Hey,
I have been building a Mark-8 Minicomputer since early this year. It has been taking me a long time to finish as I have a number of retro computing projects on the go and don't work on the same project all the time.
Another issue that has delayed me is having to deal with LS logic ICs that don't seem to be happy to work in the Mark-8.
For example here was my first pass at the CPU clock:
Yes, you read that right 4MHz and yes a very messy sine wave. Also you will see that both phases of the clock are in sync.
So I eventually traced part of this fault to the 74LS00s for IC 1 and 4 on the CPU. Once I removed these and replaced them with some from a different manufacturer things improved to this:
That's better. We are at 500KHz now and have something that looks like a square wave. However both phases of the clock are in sync.
So a bit more troubleshooting led me to check the 74LS76 flip flop IC3 on the CPU board.
If we look at the two outputs on pins 11 and 10 we get this:
Clearly these outputs should not be overlapping and ultimately this is causing the clock circuit to generate the two phases both in sync.
Here is the input of IC3 on pin 6 compared with each of the two outputs, input is the blue line, output is yellow:
Interesting to note that one is being triggered on the rising edge of the input while the other is being triggered on the falling edge!
Regardless to say that replacing these two flip flops at IC positions 2 and 3 with ones from a different manufacturer gets us to this:
That's much better and with a proper clock signal I am pleased to report that my Mark-8 is now happily executing code!
What's curious about this is that the "faulty" LS chips that I pulled from the Mark-8 work when I build a simple test circuit on a breadboard. Main difference is that I am testing them at slow speed rather than driving them from a 4MHz crystal.
So presumably not all LS chips are made equal? Or I got sold a bad batch?
Last up, a shot of my Mark-8:
Thoughts anyone?
Cheers,
Gavin.
Very nice write-up! Thank you for sharing your findings and insights. I think I used LS chips for my clock and it appears to be working fine. It may just be some bad chips.
Finding issues with the build is not always pleasant but it encourages the builder to dig in and learn about the inter-workings of the computer, which is a plus.
I have to say, your computer is looking amazing so far!