Page 196 - Making PIC Microcontroller Instruments and Controllers

192 COUtltl G PUISES: A PnOGRA|I|IABLE tACHOfllEtEn

So on the one hand, high-count encodels are desirable at low speeds; on the other hand
they become a prcblem at high speed. A corcllary to fhis is that it is very difiicult to run
a shaft accuately at low speed wiihout some sol1 of gearing or belt reductiorl Think
about the problems you would encounter tying to run a shaft at I revolution per year
accurately directly with a moror
If our specification calls for a maximum speed of 3600 rpm our encoder cat?r?ot have
more than 300 counts pei rcvolution if we want to maintain the counting time. Using a
shorter time will help on the high rym end but will make things harder in the low rpm
region, Of coulse, we can also use an encoder with many fewer counts and have a per-
f-ectly good tachometer
The soflware we are using employs integer math, and the largest variable we can use
is a 2-byte word. The largest number 2 bytes can accommodaie is 65535- This means
that in the COIJNT instruction:
couNt PoRTA- 2, 250, r,t1

The number that ends up in Wl cannot exceed 65535. We can accommodate this
requirement by shotening the 250 millisecond (or less) time frame, but then we will
have problems with the 60-rpm end of the specification- The solution is to use the lowest
count that will ser,/e our purposes on the shaft encoder at the lowest speed we are inter-
ested in. If we use the 20 counts per revolution we discussed earlie! we will be getting
*
(20 3600 / 60) = 1200 counts every second and a qua er of that every 0.25 seconds
or 300 counts every second, This indicates that we could use an elcoder with a few hun-
dred lines per revolution if accurate slow speed irdication was an important consider-
ation. However, adding a higher count encoder to an existing shaft may not be trivial,
and adding a f-ew (even 1) equally spaced markers onto the shaft manually with a little
epoxy or paint may be adequate for what we need.
There is also the possibility that we could write software that would use different rcu-
tines for different speeds of the signal, but for this projecr let's keep it simple. Such
sophi stications can be added after we get proficient at doing the work at hand. The most
important thing to keep in mind is that you must understand the problem in a compre-
hensive way before you can create a solution.

DETECTIOI{

Next, we must consider the components and circuitry needed to actually react rdth the
signal we are trying to measue (or collect). The simplest way to do this is to react to
the changes in the light intensity either as reflected from marke$ on the shaft or as a
disturbance of some other kind in the vicinity. Hall effect sensors are a popular way of
detecting rotation in a dity environment because they are rot afected by anything
other than magnetic fields. The signal must be converted to a TTUCMOSlevel signal
that goes high and low reliably with every change in the stimulus. Fjgure 15.5 shows
one way to create such an instrument jnterface for the hall effect device.
Since the input to the PIC are Schmidt triggers, we would not nomally have to con
dition the signal for bounce and jitter