Sunday 30 October 2022

Let's start doing something about those missing expansion ports

The MEGA65 case has a bunch of cut-outs for ports that we didn't include on the motherboard, to prevent the cost from being increased more than necessary. But by putting those cut-outs into the case, and providing the supporting screw bushes in the bottom of the case, we have the ability to easily design a PCB that can provide the missing ports.

No longer will this area of your MEGA65 need to look so empty.
 

I've been learning Kicad, the free PCB design software lately, and have made an initial design for an expansion board that would provide all the missing ports: user port, component video, C1565 external disk drive, and C64 tape port.  This means that we have all of the ports that a C64 or a C65 had: The C64 obviously didn't have a C1565 port, while the C65 lacks a tape port.

The main challenge for designing such a board is that we don't have a huge number of spare IO pins on the MEGA65.  We basically have the two PMOD ports in the trap-door slot, which give a total of 16 IO lines.  So we have to work with those.  They are also not super ideally positioned, but we will have to live with that.  A pair of ribbon cables with headers on the end, or perhaps a little adapter PCB that provides the necessary connectors and connections will be possible solutions. For prototyping, I'll just use a gaggle of jumper wires.

Getting component video and audio working will be the main challenges, and the solution I have devised is quite experimental for this: I am using 4-way resistor ladder DACs for chroma, luma and audio.  These will be clocked at something like 40.5MHz, or perhaps even 405MHz using SERDES from the FPGA, so that we can get much more than 4-bit resolution.  It's possible the initial design will have incorrect resistor values to get the correct voltage range, and that it might all require a bit of tweakage. But that's what prototyping is for!

Having chroma, luma and mono-audio with 4 bits each consumes 12 of our 16 pins, which is annoying. So we need a very low pin-count solution.  I could have used I2C, but have instead opted for using discrete 74LS and 74LVC logic chips, mainly serial shift registers.  There are a few reasons for this: 

1. The 74LS/LVC chips are readily available, even while Chipaggedon rages on (for now, at least). 

2. Having a bunch of DIP chips inside your MEGA65 case will have a nice retro feel to it.

3. I want this board to be able to be hand-soldered by just about anyone, so this means no surface mount parts.

This means that our interface consists of a clock from the MEGA65 side to the shift registers, and a data out and a data in line. That makes 4 in total, which is exactly what we had left over after our three DACs.

Some further tricks I have done to reduce the pin count, is to make the C1565 port get a number of its data lines from the 34-pin floppy connector. This is why the expansion board has 2x 34-pin floppy connectors: One goes to the MEGA65, and the other to your internal floppy drive. The two connectors are just pass-through between them. I similarly provide two floppy power headers: One to get the 12V we need for the tape port power supply (the tape drive requires 6V DC) as well as the 5V we need for a bunch of the logic chips, and the other so that you don't end up with one less power connector, if you had some other use for it in your MEGA65.

Here is what the design looks like:


The two funny cut-outs are to make it easier to your your floppy cable from the edge of this board under the internal drive, and then to allow  the cable to re-emerge to enter the back of the internal drive.

In volumes of 100 units, the bare board should cost less than US$15, and the parts probably another US$50 or so.

Otherwise, there probably aren't too many surprises with this board. Prototyping will let me check that the design works, add support for it to the MEGA65 core, and also make sure that I have all the measurements and port positions correct.

The source for it will be up soon at https://github.com/mega65/mega65-r3-expansion-board. But it is marked private for a couple of weeks, because some of the commits would otherwise spoil another surprise I have in the works.  But be assured, they will be available soon.  This PCB will be complete open hardware, meaning that the gerbers and everything else you would need to make one entirely on your own will be available there.

What I would love in the meantime, is anyone who would like to help me with the prototyping and testing.  Between this, the keyboard, and the other surprise PCB, I've put a whole pile of time as well as PCB prototyping production costs in lately, and it would be great to have some assistance with getting this all working.  I'd love for folks to be able to buy the parts for the expansion board before Christmas this year, to have something fun to work on on their MEGA65s over the break.

10 comments:

  1. Any rough idea on cost of the parts?

    ReplyDelete
    Replies
    1. I'm guessing that the PCB and parts should be around US$50 - US$80 in smallish quantities.

      Delete
  2. what a nice suprise ! :)

    ReplyDelete
  3. When I get my m65, I will be first in line for this.

    ReplyDelete
  4. I still wish there was a USB port ;) - but nice to see the progress on that, to have all C64/C65 ports available!

    ReplyDelete
  5. Surprise? IEEE-488 pins maybe? :D

    ReplyDelete
    Replies
    1. That would be interesting. Unfortunately there is no spare spot for it. Could be possible to put one in in place of the tape port, though, for example.

      Delete
    2. Or maybe the planned „desktop” form factor case mentionned at Discord could have a separate opening for the port?

      Delete
  6. Please let me know if you need any testers.

    ReplyDelete
    Replies
    1. Best is to hop on the discord server and announce your willingness in #peripherals, where activity on this will occur. It might be a few weeks before I have time again to work on this, as I have a few things in my queue.

      Delete