Please help to choose a microcontroller
Archiver
Posts: 46,084
Maybe I can help ya out...
I'm not 100% sure but I think the Higher level PIC processors from MicroChip do
support Floats.... Start up cost to program them can be pretty high. You have
to buy the Programer and either learn Assembly or buy a 3rd party compiler....
The time frame on the Stamp is in Milliseconds so you are not going to finish
anything in microseconds. The Basic Stamp is designed to be easy to learn and
get people involved in a great hobby. It does that very well....speed had to be
compromised. At 4,000 cycles a second it's not exactly slow by any means....I
can't do two things in one second!! [noparse]:)[/noparse]
There are also some Basic Stamp Clones that are pretty fast and support floats.
They are not MHZ fast...
http://www.basicx.com/Products/BX-24/bx24compare.htm
And the Atom
http://www.basicmicro.com/ProductDetails.aspx?productID=18
If your really into speed and can manage a book size PC...then the MiniITX
booting via Compact Flash with Linux or Dos can run at 933 MHZ. The interface
to the real world could be via the Stamp. You could datalog MB's of data
too...[noparse]:)[/noparse]
The biggest question is what is your application....tell us about what you
intend to build.....we can help you design it.
Tim
>>> radu@n... 11/19/03 07:52PM >>>
Hi to all,
I need some advices from you, if it is possible.
I must to develop a program for a microprocessor, microcontroller.
But, first I need to choose one of them (a model).
My problem is this:
The program must perform around of 300 aritmetical
operations ( add, sub, multiply and divide ) with real numbers (floating
point type) with a 9 digits precision. A tipical operation could be:
A1 = (B1 + C1 - 4/3 * D1) / F1
where A1...F1 are floating point numbers.
All operations must be completed in 20 microseconds.
I have this problem because I cannot evaluate the floating-point
algortihm.
Please, can somebody tell me a processor for this ? I dont need a pc
computer... Something little. The mathematical coprocessor is optional. Or
one place, one link to learn about that ?
Many thanks,
Radu Cosoveanu.
Project manager.
NETONWAVE.
To UNSUBSCRIBE, just send mail to:
basicstamps-unsubscribe@yahoogroups.com
from the same email address that you subscribed. Text in the Subject and Body
of the message will be ignored.
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
I'm not 100% sure but I think the Higher level PIC processors from MicroChip do
support Floats.... Start up cost to program them can be pretty high. You have
to buy the Programer and either learn Assembly or buy a 3rd party compiler....
The time frame on the Stamp is in Milliseconds so you are not going to finish
anything in microseconds. The Basic Stamp is designed to be easy to learn and
get people involved in a great hobby. It does that very well....speed had to be
compromised. At 4,000 cycles a second it's not exactly slow by any means....I
can't do two things in one second!! [noparse]:)[/noparse]
There are also some Basic Stamp Clones that are pretty fast and support floats.
They are not MHZ fast...
http://www.basicx.com/Products/BX-24/bx24compare.htm
And the Atom
http://www.basicmicro.com/ProductDetails.aspx?productID=18
If your really into speed and can manage a book size PC...then the MiniITX
booting via Compact Flash with Linux or Dos can run at 933 MHZ. The interface
to the real world could be via the Stamp. You could datalog MB's of data
too...[noparse]:)[/noparse]
The biggest question is what is your application....tell us about what you
intend to build.....we can help you design it.
Tim
>>> radu@n... 11/19/03 07:52PM >>>
Hi to all,
I need some advices from you, if it is possible.
I must to develop a program for a microprocessor, microcontroller.
But, first I need to choose one of them (a model).
My problem is this:
The program must perform around of 300 aritmetical
operations ( add, sub, multiply and divide ) with real numbers (floating
point type) with a 9 digits precision. A tipical operation could be:
A1 = (B1 + C1 - 4/3 * D1) / F1
where A1...F1 are floating point numbers.
All operations must be completed in 20 microseconds.
I have this problem because I cannot evaluate the floating-point
algortihm.
Please, can somebody tell me a processor for this ? I dont need a pc
computer... Something little. The mathematical coprocessor is optional. Or
one place, one link to learn about that ?
Many thanks,
Radu Cosoveanu.
Project manager.
NETONWAVE.
To UNSUBSCRIBE, just send mail to:
basicstamps-unsubscribe@yahoogroups.com
from the same email address that you subscribed. Text in the Subject and Body
of the message will be ignored.
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
Comments
They are basically PIC chips, but re-engineered
to support 1 instruction per clock-cycle, and
to support 50 to 100 MHz clocks.
The problem you then run into is getting a
high-level compiler for them. I understand
CCS has a modification on their 'C' compiler
which supports the SX.
Parallax has a special on a kit for these
right now. $180.00 for board, programming
dongle, chip, software, and docs.
--- In basicstamps@yahoogroups.com, "Radu" <radu@n...> wrote:
> Hi to all,
>
>
> I need some advices from you, if it is possible.
> I must to develop a program for a microprocessor,
microcontroller.
> But, first I need to choose one of them (a model).
> My problem is this:
>
> The program must perform around of 300 aritmetical
> operations ( add, sub, multiply and divide ) with real numbers
(floating
> point type) with a 9 digits precision. A tipical operation could be:
> A1 = (B1 + C1 - 4/3 * D1) / F1
> where A1...F1 are floating point numbers.
>
> All operations must be completed in 20 microseconds.
>
> I have this problem because I cannot evaluate the floating-
point
> algortihm.
>
> Please, can somebody tell me a processor for this ? I dont
need a pc
> computer... Something little. The mathematical coprocessor is
optional. Or
> one place, one link to learn about that ?
>
>
>
> Many thanks,
> Radu Cosoveanu.
> Project manager.
> NETONWAVE.
I need some advices from you, if it is possible.
I must to develop a program for a microprocessor, microcontroller.
But, first I need to choose one of them (a model).
My problem is this:
The program must perform around of 300 aritmetical
operations ( add, sub, multiply and divide ) with real numbers (floating
point type) with a 9 digits precision. A tipical operation could be:
A1 = (B1 + C1 - 4/3 * D1) / F1
where A1...F1 are floating point numbers.
All operations must be completed in 20 microseconds.
I have this problem because I cannot evaluate the floating-point
algortihm.
Please, can somebody tell me a processor for this ? I dont need a pc
computer... Something little. The mathematical coprocessor is optional. Or
one place, one link to learn about that ?
Many thanks,
Radu Cosoveanu.
Project manager.
NETONWAVE.