• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Experiment Platform for the MC68000

Status
Not open for further replies.

Quiki

New Member
For fun, I am resurrecting an old MC68000P on a breadboard. I've never designed a computer system from scratch before, so to start I hooked the 68000's address and control lines to some I/O from a PC-104 card on a 200MHz Linux computer. The program initializes the I/O pins, and then runs through a state machine to properly communicate with the chip. First the processor's /HLT and /RST lines are asserted for 1/2 a second to reset, then released. Next the card waits for the 68000 to assert its /AS. Upon seeing that, the card sends a /DTACK. And then it releases /DTACK once /AS is negated. The data bus is hard wired with pull-up/down resistors for a 'branch-back-to-itself' instruction. Now the problem...

I can start the processor and, after a few clock cycles, the processor does 2 16-bit read cycle to fetch the SP. Next it should read 2 more for the PC. But instead it hangs without ever asserting /AS again to start the next bus cycle. I made a number of minor timing changes, with no improvment. Also I notice that the behavior is not always consistent, with a bus cycle starting without the Data Strobes being asserted. (Maybe timing irregularities from the Linux board?)

I read in the documentation that there is a minimum clock frequency for the chip (that I am certainly violating) and that the clock signal must be square (which I am crudely approximately). Do clock irregularities really matter in a static processor?? If I run off a 1MHz crystal clock, the processor makes all 4 reads.

Any other ideas? Thanks a lot!
 

RCinFLA

Well-Known Member
Clock quality is very important. Multiple parallel things are happening and a corrupt clock can cause some things to sequence and other things not, resulting in an illegal state.
 

Quiki

New Member
Thanks for the reply! After playing a bit more, that's consistent with what I saw. I hooked the 68k up to an FPGA instead of the PC-104 to eliminate the very crude signal timing performance. After doing that, 68k worked fine, even down to a couple Hz. I guess the (relatively) expensive quartz crystals earn their keep.

BTW, after verifying the logic in the FPGA (avoiding tedious wiring of discrete chips), I finally broke down an wired up a version all in discrete logic, with 27256 EEPROM, 6264 SRAM, etc. This FPGA is still in there, but now as a 'chip selected' device for sound, VGA display, PC keyboard input, etc.

68k forever, right??
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top