Exact match. Not showing close matches.
'[OT?] Soliciting PIC project sensor suggestions'
|Any help or comments would be very much appreciated.
I am working on a PIC based project utilizing multiple 16F84's. (Other
device recommendations cheerfully accepted :) It is basically a system to
determine the sequencing of six discrete objects moving through a
rectangular hole (or tunnel really). I am planning to attach a 'sender' to
each of the objects. The objects may be moving at up to 30 MPH and
separated by as little as an inch. The hole is about 6 inches wide, 14
inches high, and I have about 3 feet in length where I can deploy some kind
of sensor array. The whole assembly is outside where it is exposed to
sunlight. Some of it may be shaded. It does not need to be weatherproof.
I am considering a RF based solution, but am not certain how to solve the
various problems. For example, how to solve the problem of the second
device having a stronger signal and erroneously being picked up before the
first. i.e. How do I make sure that the one sensed first is actually the
one physically first in line?
I have also considered infra-red. I am not sure of the complications that
sunlight might cause. I have found this mentioned in a number of places,
but not much useful information. Dust may also be a concern, as is reliably
detecting the signal. So far, it seems like the infra-red projects I've
seen use a relatively low frequency pulse train that would not generate
enough pulses to reliably sample in the approximately 2ms that 1 inch at 30
For either approach I am thinking of using 6 receivers each 'tuned' to a
particular pulse train. Each receiver would toggle and latch it's output
when it senses it's particular pulse pattern. Another stage would simply
scan the 6 inputs rapidly and determine which of them was first, second,
etc. It's all straight forward from there.
TIA for your input.
Sr. Systems Engineer
OEM, Carrier Sales, Healthcare
184 Shuman Blvd., Suite 200
Naperville, IL. 60563
Phone: 630-717-2960 x 3072
Why not just slap a modified bar code onto the target area?
By modified, I mead one with wider than normal stripes so that
it can be read consistently at the 30 mph velocity. Since
there are only six objects you need to determine the sequence
of, why not use a simple coding scheme consisting of a
start code and a 3 or 4 bit binary identification code?
The start code could consist of something like a double-width
black bar followed by a double width white space followed
by a 4 bit code in which a double width black bar followed
by a single width white space is a "1", and a single width
black bar followed by a double width space is considered to
be a "0". Such a scheme can be made to be tolerant of wide
variations in bit timing due to changes in velocity. The initial
start code is used to teach the PIC the *current* value of
a "double width". A simple right-shift will give us the
value of a "single-width". Shift the single-width value
right (dividing it by 2), and add the result to the single-
width value to arrive at a 1.5 value. When the beginning of a
black bar is detected, time out for 1.5 value and then look
at the current bar. If it is black, then the value is a "1".
If it is white, then the value is a "0". Very simple to
A visible laser beam could be the light source, and a photo-
diode and preamplifier followed by a simple comparator could
be configured as the detector circuit. Use a filter and
cylinder with recessed lens/pickup to eliminate as much
ambient light as possible. Shade from direct sunlight!
Adjust the comparator so that it only responds to the laser
You would only need a single laser/photodiode assembly.
The items would come by sequentially, and each would cause
its own unique ID code to be sensed in the order received.
A PIC should be capable of reading the desired ID code within
the 2 ms window with ease.
Hope this helps.
Fr. Tom McGahee
More... (looser matching)
- Last day of these posts
- In 1999
, 2000 only
- New search...