Loading ...
Sorry, an error occurred while loading the content.

Re: [cosmacelf] Re: CDP1806ACE

Expand Messages
  • Chuck Bigham
    When I simulated the Johnson counter that drives the internal state machine, there were 16 distinct states generated by the counter. I was surprised by that,
    Message 1 of 25 , Apr 12, 2013
    • 0 Attachment
      When I simulated the Johnson counter that drives the internal state machine, there were 16 distinct states generated by the counter. I was surprised by that, but I haven’t had time to simulate the rest of the control circuitry to see how the states are used.
       
      Chuak
       
      From: Lee Hart
      Sent: ‎Thursday‎, ‎April‎ ‎11‎, ‎2013 ‎9‎:‎22‎ ‎PM
      To: cosmacelf@yahoogroups.com
       
       

      David Keith wrote:

      > Just a quick reply, use an internal freq doubler...

      That could work. Though I think they found it simpler to just run the
      oscillator at a high enough frequency so simply dividing it down was
      sufficient.

      William Donnelly wrote:
      > That quoted comment doesn't really make sense to me. if the
      > instruction cycle is 8, how does it do all of the other things it needs to
      > do, like getting the value(s), etc., and then storing the results, as
      > well as perform an 8 x 1-bit ALU. It looks like you need more cycles.

      As many things as possible are done in parallel. In most cases, this is
      easy, like loading all 16 bits of a register at once. In other cases,
      they can be allowed to overlap. For example, if we tell the ALU to add
      two numbers,

      - The Fetch cycle (8 clock cycles) gets the instruction.
      Aha; it is ADD.
      - The Execute cycle (8 clock cycles) puts the memory address
      out on the bus, and sets /MRD low to read. The byte
      won't be back from memory until the end of this cycle.
      - The *next* Fetch cycle (8 clock cycles) begins. While it
      is getting the next instruction, the ALU has all 8
      clock cycles to serially add D and the memory byte.
      - The next Execute cycle thus begins with the ALU result
      complete and available for use.

      8 clock cycles is enough to add two 8-bit binary numbers with a 1-bit
      serial adder, because it can do it during the Fetch time of the
      following instruction.

      --
      Ring the bells that still can ring
      Forget your perfect offering
      There is a crack in everything
      That's how the light gets in.
      -- Leonard Cohen, from "Anthem"
      --
      Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm

    • David Keith
      I admit,   I was quite taken back when this 1802 ALU serialization was brought back into the light again.  But future implementations of the 1802 Instruction
      Message 2 of 25 , Apr 12, 2013
      • 0 Attachment
        I admit,   I was quite taken back when this 1802 ALU serialization was brought back into the light again.  But future implementations of the 1802 Instruction Set does not have to use this technique.  I proposed the ISA committee so we as a group of 1802 assembler programmers can have a common future version to use and share program.  There are many ways to implement it, and several have already been done by this group. I believe even Lee suggested implementing an 1802 using fet transistors :-)  a month or so back.  But what matters is that we have a standard expanded set of instructions that would ease certain programming issues. 

        -- David Keith



        From: Chuck Bigham <chuck@...>
        To: "cosmacelf@yahoogroups.com" <cosmacelf@yahoogroups.com>
        Sent: Friday, April 12, 2013 2:18 PM
        Subject: Re: [cosmacelf] Re: CDP1806ACE



        When I simulated the Johnson counter that drives the internal state machine, there were 16 distinct states generated by the counter. I was surprised by that, but I haven’t had time to simulate the rest of the control circuitry to see how the states are used.
         
        Chuak
         
        From: Lee Hart
        Sent: ‎Thursday‎, ‎April‎ ‎11‎, ‎2013 ‎9‎:‎22‎ ‎PM
        To: cosmacelf@yahoogroups.com
         
         
        David Keith wrote:
        > Just a quick reply, use an internal freq doubler...

        That could work. Though I think they found it simpler to just run the
        oscillator at a high enough frequency so simply dividing it down was
        sufficient.

        William Donnelly wrote:
        > That quoted comment doesn't really make sense to me. if the
        > instruction cycle is 8, how does it do all of the other things it needs to
        > do, like getting the value(s), etc., and then storing the results, as
        > well as perform an 8 x 1-bit ALU. It looks like you need more cycles.

        As many things as possible are done in parallel. In most cases, this is
        easy, like loading all 16 bits of a register at once. In other cases,
        they can be allowed to overlap. For example, if we tell the ALU to add
        two numbers,

        - The Fetch cycle (8 clock cycles) gets the instruction.
        Aha; it is ADD.
        - The Execute cycle (8 clock cycles) puts the memory address
        out on the bus, and sets /MRD low to read. The byte
        won't be back from memory until the end of this cycle.
        - The *next* Fetch cycle (8 clock cycles) begins. While it
        is getting the next instruction, the ALU has all 8
        clock cycles to serially add D and the memory byte.
        - The next Execute cycle thus begins with the ALU result
        complete and available for use.

        8 clock cycles is enough to add two 8-bit binary numbers with a 1-bit
        serial adder, because it can do it during the Fetch time of the
        following instruction.

        --
        Ring the bells that still can ring
        Forget your perfect offering
        There is a crack in everything
        That's how the light gets in.
        -- Leonard Cohen, from "Anthem"
        --
        Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm




      • William Donnelly
        I think of the expansion of the 1802 instruction set as mostly thought-crimes that are fun to discuss and rumentate about. But they could be fun and
        Message 3 of 25 , Apr 12, 2013
        • 0 Attachment
          I think of the expansion of the 1802 instruction set as mostly thought-crimes
          that are fun to discuss and rumentate about. But they could be fun and interesting
          to implement. Most likely on a one-off type of situation or so. I doubt if any of them
          would ever take off too much or become very popular.

          If I really wanted a better instruction set or faster CPU, I would use another, newer CPU,
          like an 8085, or one of the other microcontrollers, like Arduino, PIC, etc.

