Feel and see how computation works - with Konrad Zuse style.
There are NO diodes involved - a relay computer with semiconductor logic is not a realay computer, so all logic will be based solely on relays. One motor will be driving the program tape.
However, the relays have LEDs to show if they are on - but they don´t have any logic function, they can be replaced by lightbulbs or left out completely without disturbing the funcionality of the computer.
This machine is intended to be some kind of educational art with a fun factor. The hat rail box can be put onto the wall to make it a playable decoration.
The memory addressing and read/write operations work after having finished the wiring. You can select a 5-bit address by clocking +24V or GND to the 5 address relays to set or reset them. The memory multiplexer selects the one of the 32 cells and puts its state to the ALU-Multiplexer which can store the bit into the ALU a,b,c cells. A +24/GND clock on the write wire will write 1/0 to the cell.
Row 3 and 4 are the memory demultiplexer and the memory itself. To save wires and relays, I use resistors to latch the relays. The advantage of it is that clocking +24V will latch a bit ON and clocking the GROUND will latch it OFF. A relay can still hold the on-position with less than half of the nominal coil voltage, so this hack is very useful.
I´ll modify my actual design to this kind of latching also for the address- and ALU-multiplexers. It will make the computer more powerful with a minimal amount of effort.
Hi Paul, I saw a replica of one of Zuse's computers when I was in Berlin in 2009. I've never seen a relay computer running in real life though - I imagine they must be quite loud :) What are you planning in terms of execution speed, memory etc?
I will keep the build simple, so there will be only 32 bits of memory in this build, but scalable. It´s a 1-bit architecture, so an 8bit+8bit addition will take 8x longer than an 8bit-ALU. But it saves tons of relays :) The relays should be able to switch stable with up to 100Hz, the maximum switch depth is two, so 50 Hz should be doable.
Years ago I made the schematics - should have been a compact PCB relais computer. Never finished it due to other projects. Now I decided to make it big on hat rail. I ordered most parts already and will start to build this week.
Hi Paul, I saw a replica of one of Zuse's computers when I was in Berlin in 2009. I've never seen a relay computer running in real life though - I imagine they must be quite loud :) What are you planning in terms of execution speed, memory etc?