Go Down

Topic: Non-contact proximity sensor (Read 2840 times) previous topic - next topic

anfred38

Hi,

Can anyone recommend a non-contact proximity sensor to detect metallic objects?

I am trying to build a project to detect and count the number of  plastic balls travelling in a pvc water pipe system.  The plastic balls are embedded with small metallic bits to reduce their bouyancy in the pipe.  

I need to detect how many of the plastic balls left and return to the collection chamber, so that I can calculate the time taken by the plastic balls trip through a network of pvc water pipes.

Thanks.

Fred

RuggedCircuits

How far away will the balls be from the sensor?

The first thing I think of is an inductive proximity sensor. This is an industrial piece of equipment that is relatively expensive. Cheaper would be a Hall effect sensor but I'm not sure of the range.

drspectro

some brainstorming
If the metal is ferromagnetic, a compass module might be more sensitive.  I would expect the response time would be on the slow side.

For non magnetic metals, you could do an oscillator with a coil and measure the frequency change.  Something like a Joule Thief with no ferrite core.  Or maybe modify a simple metal detector circuit then use the arduino as a frequency counter.  

OR put in a small section of food grade clear PVC and use  visible light LED and photo transistor.  The electronics for that are all over the forum.

OR if the balls make a "click" when they hit an elbow use a piezo and an amplifier and do a "knock" sensor.  glue the piezo to the PVC pipe.  I played with this to see if I could tell when a faucet was left running and I could easily detect a small drip.  


ohnoezitasploded

Or capacitive sensing?
http://www.arduino.cc/playground/Main/CapSense

anfred38

#4
Jun 07, 2009, 11:13 am Last Edit: Jun 07, 2009, 12:22 pm by anfred38 Reason: 1
Hi,

Thanks for all the suggestions.

The pipe diameter is 50mm at the outlet and the inlet, so I assume the farthest distance the object is away from the proximity sensor would be no more than 50mm.  

The diameter of the metallic core plastic ball is 15mm.  The metallic core is non-ferromagnetic.  

When the pump starts, the containing chamber will open the gate, and release all the plastic balls, then it will close the gate at the outlet and open the inlet gate, after the proximity sensors has detected all the plastic balls had left the chamber.  The circuit will record the time the first plastic ball and the last one left the chamber, likewise for the returning plastic balls, before closing the gate at the inlet, and shutting down the pump, after the last plastic ball return to the chamber.

As the pipe diameter is 50mm, there would be more than 1 plastic balls will passing by the proximity sensor at the same time.

Here is a brief sketch, maybe it would be easier to visualize.



Thanks.

Regards
Fred

ohnoezitasploded

Just out of idle curiosity, what's this for?

Grumpy_Mike

#6
Jun 08, 2009, 01:48 pm Last Edit: Jun 08, 2009, 01:50 pm by Grumpy_Mike Reason: 1
Quote
there would be more than 1 plastic balls will passing by the proximity sensor at the same time.


That is the killer which makes this project very difficult to do. Inductive proximity detectors won't distinguish between multiple targets and a single target.

The other big problem is that as there is water flowing a lot of metal detector / RFID type solutions won't work because of the conductivity of the fluid.

Magnets and hall effect sensors won't work due to the fact that you can't guarantee the orientation of the magnet to the sensor.

The solution that would work is not open to you. Load a different radioactive isotope into each ball and have peak gamer detector identify what combination of signatures are passing the detector.

AWOL

#7
Jun 08, 2009, 02:32 pm Last Edit: Jun 08, 2009, 02:34 pm by AWOL Reason: 1
Very bright light and a line-scan camera?

X-ray source and scintillator/photodiode array.
"Pete, it's a fool looks for logic in the chambers of the human heart." Ulysses Everett McGill.
Do not send technical questions via personal messaging - they will be ignored.

mircho

This guy scrambles and then counts 1.3 million dice rolls a day.
http://gamesbyemail.com/News/DiceOMatic

Maybe if you have a clear tube and color (black) balls it might be possible to analyze very fast shots of the tube to see if there are black spots in them and then count them. You'll need a more powerful processor though.

anfred38

The only thing I can do, is to make a pipe reducer, so that only 1 ball enters or leaves the chamber.  Maybe that way it would be easier.

This is actually a class project I am doing on probabilty.



RuggedCircuits

Oh absolutely. If you can reduce the distance to the balls and guarantee no more than 1 ball at a time you could even use a simple optical reflector.

You'll just have to make sure the balls don't jam up at the input of the reducer.

anfred38

We will need to experiment with a reducer that won't cause the balls to jam up.  Back to the drawing board.

Go Up