          – Bill


          On 4/12/2013 5:50 PM, David Keith wrote:
           
          {snip}

        • Lee Hart
          ... My guess is that if the instruction set were opened up to being changed by committee, no telling what kind of strange mess might result. It would also be
          Message 4 of 25 , Apr 12, 2013
          • 0 Attachment
            On 4/12/2013 7:50 PM, David Keith wrote:
            > I admit, I was quite taken back when this 1802 ALU serialization was
            > brought back into the light again. But future implementations of the
            > 1802 Instruction Set does not have to use this technique. I proposed the
            > ISA committee so we as a group of 1802 assembler programmers can have a
            > common future version to use and share program. There are many ways to
            > implement it, and several have already been done by this group.

            My guess is that if the instruction set were opened up to being changed
            by committee, no telling what kind of strange mess might result. It
            would also be likely to be incompatible with existing 1802 software. It
            would rapidly diverge into designing an entirely different CPU. That's
            not bad; but it's not an 1802.

            As an alternative, it might be more practical to think about ways that
            the 1802 hardware architecture could be expanded. It is possible to
            implement an 1802 with an FPGA. It is also possible that such an FPGA
            could have changes and improvements, while still being compatible with
            existing hardware and software.

            Here are some thoughts about an 180X architecture:

            1. 1802 opcode 68h is used as a "shift" for the 1804/5/6. How about
            adding some microcode RAM to the 180X, so you can *program* what
            the extended opcodes 68 00 thru 68 FF will do? The default would
            be to load them with the standard 1804 instructions. Or, you could
            invent new instructions that were specifically useful for the task
            at hand.

            The new instructions would not be fixed and in ROM; they could be
            loaded into the CPU at run time. This would be a wonderful tool
            for testing and experimenting with new instruction sets.

            2. There is no point in changing the serial ALU for a parallel one
            unless you're also going to reduce the clock cycles needed for
            *every* instruction. In the extreme, it would be possible have
            one clock cycle = one bus cycle. For example, the first half of
            the cycle (clock high) outputs the high address byte on MA0-7,
            and the second half (clock low) outputs the low byte on MA0-7.
            Even at the same clock speed, the 180X would be 8x faster.

            3. Pipelining. The 1802 has many bus cycles when nothing is going
            on with the bus. An INC Rx instruction takes 2 bus cycles, but
            the bus is idle during the second Execute cycle. Use synchronous
            registers that can be loaded, incremented, or decremented in a
            single clock cycle, and thus don't require a second execute
            cycle. The 180X could be fetching the next instruction instead.
            This would mean we'd have 1, 2, and 3-cycle instructions.

            4. Cache. It may also be practical to add a small cache. The cache
            gets loaded with recent memory data. If a branch instruction
            jumps to an address already in the cache, it can get the data
            from there, eliminating a bus cycle.

            5. Expanded I/O. Most modern micros use their registers as I/O
            ports. The 1802 architecture already makes it possible to "view"
            the internal 16-bit registers externally -- all it takes is an
            external latch to grab the register value from MA0-7 when any
            register access instruction is executed. A tiny bit of external
            hardware could make any register an output port.

            Expand on this concept, so you can do inputs as well as outputs.
            Some of the shifted 68 xx opcodes could *read* the address bus,
            and load it into a designated register. For example, the registers
            in an 8250 UART could appear as 180X internal registers, where
            they can be manipulated by the instruction set directly!

            These sorts of changes are in line with modern CPU chip design, yet
            don't change the instruction set. Only the timing changes; software
            timing loops may have to be tweaked, but otherwise older software should
            still work.

            --
            Never doubt that the work of a small group of thoughtful, committed
            citizens can change the world. Indeed, it's the only thing that ever
            has! -- Margaret Mead
            --
            Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm
          • country_robot
            ... There s a guy who did just that for the 65C02, though without an FPGA. He used techniques very similar to what I m proposing to use to make a 16-bit 1802:
            Message 5 of 25 , Apr 12, 2013
            • 0 Attachment
              --- In cosmacelf@yahoogroups.com, Lee Hart <leeahart@...> wrote:
              >
              > My guess is that if the instruction set were opened up to being changed
              > by committee, no telling what kind of strange mess might result. It
              > would also be likely to be incompatible with existing 1802 software. It
              > would rapidly diverge into designing an entirely different CPU. That's
              > not bad; but it's not an 1802.
              >
              > As an alternative, it might be more practical to think about ways that
              > the 1802 hardware architecture could be expanded. It is possible to
              > implement an 1802 with an FPGA. It is also possible that such an FPGA
              > could have changes and improvements, while still being compatible with
              > existing hardware and software. ...

              There's a guy who did just that for the 65C02, though without an FPGA. He used techniques very similar to what I'm proposing to use to make a 16-bit 1802:

              http://laughtonelectronics.com/arcana/Kimklone_short_summary.html

              -Bobby
            • David Keith
              Here is what I have been looking at to implement the 1802++ in a FPGA $48 +$5 shipping
              Message 6 of 25 , Apr 12, 2013
              • 0 Attachment
                Here is what I have been looking at to implement the 1802++ in a FPGA $48 +$5 shipping
                http://www.ebay.com/itm/251067943099?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649

                EP3C16: onboard FPGA device which features:
                Operating Frequency: 50MHz
                Operating Voltage: 1.5-3.3V
                Package: QFP240
                I/Os: 160
                LEs: 15408
                RAM: 504kb
                Multipliers: 56
                PLLs: 4
                Debugging/Programming: JTAG
                Board runs from a 5V supply, regulated down to 1.5-3.3




                From: country_robot <rlnansel@...>
                To: cosmacelf@yahoogroups.com
                Sent: Friday, April 12, 2013 10:57 PM
                Subject: [cosmacelf] Re: CDP1806ACE - ISA Standard committee



                --- In cosmacelf@yahoogroups.com, Lee Hart <leeahart@...> wrote:
                >
                > My guess is that if the instruction set were opened up to being changed
                > by committee, no telling what kind of strange mess might result. It
                > would also be likely to be incompatible with existing 1802 software. It
                > would rapidly diverge into designing an entirely different CPU. That's
                > not bad; but it's not an 1802.
                >
                > As an alternative, it might be more practical to think about ways that
                > the 1802 hardware architecture could be expanded. It is possible to
                > implement an 1802 with an FPGA. It is also possible that such an FPGA
                > could have changes and improvements, while still being compatible with
                > existing hardware and software. ...

                There's a guy who did just that for the 65C02, though without an FPGA. He used techniques very similar to what I'm proposing to use to make a 16-bit 1802:

                http://laughtonelectronics.com/arcana/Kimklone_short_summary.html

                -Bobby




                ------------------------------------

                ========================================================
                Visit the COSMAC ELF website at http://www.cosmacelf.comYahoo! Groups Links

                <*> To visit your group on the web, go to:
                    http://groups.yahoo.com/group/cosmacelf/

                <*> Your email settings:
                    Individual Email | Traditional

                <*> To change settings online go to:
                    http://groups.yahoo.com/group/cosmacelf/join
                    (Yahoo! ID required)

                <*> To change settings via email:
                    cosmacelf-digest@yahoogroups.com
                    cosmacelf-fullfeatured@yahoogroups.com

                <*> To unsubscribe from this group, send an email to:
                    cosmacelf-unsubscribe@yahoogroups.com

                <*> Your use of Yahoo! Groups is subject to:
                    http://docs.yahoo.com/info/terms/



              • urrossum@att.net
                ... ... ... All of these sound like great ideas. In my mind, one could keep verbatim the existing instruction set, but just speed it the heck up,
                Message 7 of 25 , Apr 12, 2013
                • 0 Attachment
                  > 2. There is no point in changing the serial ALU for a parallel one
                  > unless you're also going to reduce the clock cycles needed for
                  > *every* instruction. In the extreme, it would be possible have
                  > one clock cycle = one bus cycle. For example, the first half of
                  <..>
                  > 3. Pipelining. The 1802 has many bus cycles when nothing is going
                  > on with the bus. An INC Rx instruction takes 2 bus cycles, but
                  > the bus is idle during the second Execute cycle. Use synchronous
                  <..>
                  > 4. Cache. It may also be practical to add a small cache. The cache

                  All of these sound like great ideas. In my mind, one could keep verbatim the existing instruction set, but just speed it the heck up, much like Dallas Semi has done with the 8051. (Their 87C520 runs at 33 MHz, with nearly every instruction taking only one clock cycle. In other words, nearly 33 MIPS from an 8051!) How about a 50 MHz 1802, with every instruction only taking one clock cycle?

                  > 5. Expanded I/O. Most modern micros use their registers as I/O
                  > ports. The 1802 architecture already makes it possible to "view"
                  > the internal 16-bit registers externally -- all it takes is an
                  > external latch to grab the register value from MA0-7 when any
                  > register access instruction is executed. A tiny bit of external
                  > hardware could make any register an output port.

                  This is an interesting way to do it. One of the fundamental limitations of the 1802 for me is the very limited I/O space. Sure, you can get around it by using memory mapped I/O, but if one has the fire breathing monster described above, you're gonna need more I/O (so you can run a server farm of Linux machines on 1802s). I was just thinking of condensing the I/O instructions into a two-byte set with the second byte being the I/O port, but this is an interesting way to accomplish a lot of that without breaking the existing instruction set.
                  ~~
                  Mark Moulding
                • country_robot
                  I know this may seem esoteric, but I m operating under a different technical regime -- an aesthetic one, to be accurate. I want to create an
                  Message 8 of 25 , Apr 12, 2013
                  • 0 Attachment
                    I know this may seem esoteric, but I'm operating under a different technical regime -- an aesthetic one, to be accurate.  I want to create an 180X-upward-compatible design that a (perhaps gifted) computer hobbyist could have designed and built circa 1983-1984. I'm a sculptor, so the point of this is explicitly artistic, and I have a set of rules I'm following. The rules don't have to make sense for here and now, but they do have to make at least some sense for the target era.

                    In simplest terms, if the part wasn't in data books and available for sale to mere mortals in that period, then I'm not using it. The computer is going to be used to control a retro-robot sculpture, also designed under this same rule.

                    In the long term I'm sculpting a whole series of robots that I call (rather unimaginatively) Twentieth Century Robots, each of the robots set within what I find to be an interesting time period roughly corresponding to each decade of the twentieth century. Last year I built the mechanical shell of a retrobot set in the sixties, and next year I'm building one firmly set in the fifties (complete with electron valve electronics). This year I'm doing a robot straddling 1975-1985, hence the 1802 focus.

                    For each of these robots in this series I also create ancillary artefacts to go with the installation, things such as photos and documentation, even creating a fictional builder with a back story to explain why he/she went to all the trouble. The psychology of the builder and his/her goals are just as important as the technology involved.

                    Each robot will technically function properly, in some sense, but perhaps not as well as the fictional builder might have hoped.

                    Anyway, like I said, it's an aesthetic thing.

                    -Bobby


                    --- In cosmacelf@yahoogroups.com, David Keith <beloved_wind@...> wrote:
                    >
                    > I admit,   I was quite taken back when this 1802 ALU serialization was brought back into the light again.  But future implementations of the 1802 Instruction Set does not have to use this technique.  I proposed the ISA committee so we as a group of 1802 assembler programmers can have a common future version to use and share program.  There are many ways to implement it, and several have already been done by this group. I believe even Lee suggested implementing an 1802 using fet transistors :-)  a month or so back.  But what matters is that we have a standard expanded set of instructions that would ease certain programming issues.  
                    > -- David Keith
                  • William Donnelly
                    I like the sound of that. -- Bill http://groups.yahoo.com/group/cravcomp/ ... {snip}
                    Message 9 of 25 , Apr 12, 2013
                    • 0 Attachment
                      I like the sound of that.

                      On 4/12/2013 9:34 PM, country_robot wrote:
                       

                      I know this may seem esoteric, but I'm operating under a different technical regime -- an aesthetic one, to be accurate.  I want to create an 180X-upward-compatible design that a (perhaps gifted) computer hobbyist could have designed and built circa 1983-1984. I'm a sculptor, so the point of this is explicitly artistic, and I have a set of rules I'm following. The rules don't have to make sense for here and now, but they do have to make at least some sense for the target era.

                      In simplest terms, if the part wasn't in data books and available for sale to mere mortals in that period, then I'm not using it. The computer is going to be used to control a retro-robot sculpture, also designed under this same rule.

                      In the long term I'm sculpting a whole series of robots that I call (rather unimaginatively) Twentieth Century Robots, each of the robots set within what I find to be an interesting time period roughly corresponding to each decade of the twentieth century. Last year I built the mechanical shell of a retrobot set in the sixties, and next year I'm building one firmly set in the fifties (complete with electron valve electronics). This year I'm doing a robot straddling 1975-1985, hence the 1802 focus.

                      For each of these robots in this series I also create ancillary artefacts to go with the installation, things such as photos and documentation, even creating a fictional builder with a back story to explain why he/she went to all the trouble. The psychology of the builder and his/her goals are just as important as the technology involved.

                      Each robot will technically function properly, in some sense, but perhaps not as well as the fictional builder might have hoped.

                      Anyway, like I said, it's an aesthetic thing.

                      -Bobby

                      {snip}

                    • country_robot
                      Thanks. It s to be sort of a series of love letters to hackers of the 20th century, their triumphs as well as their foibles. -Bobby
                      Message 10 of 25 , Apr 12, 2013
                      • 0 Attachment
                        Thanks. It's to be sort of a series of love letters to hackers of the 20th century, their triumphs as well as their foibles.

                        -Bobby

                        --- In cosmacelf@yahoogroups.com, William Donnelly <william@...> wrote:
                        >
                        > I like the sound of that.
                        >
                        > -- Bill
                      • Lee Hart
                        ... That s an interesting approach. He added a coprocessor to a 6502. The coprocessor detected unused and illegal 6502 opcodes, and took over execution of
                        Message 11 of 25 , Apr 13, 2013
                        • 0 Attachment
                          country_robot wrote:
                          > There's a guy who did just that for the 65C02, though without an FPGA. He used techniques very similar to what I'm proposing to use to make a 16-bit 1802:
                          > http://laughtonelectronics.com/arcana/Kimklone_short_summary.html

                          That's an interesting approach. He added a coprocessor to a 6502. The
                          coprocessor detected unused and illegal 6502 opcodes, and "took over"
                          execution of them.

                          On an 1802, this would mean adding hardware that detected a Fetch 68h,
                          fed the 1802 something else instead (like a NOP), and then used the
                          resulting 3 cycles to do some new operation.

                          You'd have to get really intimate with how the 1802 works at the
                          hardware level, and "channel" Weisbecker" to pull this off. But it does
                          seem promising. It would be fairly hard to change the internal 1802
                          registers, but you certainly could add new external registers and
                          manipulate them extensively.
                          --
                          Failure is only the opportunity to begin again more intelligently.
                          -- Henry Ford
                          --
                          Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm
                        • Lee Hart
                          ... Wow. Only to implement *one* 1802, or a dozen? Or an 1802 and a dozen other chips as well? -- There are few industries with more BS than the battery
                          Message 12 of 25 , Apr 13, 2013
                          • 0 Attachment
                            David Keith wrote:
                            > Here is what I have been looking at to implement the 1802++ in a FPGA
                            > $48 +$5 shipping
                            > http://www.ebay.com/itm/251067943099?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649

                            Wow. Only to implement *one* 1802, or a dozen? Or an 1802 and a dozen
                            other chips as well?

                            --
                            There are few industries with more BS than the battery industry.
                            Elon Musk
                            --
                            Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm
                          • William Donnelly
                            That is pretty interesting. The way to modify anything internal to the 1802 is to feed it instructions to execute. Like LDI and PHI/PLO for registers, etc.
                            Message 13 of 25 , Apr 13, 2013
                            • 0 Attachment
                              That is pretty interesting.

                              The way to modify anything internal to the 1802 is to feed it instructions to execute.
                              Like LDI and PHI/PLO for registers, etc. Then fix the D register or anything changed.
                              You would have to reset the PC, but even that would just be a short or long jump
                              back to where it was supposed to be for its next instruction. You don't necessarily
                              even have to go along with the 1802's signals, You could put it in a wait state until
                              you are done. It depends on what you need to do.

                              It would be an interesting project.

                              Also, check out this page: The Diablo (printer) Proprietary CPU
                                  http://laughtonelectronics.com/arcana/DiabloCPU.html

                              His website is really nice. Nicely designed and well-implemented, where everything
                              is logical and easily seen and found. (more so than too many sites)

                              – Bill

                              On 4/13/2013 11:25 AM, Lee Hart wrote:
                               

                              country_robot wrote:
                              > There's a guy who did just that for the 65C02, though without an FPGA. He used techniques very similar to what I'm proposing to use to make a 16-bit 1802:
                              > http://laughtonelectronics.com/arcana/Kimklone_short_summary.html

                              That's an interesting approach. He added a coprocessor to a 6502. The
                              coprocessor detected unused and illegal 6502 opcodes, and "took over"
                              execution of them.

                              On an 1802, this would mean adding hardware that detected a Fetch 68h,
                              fed the 1802 something else instead (like a NOP), and then used the
                              resulting 3 cycles to do some new operation.

