[EE]:Looking for Fantasy Microcontroller
Dipperstein, Michael
microcontroller that will get the job done. In the fantasy world this
microcontroller would:
- Have infinite I/O
- Consume no current
- Be in-circuit reprogrammable
- Be given away for free (better yet, I'd get paid to use them)
- Have infinite memory, (RAM, ROM, and EEPROM)
- Have performance measured in the giga FLOPS
If any of you have such a microcontroller let me know.
All kidding aside, the project requires in-circuit reprogrammablity, and long
battery life. At present I have identified a requirement for 39 I/O general
purpose digital pins, and one low resolution ADC. Most of the I/O is not well
suited for I/O expanders. I could probably get away with 8K of instruction
space, 256 bytes of data space, and 2 bytes of non-volatile data space for
storing configuration information.
A PIC16LF877 with two more 8-bit I/O ports would be great. I've tried to come
up with a reasonable way to use a the PIC16LF877 with I/O expanders. It can be
done, but processing time would increase immensely. Plus, when feature creep
shows it's face, and I need additional I/O, I'd have to scramble even harder.
I've been looking at TI's MSP430F133. It seems capable, cheap enough, and there
is room for upward migration in the family. I'd appreciate hearing about any
good or bad experiences that people have had with this or any other MSP430 part.
Also if you know of any other microcontrollers that might be suited for my
requirements, I'd like to hear about them.
-Mike
--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email list...@mitvma.mit.edu with SET PICList DIGEST in the body
Barry Gershenfeld
>microcontroller that will get the job done.
>A PIC16LF877 with two more 8-bit I/O ports would be great. I've tried to come
>up with a reasonable way to use a the PIC16LF877 with I/O expanders. It can be
>done, but processing time would increase immensely. Plus, when feature creep
>shows it's face, and I need additional I/O, I'd have to scramble even harder.
How about two processors? It works for intel :) The second one
wouldn't even need to be an '877. You get more horespower at
the same time. Like having a "smart" I/O expander.
Barry
Dipperstein, Michael
>
> >I've just been given a new project to work on, and I'm
> searching for a
> >microcontroller that will get the job done.
>
> >A PIC16LF877 with two more 8-bit I/O ports would be great.
> I've tried to come
> >up with a reasonable way to use a the PIC16LF877 with I/O
> expanders. It can be
> >done, but processing time would increase immensely. Plus,
> when feature creep
> >shows it's face, and I need additional I/O, I'd have to
> scramble even harder.
>
> How about two processors? It works for intel :) The second one
> the same time. Like having a "smart" I/O expander.
It's a nice idea, and might actually work. I don't know how it compares to a
single TI processors as far cost space and power consumption, but I do have a
fairly clean partition of tasks (and I/O). The communication between each task
is minor and can easily be done over I2C. An '877 and an '872 would do the job
just fine.
-Mike
Alice Campbell
processing architecture' in the advert.
a.
> >I've just been given a new project to work on, and I'm searching for a
> >microcontroller that will get the job done.
>
> >A PIC16LF877 with two more 8-bit I/O ports would be great. I've tried to come
> >up with a reasonable way to use a the PIC16LF877 with I/O expanders. It can be
> >done, but processing time would increase immensely. Plus, when feature creep
> >shows it's face, and I need additional I/O, I'd have to scramble even harder.
>
> How about two processors? It works for intel :) The second one
> wouldn't even need to be an '877. You get more horespower at
> the same time. Like having a "smart" I/O expander.
>
> Barry
Dipperstein, Michael
>
> oooo, cool, and you could call it 'massive parallel
> processing architecture' in the advert.
>
> a.
Don't say that too loudly. Our marketing people will be calling the people that
produce all the glossy paper, trying to get a picture of our tester next to an
SGI Origin 2K.
-Mike
> > >I've just been given a new project to work on, and I'm
> searching for a
> > >microcontroller that will get the job done.
> >
> > >A PIC16LF877 with two more 8-bit I/O ports would be great.
> I've tried to come
> > >up with a reasonable way to use a the PIC16LF877 with I/O
> expanders. It can be
> > >done, but processing time would increase immensely. Plus,
> when feature creep
> > >shows it's face, and I need additional I/O, I'd have to
> scramble even harder.
> >
> > How about two processors? It works for intel :) The second one
> > wouldn't even need to be an '877. You get more horsepower at
Brian Aase
(at the risk of sounding like a heretic)
>I've just been given a new project to work on, and I'm searching for a
>microcontroller that will get the job done.
>
>A PIC16LF877 with two more 8-bit I/O ports would be great. I've tried to come
>up with a reasonable way to use a the PIC16LF877 with I/O expanders. It can be
>done, but processing time would increase immensely. Plus, when feature creep
>shows it's face, and I need additional I/O, I'd have to scramble even harder.
--
William Chops Westfield
and long battery life. At present I have identified a requirement for
39 I/O general purpose digital pins, and one low resolution ADC. Most
of the I/O is not well suited for I/O expanders. I could probably get
away with 8K of instruction space, 256 bytes of data space, and 2
bytes of non-volatile data space for storing configuration information.
TI's MSP430 series sounds like a good match to me...
BillW
Dan Michaels
>>I've just been given a new project to work on, and I'm searching for a
>>microcontroller that will get the job done.
>
>>A PIC16LF877 with two more 8-bit I/O ports would be great. I've tried to come
>>up with a reasonable way to use a the PIC16LF877 with I/O expanders. It
can be
>>done, but processing time would increase immensely. Plus, when feature creep
>>shows it's face, and I need additional I/O, I'd have to scramble even harder.
>
>the same time. Like having a "smart" I/O expander.
I think Mchp has an old appnote about using a PIC73 as an I/O port
expander running as an I2C slave to the main processor.
http://www.protean-logic.com/ sells one of these, in case you want
to go take a look.
- dan michaels
www.oricomtech.com
==================
Kris Wilk
>I've been looking at TI's MSP430F133. It seems capable, cheap enough, and
>there
>is room for upward migration in the family. I'd appreciate hearing about any
>good or bad experiences that people have had with this or any other MSP430
>part.
Only good experiences with the MSP430s so far. I've only used the
MSP430F11x1 flavours, but the 133 (and it's close siblings) are all
essentially the same chips with more of everything.
Clean instruction set, excellent (if perhaps a bit complicated) clock
configurability, in-circuit programmable, flash, HW multiplier in the
bigger chips, etc. And the F149 has 60 kilobytes (!!!) program memory and 2
kilobytes of RAM (!!!). Quite a beast!
You can't beat the $49 flash development kit for the F11x1 either...the ZIF
socket alone is worth at least that much.
Kris Wilk
ReefNet inc.
www.reefnet.on.ca