[net.micro.6809] M6829 Memory Management Chip

davecl@mako (12/10/84)

For two or three years the Motorola data books have had data sheets
on the 6829, a memory management chip for the 6809.  However, I have
never seen the 6829 at any of the "commercial" distributors
(i.e., back of Byte ads, etc.).

Was the 6829 ever produced?

Dave Clemans
tektronix!tekecs!davecl

hammond@petrus.UUCP (12/11/84)

> ...
> Was the 6829 ever produced?
> ...

I received samples from a Motorola Rep of the 6829 well over a year ago.
I don't remember whether they were stamped 6829 or were preliminary.
I never actually used it, since the prices on the 68000 were dropping so
fast that we went with it and a homebrew MMU instead.  If the 6829 had
been available when the 6809 came out, the combination would have been
the neatest thing since sliced bread.  As it is, it is too little, too late.
Rich Hammond Bell Communications Research (decvax!bellcore!hammond)

bmw@aesat.UUCP (Bruce Walker) (12/12/84)

/*NOTEATEN*/

The MC6829 does exist, I have one and have had it for about three years.  The
chip is not particularly popular, I suspect, because SWTPC and Gimix invented
their own MMU when the 6809 systems were first brought out.  You should check
with your local Hamilton-Avnet dealer if you'd really like to get one.  Be
prepared to spend $30 on it.  It is a really neat chip if you are building a
real-time kernel for process control or something.  Not quite up to the
Z8001 & Z8010 combo but simpler to understand and build for hobbyists
certainly.

Bruce Walker     {allegra,ihnp4,linus,decvax}!utzoo!aesat!bmw

david@bragvax.UUCP (David DiGiacomo) (12/18/84)

I think the 6829's unpopularity is well justified; by today's standards
it is a very small, very slow, very expensive RAM.  Motorola would have
been better off building a tiny chip with just the "fuse" logic (and
maybe a single step NMI generator).  Of course, they could have just put
another stack pointer in the 6809.

Anyway, any hobbyist who is tempted to use the 6829 is probably better
off duplicating the "fuse" logic (maybe in a PAL) and using modern fast
static RAMs for the MMU (maybe 2148s-- not really modern).  It might be
better to build an adder type MMU a la PDP-11 if the additional delay is
tolerable, unless there's someone crazy enough to do paging on a 6809.

-- 
David DiGiacomo, BRAG Systems Inc., San Mateo CA  (415) 342-3963
(...decvax!ucbvax!hplabs!bragvax!david)