                              You'd have to get really intimate with how the 1802 works at the
                              hardware level, and "channel" Weisbecker" to pull this off. But it does
                              seem promising. It would be fairly hard to change the internal 1802
                              registers, but you certainly could add new external registers and
                              manipulate them extensively.
                              --
                              Failure is only the opportunity to begin again more intelligently.
                              -- Henry Ford
                              --
                              Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm

                            • David Keith
                              Very Interesting Bobby, I too am interested in Retro Vintage Logic, I have a long term project to rebuild the ALU logic in the Atanasoff–Berry Computer
                              Message 14 of 25 , Apr 13, 2013
                              • 0 Attachment
                                Very Interesting Bobby,

                                I too am interested in Retro Vintage Logic, I have a long term project to rebuild the ALU logic in the Atanasoff–Berry Computer (ABC). It is mainly Resistor-Triode Logic, but I have created some retro-50s artwork that you may like, it is either 8 AOI Gates or 4 FFs with Neon lamp output, depending on how I place the components.

                              • David Keith
                                The 1802++ will be my first soft core, small steps :-)   We have plenty of room to expand the ISA.  David Keith ________________________________ From: Lee
                                Message 15 of 25 , Apr 13, 2013
                                • 0 Attachment
                                  The 1802++ will be my first soft core, small steps :-)   We have plenty of room to expand the ISA. 

                                  David Keith



                                  From: Lee Hart <leeahart@...>
                                  To: cosmacelf@yahoogroups.com
                                  Sent: Saturday, April 13, 2013 2:26 PM
                                  Subject: Re: [cosmacelf] Re: CDP1806ACE - ISA Standard committee

                                  David Keith wrote:
                                  > Here is what I have been looking at to implement the 1802++ in a FPGA
                                  > $48 +$5 shipping
                                  > http://www.ebay.com/itm/251067943099?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649

                                  Wow. Only to implement *one* 1802, or a dozen? Or an 1802 and a dozen
                                  other chips as well?

                                  --
                                  There are few industries with more BS than the battery industry.
                                      Elon Musk
                                  --
                                  Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm


                                  ------------------------------------

                                  ========================================================
                                  Visit the COSMAC ELF website at http://www.cosmacelf.comYahoo! Groups Links

                                  <*> To visit your group on the web, go to:
                                      http://groups.yahoo.com/group/cosmacelf/

                                  <*> Your email settings:
                                      Individual Email | Traditional

                                  <*> To change settings online go to:
                                      http://groups.yahoo.com/group/cosmacelf/join
                                      (Yahoo! ID required)

                                  <*> To change settings via email:
                                      cosmacelf-digest@yahoogroups.com
                                      cosmacelf-fullfeatured@yahoogroups.com

                                  <*> To unsubscribe from this group, send an email to:
                                      cosmacelf-unsubscribe@yahoogroups.com

                                  <*> Your use of Yahoo! Groups is subject to:
                                      http://docs.yahoo.com/info/terms/



                                • David Keith
                                  email truncated the embedded image, this time with the image attached ________________________________ From: David Keith To:
                                  Message 16 of 25 , Apr 13, 2013
                                  • 1 Attachment
                                  • 124 KB
                                  email truncated the embedded image, this time with the image attached


                                  From: David Keith <beloved_wind@...>
                                  To: "cosmacelf@yahoogroups.com" <cosmacelf@yahoogroups.com>
                                  Sent: Saturday, April 13, 2013 7:42 PM
                                  Subject: Re: [cosmacelf] Re: CDP1806ACE - ISA Standard committee

                                  Very Interesting Bobby,

                                  I too am interested in Retro Vintage Logic, I have a long term project to rebuild the ALU logic in the Atanasoff–Berry Computer (ABC). It is mainly Resistor-Triode Logic, but I have created some retro-50s artwork that you may like, it is either 8 AOI Gates or 4 FFs with Neon lamp output, depending on how I place the components.


                                  There is Art in all things
                                  Dave


                                  From: country_robot <rlnansel@...>
                                  To: cosmacelf@yahoogroups.com
                                  Sent: Saturday, April 13, 2013 12:34 AM
                                  Subject: [cosmacelf] Re: CDP1806ACE - ISA Standard committee



                                  I know this may seem esoteric, but I'm operating under a different technical regime -- an aesthetic one, to be accurate.  I want to create an 180X-upward-compatible design that a (perhaps gifted) computer hobbyist could have designed and built circa 1983-1984. I'm a sculptor, so the point of this is explicitly artistic, and I have a set of rules I'm following. The rules don't have to make sense for here and now, but they do have to make at least some sense for the target era.

                                  In simplest terms, if the part wasn't in data books and available for sale to mere mortals in that period, then I'm not using it. The computer is going to be used to control a retro-robot sculpture, also designed under this same rule.

                                  In the long term I'm sculpting a whole series of robots that I call (rather unimaginatively) Twentieth Century Robots, each of the robots set within what I find to be an interesting time period roughly corresponding to each decade of the twentieth century. Last year I built the mechanical shell of a retrobot set in the sixties, and next year I'm building one firmly set in the fifties (complete with electron valve electronics). This year I'm doing a robot straddling 1975-1985, hence the 1802 focus.

                                  For each of these robots in this series I also create ancillary artefacts to go with the installation, things such as photos and documentation, even creating a fictional builder with a back story to explain why he/she went to all the trouble. The psychology of the builder and his/her goals are just as important as the technology involved.

                                  Each robot will technically function properly, in some sense, but perhaps not as well as the fictional builder might have hoped.

                                  Anyway, like I said, it's an aesthetic thing.

