I've made some progress on the Mark 2 project and started the Mark 3!! First the Mark 2, I've built the system that I detailed in the articles on the Mark 2. There are a few odd bugs that I spotted as I tidied the circuit diagrams to add to the articles, these will need minor re-wiring of a bit of the glue logic. I kept the system on two bits of strip-board with the CPU and the PIC on the top layer with a bit of glue logic, the rest of the system (I/O decode, memory and peripherals are all on a lower layer, interfaced with 4 ten-way pin headers and receptacles. (Pics will follow soon). The pin-outs of the connectors can be seen in the circuit diagrams.
I've done electrical testing on the system so far. All of the power pins have the right voltages on them, including the +/-10V outputs on the MAX232 chip. The clock output on the 74HC4060 chip that provides the clock for the UART is stable and at the correct frequency. The PIC also registers with the correct part number when connected to a PicKit2 programmer.
The next step I plan to do is to write software for the PIC. My plan is to write a debug kernel for the PIC that can drive the system bus without the Z80 fitted to probe the peripherals and memory to ensure there are no bus shorts or mistakes in peripheral configuration. This code will be the basis for the debug mode of the PIC once the Z80 is fitted, with the DMA mode being enabled before the PIC drives the bus.
Beginnings of the Mark 3 I've already started on the Mark 3 version of the system. It's going to be based on the Mark 2 functionally, but the design of the hardware will be very different. I've come across the DIN41612 96-way connector and realised the possibility of a back-plane based system. This is not a new idea of course, and I'd seen DIN41612 backplanes before but never connected it with a Z80 computer. A notable example is the n8vem project.
The idea being that I can upgrade parts of the system without replacing the whole PCB. This new design will replace the massive PCB that I had been working on for the Mark 2 which got shelved because of the massive cost of getting that size PCB made and the fact that it would have been impossible to upgrade. I've managed to get hold of a second hand sub-rack from a prototype at my dad's work. I've mounted a PC power supply and standard set of ATX front panel indicators and controls in aluminium panels at one end. I've just finished routing a backplane for it.
The 96 connections in the back-plane are mainly just connected together, but there is a full row (of 3 connections) on the end of each connector assigned to +5V, another row to 0V and one to +3.3V. The backplane has the ATX power-supply connector on it with a 3 pin header to wire off the control signals to and from it for software controlled power supply, as well as an over-ride jumper which when fitted forces the PSU on so it is controlled by the mains switch on the back. +/-12V and the 5V standby power signals are also routed onto the bus but only for low current applications. There are no bus terminations on the backplane so each line can be fitted with pull-up or down resistors depending on the application. This gives maximum flexibility because I am considering the use of alternate processors as a future possibility.