
> >> >>debug
> >> >> software before it was placed in rom on the embed
> >>board!
> >> >>
> >> >> If anyone is interested I will design the ckt board
> >>and
> >> >>place
> >> >> the design in public domain or even post it here if
> >> >>allowed.
> >> >> Heck I may design this thing anyway just for myself.
> >>Any
> >> >> takers??
> >> >>
> >> >> Eric
> >> >>
> >> >> --
> >> >> Coco mailing list
> >> >> Coco at maltedmedia.com
> >> >> http://five.pairlist.net/mailman/listinfo/coco
> >> >
> >> >
> >> >
> >> >--
> >> >Coco mailing list
> >> >Coco at maltedmedia.com
> >> >http://five.pairlist.net/mailman/listinfo/coco
> >>
> >>
> >> --
> >> Coco mailing list
> >> Coco at maltedmedia.com
> >> http://five.pairlist.net/mailman/listinfo/coco
> >
> >
> >
> >--
> >Coco mailing list
> >Coco at maltedmedia.com
> >http://five.pairlist.net/mailman/listinfo/coco
>
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco
More information about the Coco
mailing list