                                  -Bobby


                                  --- In cosmacelf@yahoogroups.com, David Keith <beloved_wind@...> wrote:
                                  >
                                  > I admit,   I was quite taken back when this 1802 ALU serialization was brought back into the light again.  But future implementations of the 1802 Instruction Set does not have to use this technique.  I proposed the ISA committee so we as a group of 1802 assembler programmers can have a common future version to use and share program.  There are many ways to implement it, and several have already been done by this group. I believe even Lee suggested implementing an 1802 using fet transistors :-)  a month or so back.  But what matters is that we have a standard expanded set of instructions that would ease certain programming issues.  
                                  > -- David Keith






                                • William Donnelly
                                  Yeah. I didn t frame my Elf2K board like I threatened to, but it s sitting here on my desk as a piece of art at the moment. – Bill ... {snip} Yeah. I didn t
                                  Message 17 of 25 , Apr 13, 2013
                                  • 0 Attachment
                                    Yeah.
                                    I didn't frame my Elf2K board like I threatened to, but it's sitting here on my desk
                                    as a piece of art at the moment.

                                    – Bill


                                    On 4/13/2013 4:42 PM, David Keith wrote:
                                     
                                    {snip}

                                  • jdrose_8_bit
                                    ... Sounds wonderful.
                                    Message 18 of 25 , Apr 13, 2013
                                    • 0 Attachment
                                      --- In cosmacelf@yahoogroups.com, "urrossum@..." <mark@...> wrote:
                                      >
                                      > All of these sound like great ideas. In my mind, one could keep verbatim the existing instruction set, but just speed it the heck up
                                      >
                                      > How about a 50 MHz 1802, with every instruction only taking one clock cycle?
                                      >
                                      >

                                      Sounds wonderful.
                                    • David Keith
                                      Lee, The ABC was amazing for it s time. They did not get credit for being the first electronic computer until much later.  I have attached two circuit diagram
                                      Message 19 of 25 , Apr 14, 2013
                                      Lee,

                                      The ABC was amazing for it's time. They did not get credit for being the first electronic computer until much later.  I have attached two circuit diagram which I am not following exactly for my PCB layout.  I'm trying to do it with one VCC source and ground, so I may have to redo the PCB if my prototyping of the circuit fails.  Using a PCB is me just being lazy, I am not worried about an exact replica, someone else has already done that.  They do make 9 pin PCB tube sockets that hold the 12AU7 tubes about 3/8 inch above the board for cooling. However, I am concern with the thickness of the power supply leads, so they are both on both top & bottom layers. 

                                      --David Keith


                                      From: Lee Hart <leeahart@...>
                                      To: beloved_wind@...
                                      Sent: Sunday, April 14, 2013 1:03 AM
                                      Subject: Re: [cosmacelf] Re: CDP1806ACE - ISA Standard committee

                                      David Keith <beloved_wind@...> wrote:
                                      > Atanasoff–Berry Computer...

                                      This is amazing! I had never heard of this computer before. It is quite
                                      a remarkable invention, and way ahead of its time.

                                      The circuit board you showed: Where did you find the circuit diagrams of
                                      it? Are you building a reproduction of the ABC? Are you worried that
                                      PCBs weren't invented at the time, and in fact don't work all that well
                                      with tubes (because the heat of the tubes tends to damage the board and
                                      solder joints).

                                      --
                                      If we knew what we were doing, it wouldn't be called research, would it?
                                          -- Albert Einstein
                                      --
                                      Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm




                                    • ajparent1
                                      ... the power as in plate voltage is low and in the milliamps area. the killer is the heater power as even at 12.6V that .15A per tube. Put ten of them on the
                                      Message 20 of 25 , Apr 14, 2013
                                      • 0 Attachment
                                        --- In cosmacelf@yahoogroups.com, David Keith <beloved_wind@...> wrote:
                                        >
                                        > Lee,
                                        >
                                        >
                                        > The ABC was amazing for it's time. They did not get credit for being the first electronic computer until much later.  I have attached two circuit diagram which I am not following exactly for my PCB layout.  I'm trying to do it with one VCC source and ground, so I may have to redo the PCB if my prototyping of the circuit fails.  Using a PCB is me just being lazy, I am not worried about an exact replica, someone else has already done that.  They do make 9 pin PCB tube sockets that hold the 12AU7 tubes about 3/8 inch above the board for cooling. However, I am concern with the thickness of the power supply leads, so they are both on both top & bottom layers. 
                                        >
                                        >

                                        the power as in plate voltage is low and in the milliamps area.

                                        the killer is the heater power as even at 12.6V that .15A per tube.
                                        Put ten of them on the board and the lines need to be large enough for 1.5A. G10 board will tolerate the heat fine with good airflow.

                                        Cooling, fans are a great idea. you want to remove the heat as
                                        tubes themselves were fairly reliable but the heat could cook the local electronics (caps, resistors, and the like.) Orienting the board so the tubes are horizontal and the board vertical with good airflow upward (fan forced will help) can mtigate the heat things.
                                        build strong (use 2W resistor where 1W will work).

                                        If you have never worked with tubes, two things running them at 12.3V
                                        (tightly regulated) will greatly increase life especially if you bring them up slowly. Power cycling them tends to kill the heater.

                                        The other item is they do not like to be sitting at cutoff for long periods of time (cumulative days) as that poisons the cathode and emission goes down. There were special variants that were more tolerant of that.

                                        Doing a functional tube computer in this day is very possible.
                                        I understand a British group rebuilt and flat out made a copy of
                                        one of the code breaking machines (tubes).

