After some time, on hot summer days of 2014 I returned to DolphinDOS 2 as there were still things that I wanted to improve in the design. The most important one was the method of enabling and disabling the expansion. Originally Mr. Jilg (I think) decided to pass the A15 address line through a mechanical switch and used two, relatively strong "push-down" resistors to keep the line in LO state unless actively driven otherwise. This brings quite a few problems with it. First the HF address line is run through far less than optimal environment of potentially long cable to the switch, then it has to go through the said switch, while cheap (and even less cheap) switches are notorious for their HF unreliability. Therefore the cable has to be kept short and carefully routed, while the push-down resistors have to be of low resistance in order to avoid numerous problems related to high frequency operation. This – OTOH – should be avoided whenever possible, especially with NMOS drivers, which are not the most powerful when it comes to driving HI. Summing up – I didn't like this approach from the moment I understood what the resistors were there for. What I did were some tests and I found that originally used 470Ω resistors can be replaced with weaker 1kΩ ones, without causing operational problems. At least on the boards I designed. That's what I did but I still didn't want to leave it that way forever. Therefore the remaining, originally unused gates were put into a good use and connected in such way that enabling and disabling of the expansion can be done by simple shorting of a single static line with one weak pull-up attached to it. When discussing this with Gerrit, we had some doubts regarding the apparent difference in propagation paths / times so I thought that the best way to check this is to simply build a prototype and verify. It worked :-)



With the most important schematic change out of the way I moved on to laying the new board out, introducing the changes I had in mind for some time. I re-routed the whole PCB from scratch, increased both signal and power tracks widths and grouped all (or so I initially thought) jumpers into one jumper block on the lower right corner of the routing area. After numerous rounds of routing, optimising and polishing the layout, when I was basically ready, I got several, virtually simultaneous requests from people who wanted to run another DOS (either SpeedDOS or Jiffy DOS) with my boards. I promised them I'll try to introduce this possibility to the new design. I quickly added the requested option to the schematic but I didn't feel like re-routing the board again in order to place the additional jumper together with all the other, I just successfully grouped. So, please bear with me, will you?




  1. ID pins (JP1 3-4, 5-6) have to be both CLOSED or you will get your device working as drive 11 rather than - what you probably expect - 8
  2. CON4-0 and CON4-1 are 20-pin pinheads, which I use for connecting to the main controller board’s 6502 soldering pads. CON3-0 and CON3-1 do the same for the 6522 VIA. I use "goldpin" type of connectors for them. If you plan to use the same, please note that the pins need to be kept straight when mounting. I myself use socket constructions to keep the strips straight and aligned when soldering. For this revision I dropped the idea of reduced drills diameters I used in the previous revision. It worked well for me and most of the others but not for everyone.
  3. Care has to be taken when mounting the hardware inside the drive in order to make it sit low and fit properly inside the drive’s case, yet remain elevated above the level of other elements on the controller’s main board!
  4. Pin "1" of the parallel connector is marked with a pointing triangle on the PCB. This means that if your cable is correctly built and has one wire marked as wire #1, the notches will force it to become wire #10 rather than #1 if you use a regular socket. Therefore – for now – it is probably best to use a right-angled pinhead strip rather than a full socket.

For Sale: