All technical project is developed in cycles 'design, testing, design refinement' that are repeated in steps to achieve the objectives. I say "staggered" because a project Technical typically consists of several components dependent on each other, so that should be put in place before designing a few others who depend on them.
Heritage / 1 is a particularly complex project in that (1) comprises a large number of components and (2) the interdependencies between them is quite intricate. That is why it has been difficult to arrive at the desired "phased development phase", finding it stuck in a phase where there are many drawings, but almost nothing has been built, much less proven.
The design, however, can be considered definitive but only in general terms. This includes (among others) the following:
- Architecture of the machine in general and in particular CPU, including details such as provision of records and buses as well as a well-defined philosophy of operation that includes interrupt and arbitration mechanisms.
- Backplane CPU pinout, which includes the detailed design of the internal buses and the allocation of cards to connect directly to the Backplane.
- Detailed design of the interface between the backplane and CPU and other components of the Console, the External Bus and power supply.
- Detailed design of the external buses.
- Complete design of the Console, is composed of four credit cards as well buttons, switches and lamps (LEDs) command.
- Complete design of H11R-GEN card containing two general registers (BC or DE).
- Design part of other circuits such as the Master Controller (H11C-MC card).
- Outline of the other circuits such as ALU, A, PC, SP, decodifucadores instructions (IDS cards) and others.
- Outline of input-output techniques and system software.
It would have been to build these circuits one by one, test (and refine) separately, then connect them together and test their joint operation. The problem is that none of these circuits can work independently; in particular, none of them can work without the existence of Backplane and Console. That is why I put the priority on producing both the backplane and the four cards that make up the Console. At the time of writing I find myself squarely in that task.
The problem of this "no step" is the inefficiency. Backplane wiring the take (estimate) about a month, building the cards will take a week each, so that building this "infrastructure" will take at least two months (probably much more since they do not have much free time) and only then I will be ready to start testing ... All set!
My forecast for the first tests can not be other than "total disaster", so I will take at least another two months to get all that comes to work properly. Upon expiration
this difficult period, ie once set-up the infrastructure of CPU, I can then go to the desired stage of "phased development."
You see this is hardly take place before August of this year. A Heritage / 1 will not be fully functional as available before January 2011.
0 comments:
Post a Comment