                                        Allison



                                        > --David Keith
                                        >
                                        >
                                        >
                                        > ________________________________
                                        > From: Lee Hart <leeahart@...>
                                        > To: beloved_wind@...
                                        > Sent: Sunday, April 14, 2013 1:03 AM
                                        > Subject: Re: [cosmacelf] Re: CDP1806ACE - ISA Standard committee
                                        >
                                        >
                                        > David Keith <beloved_wind@...> wrote:
                                        > > Atanasoffâ€"Berry Computer...
                                        >
                                        > This is amazing! I had never heard of this computer before. It is quite
                                        > a remarkable invention, and way ahead of its time.
                                        >
                                        > The circuit board you showed: Where did you find the circuit diagrams of
                                        > it? Are you building a reproduction of the ABC? Are you worried that
                                        > PCBs weren't invented at the time, and in fact
                                        > don't work all that well
                                        > with tubes (because the heat of the tubes tends to damage the board and
                                        > solder joints).
                                        >
                                        > --
                                        > If we knew what we were doing, it wouldn't be called research, would it?
                                        >     -- Albert Einstein
                                        > --
                                        > Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm
                                        >
                                      • jdrose_8_bit
                                        ... What are the differences between regular 2N2222 transistors and Field Effect transistors when implementing logic gates?
                                        Message 21 of 25 , Apr 17, 2013
                                        • 0 Attachment
                                          > --- In cosmacelf@yahoogroups.com, David Keith
                                          >
                                          > I proposed the ISA committee so we as a group of 1802
                                          > assembler programmers can have a common future version to use and share.
                                          >
                                          > There are many ways to implement it, and several have already
                                          > been done by this group. I believe even Lee suggested implementing an
                                          > 1802 using fet transistors :-) a month or so back.
                                          >

                                          What are the differences between "regular" 2N2222 transistors and Field Effect transistors when implementing logic gates?
                                        • Lee Hart
                                          ... Quite a bit! Bipolar transistors are current-controlled; a current into the base (at negligible voltage) turns them on. FETs (including MOSFETs) are
                                          Message 22 of 25 , Apr 17, 2013
                                          • 0 Attachment
                                            On 4/17/2013 8:14 PM, jdrose_8_bit wrote:
                                            >
                                            >> --- In cosmacelf@yahoogroups.com, David Keith
                                            >>
                                            >> I proposed the ISA committee so we as a group of 1802
                                            >> assembler programmers can have a common future version to use and share.
                                            >>
                                            >> There are many ways to implement it, and several have already
                                            >> been done by this group. I believe even Lee suggested implementing an
                                            >> 1802 using fet transistors :-) a month or so back.
                                            >>
                                            >
                                            > What are the differences between "regular" 2N2222 transistors and Field Effect transistors when implementing logic gates?

                                            Quite a bit! Bipolar transistors are current-controlled; a current into
                                            the base (at negligible voltage) turns them on. FETs (including MOSFETs)
                                            are voltage-controlled; a voltage into the gate (at negligible current)
                                            turns them on.

                                            You can built the same sort of circuits with either type, but the
                                            details of the circuit are quite different. Compare the internal
                                            schematic for a bipolar gate (7400 etc.) which is built with bipolar
                                            transistors, to that of a CMOS gate (4011 etc. built with MOSFETs. (The
                                            internal circuits are on the older data sheets.)

                                            If you wanted to build an 1802 from individual transistors, n-channel
                                            and p-channel MOSFETs would work with exactly the same circuits used in
                                            the 1802. The MOSFETs would cost you around 5-10 cents each.

                                            If you wanted to build an 1802 with bipolar transistors, you would have
                                            to replace each type of gate (inverter, NAND, NOR, flip-flop, etc.) with
                                            its equivalent bipolar circuit. It would wind up requiring more
                                            transistors and be faster; but use more power. Bipolar transistors are
                                            3-5 cents each, but you'd need more of them so the price probably comes
                                            out about the same.

                                            Either way, it would be a major project; like building a model Eiffel
                                            tower out of toothpicks. :-)

                                            --
                                            If you would not be forgotten
                                            When your body's dead and rotten
                                            Then write of great deeds worth the reading
                                            Or do these great deeds, worth repeating.
                                            -- Ben Franklin, from Poor Richard's Almanac
                                            --
                                            Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm
                                          • David Keith
                                            A couple months ago I was recovering from double knee replacements (6 weeks apart) I had a lot of time on my hands that I could not work, I have been doing a
                                            Message 23 of 25 , Apr 17, 2013
                                            A couple months ago I was recovering from double knee replacements (6 weeks apart) I had a lot of time on my hands that I could not work,  I have been doing a lot experimenting with Diode Transistor Logic for a while now, so I decided to design a PCB and have some madeDTL is much easier and cost effective than a pure transistor circuit.  I decided to redesign an IBM SMS card using today's discreet components LOL.(I was on some very good drugs)  The original only SMS card I have contained 6 NAND circuits, while my new design uses 48 AOI (and-or-invert) gates (I prefer AOIs because I can design much tighter logic, with much less transistors). I have attached several pictures, of comparing the SMS basic form to my AOI board.  With another with a card almost populated, The last is the layout showing a 1 bit-slice CPU with 8 registers, full functional ALU (74181 equivalent), and data register to memory interface.  Overall it was just something to do that was not work related.  I plan to populate a board a month to get an 8-bit CPU/ALU core.

                                            --David Keith


                                            From: Lee Hart <leeahart@...>

                                            On 4/17/2013 8:14 PM, jdrose_8_bit wrote:
                                            >
                                            >> --- In cosmacelf@yahoogroups.com, David Keith
                                            >>
                                            >>  I proposed the ISA committee so we as a group of 1802
                                            >> assembler programmers can have a common future version to use and share.
                                            >>
                                            >>  There are many ways to implement it, and several have already
                                            >> been done by this group. I believe even Lee suggested implementing an
                                            >> 1802 using fet transistors :-)  a month or so back.

                                            >> What are the differences between "regular" 2N2222 transistors and Field Effect transistors when implementing logic gates?

                                            >You can built the same sort of circuits with either type, but the
                                            >details of the circuit are quite different. Compare the internal
                                            >schematic for a bipolar gate (7400 etc.) which is built with bipolar
                                            >transistors, to that of a CMOS gate (4011 etc. built with MOSFETs. (The
                                            >internal circuits are on the older data sheets.)

                                            >If you wanted to build an 1802 from individual transistors, n-channel
                                            >and p-channel MOSFETs would work with exactly the same circuits used in
                                            >the 1802. The MOSFETs would cost you around 5-10 cents each.

