> Harry H. Arends wrote:
>> Is there a need for a piece of debouncing check on slow 
>> moving HALL
>> sensors connected to a MCU?

Olin escrit
> Any signal can have noise.  It would be a good idea to use 
> hysteresis at
> least a bit greater than the noise.

What he says.

While in many cases the hysteresis which is often included 
in Hall sensors will be enough, the result will vary with 
application. Depending on how the magnetic field variation 
is produced you may get "noise" or may not. If you are 
monitoring a rotating shaft (or similar object) then 
inertia/momentum/natural cussedness will mean that you 
probably only see 'vibration' bucking the trend and most 
Hall sensors will probably work OK. If you are monitoring 
something like a door / safety interlock etc then you may be 
more at risk. But even with a rotating shaft "vibration" may 
be due to eg bearings beginning to fail or an out of balance 
load, so what works this week as an RPM monitor may work 
next month as a bearing failure or out of balance load 
detector.

Also, as Olin suggests, noise may be injected electrically 
in some manner - circuit impedance, inductive pickup and 
signal strength (and more) will help affect how great the 
resultant noise is. "Relay turning off / inductive kick" 
transients, which love to ride in on all sorts of other 
signals may find a magnetically coupled path for this one.

Looking at the signal on a scope and comparing the observed 
variations with the sensor's hysteresis spec  (or measured 
hysteresis characteristic in the absence of a spec) will 
give you a rough feel for what is or isn't required. Failure 
in service in equipment 4000 km away will give you a finer 
feel :-).


        Russell




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