On 1/17/2014 12:30 PM, Jon Hulatt wrote:
> I definitely will try assembly once I have the logic figured out. I wrote=
 the above code and I just have this niggling doubt that that is the best w=
ay to do it, i wonder if i=92m missing some major optimisation or something=
..
>
>> Your code aside, if you're looking for speed maybe you should be writing=
 it in assembly?
Counting quadrature tachometer signals requires the OLD data as well as=20
the NEW data, cleaned up to prevent random errors from noisy contacts.

1. read the tachometer data repeatedly until the data is identical to=20
the previous read; this eliminates noise from the system.

2. combine the LAST two bits and the NOW two bits to form the address of=20
a 16-position jump table. ONE of the addresses means down-count and=20
another means up-count; which one depends on how you have organized the=20
data. The other 14 addresses are meaningless so are simply ignored.

3. I would interrupt on ANY data change by the two bits, by using the=20
INT ON DATA CHANGE function of most PICs.

The faster you can sample the data the faster the speed of the shaft can=20
be read, so your PIC needs to run as fast as possible and be written in=20
assembly. BUT you can write in C or whatever to verify your approach.

its more complex than it looks.

-- Bob Axtell

The only place success comes before work is in the dictionary.

VINCE LOMBARDI

--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.