                                            >If you wanted to build an 1802 with bipolar transistors, you would have
                                            >to replace each type of gate (inverter, NAND, NOR, flip-flop, etc.) with
                                            >its equivalent bipolar circuit. It would wind up requiring more
                                            >transistors and be faster; but use more power. Bipolar transistors are
                                            >3-5 cents each, but you'd need more of them so the price probably comes
                                            >out about the same.


                                            --
                                          • bill rowe
                                            Keith: what does the card do? To: cosmacelf@yahoogroups.com From: beloved_wind@yahoo.com Date: Wed, 17 Apr 2013 21:19:46 -0700 Subject: Re: [cosmacelf] Re:
                                            Message 24 of 25 , Apr 18, 2013
                                            • 0 Attachment
                                              Keith: what does the card do?


                                              To: cosmacelf@yahoogroups.com
                                              From: beloved_wind@...
                                              Date: Wed, 17 Apr 2013 21:19:46 -0700
                                              Subject: Re: [cosmacelf] Re: 2N2222 ++ 1N914 DTL (off topic) [4 Attachments]

                                               
                                              [Attachment(s) from David Keith included below]
                                              A couple months ago I was recovering from double knee replacements (6 weeks apart) I had a lot of time on my hands that I could not work,  I have been doing a lot experimenting with Diode Transistor Logic for a while now, so I decided to design a PCB and have some madeDTL is much easier and cost effective than a pure transistor circuit.  I decided to redesign an IBM SMS card using today's discreet components LOL.(I was on some very good drugs)  The original only SMS card I have contained 6 NAND circuits, while my new design uses 48 AOI (and-or-invert) gates (I prefer AOIs because I can design much tighter logic, with much less transistors). I have attached several pictures, of comparing the SMS basic form to my AOI board.  With another with a card almost populated, The last is the layout showing a 1 bit-slice CPU with 8 registers, full functional ALU (74181 equivalent), and data register to memory interface.  Overall it was just something to do that was not work related.  I plan to populate a board a month to get an 8-bit CPU/ALU core.

                                              --David Keith


                                              From: Lee Hart <leeahart@...>

                                              On 4/17/2013 8:14 PM, jdrose_8_bit wrote:
                                              >
                                              >> --- In cosmacelf@yahoogroups.com, David Keith
                                              >>
                                              >>  I proposed the ISA committee so we as a group of 1802
                                              >> assembler programmers can have a common future version to use and share.
                                              >>
                                              >>  There are many ways to implement it, and several have already
                                              >> been done by this group. I believe even Lee suggested implementing an
                                              >> 1802 using fet transistors :-)  a month or so back.

                                              >> What are the differences between "regular" 2N2222 transistors and Field Effect transistors when implementing logic gates?

                                              >You can built the same sort of circuits with either type, but the
                                              >details of the circuit are quite different. Compare the internal
                                              >schematic for a bipolar gate (7400 etc.) which is built with bipolar
                                              >transistors, to that of a CMOS gate (4011 etc. built with MOSFETs. (The
                                              >internal circuits are on the older data sheets.)

                                              >If you wanted to build an 1802 from individual transistors, n-channel
                                              >and p-channel MOSFETs would work with exactly the same circuits used in
                                              >the 1802. The MOSFETs would cost you around 5-10 cents each.

                                              >If you wanted to build an 1802 with bipolar transistors, you would have
                                              >to replace each type of gate (inverter, NAND, NOR, flip-flop, etc.) with
                                              >its equivalent bipolar circuit. It would wind up requiring more
                                              >transistors and be faster; but use more power. Bipolar transistors are
                                              >3-5 cents each, but you'd need more of them so the price probably comes
                                              >out about the same.


                                              --

                                            • jdrose_8_bit
                                              Thank you. Great explanation. I knew how bipolar transistors worked but FETs were a mystery to me.
                                              Message 25 of 25 , Apr 18, 2013
                                              • 0 Attachment
                                                Thank you. Great explanation. I knew how bipolar transistors worked but FETs were a mystery to me.

                                                --- In cosmacelf@yahoogroups.com, Lee Hart <leeahart@...> wrote:

                                                > >
                                                > > What are the differences between "regular" 2N2222 transistors and Field Effect transistors when implementing logic gates?
                                                >
                                                > Quite a bit! Bipolar transistors are current-controlled; a current into
                                                > the base (at negligible voltage) turns them on. FETs (including MOSFETs)
                                                > are voltage-controlled; a voltage into the gate (at negligible current)
                                                > turns them on.
                                                >
                                                > You can built the same sort of circuits with either type, but the
                                                > details of the circuit are quite different. Compare the internal
                                                > schematic for a bipolar gate (7400 etc.) which is built with bipolar
                                                > transistors, to that of a CMOS gate (4011 etc. built with MOSFETs. (The
                                                > internal circuits are on the older data sheets.)
                                                >
                                                > If you wanted to build an 1802 from individual transistors, n-channel
                                                > and p-channel MOSFETs would work with exactly the same circuits used in
                                                > the 1802. The MOSFETs would cost you around 5-10 cents each.
                                                >
                                                > If you wanted to build an 1802 with bipolar transistors, you would have
                                                > to replace each type of gate (inverter, NAND, NOR, flip-flop, etc.) with
                                                > its equivalent bipolar circuit. It would wind up requiring more
                                                > transistors and be faster; but use more power. Bipolar transistors are
                                                > 3-5 cents each, but you'd need more of them so the price probably comes
                                                > out about the same.
                                                >
                                                > Either way, it would be a major project; like building a model Eiffel
                                                > tower out of toothpicks. :-)
                                                >
                                                >
                                                > Lee A. Hart, http://www.sunrise-ev.com/LeesEVs.htm
                                                >
                                              Your message has been successfully submitted and would be delivered to recipients shortly.