Simple project using 68K microprocessor

[BobG]

I see people make a 3 order of magnitude mistake all the time when they
say billion and mean million or say nf and mean pf. I think this is a
simple 1 order of magnitude misspeak and the instructor told him to use
a 6800.

 

[Fred Bartoli]

BobG a écrit :

I see people make a 3 order of magnitude mistake all the time when they
say billion and mean million or say nf and mean pf. I think this is a
simple 1 order of magnitude misspeak and the instructor told him to use
a 6800.

No, 68K and 3 orders of magnitude make for 68. Maybe he forgot one
because 69 is just what's needed to taste flip flaps.

 

[Michael Black]

I see people make a 3 order of magnitude mistake all the time when they
say billion and mean million or say nf and mean pf. I think this is a
simple 1 order of magnitude misspeak and the instructor told him to use
a 6800.

But even then, why in the world would nayone need a CPU to test a few
simple gates?

And if the point is to put together a small computer, then surely one
would come up with a more recent 8bit CPU than the 6800 (or for that
matter a more recent 16bit CPU than the 68000).

I'd use the old CPUs for some project, but that's because I'm loaded
with them, taken out of old junk. Unless the school is supplying parts,
and has lots of old parts, using the 68000 will be an impediment, apart
from overkill.

Michael

 

[Spehro Pefhany]

But even then, why in the world would nayone need a CPU to test a few
simple gates?

Ask Teradyne or Advantest.

And if the point is to put together a small computer, then surely one
would come up with a more recent 8bit CPU than the 6800 (or for that
matter a more recent 16bit CPU than the 68000).

True, both are approaching museum-material status.

I'd use the old CPUs for some project, but that's because I'm loaded
with them, taken out of old junk. Unless the school is supplying parts,
and has lots of old parts, using the 68000 will be an impediment, apart
from overkill.

Michael

It's an equally valuable exercise for pedagological purposes whether
the student uses a 68000 or an ARM or AVR. There is perhaps some small
advantage in using an older, more orthogonal, architecture both for
education and because nobody is likely going to be able to google the
source code off the net.



Best regards,
Spehro Pefhany

 

[Tim Auton]

But even then, why in the world would nayone need a CPU to test a few
simple gates?

They need a CPU because solutions not using a CPU can be expected to
score a big fat zero when handed in as project work for a course in
microprocessors.

It's an excercise, nothing more. Step two after flashing an LED. Perhaps
next term/semester they make something really difficult and exciting,
but you gotta learn the tools, techniques and tricks somehow.

To the OP:

Apologies if this is overly basic, I'm assuming you just don't know
where to start.

I assume the 68230 and the logic chips you will be testing are both
running off the same voltage and have compatible IO levels, so can be
connected directly. Connect the logic chip to power and to your 68230.
Make a note of which pins connect to which. That is all the circuit you
need (all the additional circuitry to do this anyway - you still need
the 68k, clock source, power, programming interface...). You can now set
the inputs of the logic and read its outputs all from software, by
controlling the 68230's IO pins. I expect they chose testing logic chips
precisely because they are so easy to interface with microcontrollers.

You need to provide known input (input to the logic, output from the
68230) and compare the output with the expected output. Both inputs of
an AND high should make the output high, etc. Testing a flip-flop is
slightly more involved, as you have state to consider too. Still, the
principle is the same.

Of course, you real code will do that for all the gates on a logic chip,
it'll handle a few different kinds of logic chip, there will be a bunch
of code to handle your 7-segment displays...


Tim

 

[David Brown]

Ask Teradyne or Advantest.


True, both are approaching museum-material status.

It's an equally valuable exercise for pedagological purposes whether
the student uses a 68000 or an ARM or AVR. There is perhaps some small
advantage in using an older, more orthogonal, architecture both for
education and because nobody is likely going to be able to google the
source code off the net.

Asking the students to use a 68000 and a 68230 is not any more
pedagogical than asking them to program it in Cobol. Use the m68k
architecture by all means - it is a solid architecture that is very much
in use today - but use a realistic device such as a ColdFire. Very few
embedded engineers design systems using pure processors and stand-alone
timer peripheral chips, so there is little point in training people to
do so. Start with something more practical, and then (if necessary)
move to a more advanced design later.

 

[Paul Burke]

Asking the students to use a 68000 and a 68230 is not any more
pedagogical than asking them to program it in Cobol.... Very few
embedded engineers design systems using pure processors and stand-alone
timer peripheral chips, so there is little point in training people to
do so. .

Mebbe so, but I know of a local firm that makes a nice steady income out
of selling 68000 trainer boards to the educational market.

Paul Burke

 

[John Woodgate]

Mebbe so, but I know of a local firm that makes a nice steady income
out of selling 68000 trainer boards to the educational market.

That says something not very good about the educational market.

 

[David Brown]

Mebbe so, but I know of a local firm that makes a nice steady income out
of selling 68000 trainer boards to the educational market.

Paul Burke

That's a little different - programming the 68000 is perfectly suitable
for educational purposes, since the architecture is very nice, there are
plenty of tools, and the architecture is still widely used in new
systems. It's asking students to design hardware boards based on the
68000 that I think is a waste.

 

[Frank Bemelman]

That's a little different - programming the 68000 is perfectly suitable
for educational purposes, since the architecture is very nice, there are
plenty of tools, and the architecture is still widely used in new systems.
It's asking students to design hardware boards based on the 68000 that I
think is a waste.

Why is that a waste? Let those students worry a bit about addressing and
bus conflicts, perhaps program a PLD for decoding too, instead of just
connecting a led to an I/O pin on a microcontroller.

 

[David Brown]

Why is that a waste? Let those students worry a bit about addressing and
bus conflicts, perhaps program a PLD for decoding too, instead of just
connecting a led to an I/O pin on a microcontroller.

It all boils down to what you want to teach them, and whether the
project matches their current level. Sure, working with a proper bus is
instructive - but only if that's what the course is covering at the
time. It sounds like in this case, the project is way over the heads of
the students, so either they will fail to make any headway, or they will
regurgitate someone else's design - either way, it's a waste of
everyone's time. It is also such an artificial exercise (a powerful
processor for testing a logic gate) that it is a poor example for students.

Having said that, if they want to make students use a processor (as
distinct from a microcontroller), and learn the issues of bus decoding
and everything else, then the 68000 is a good choice - the bus is well
documented, there are some standard peripherals available, and (IIRC) it
is available in DIP package.