Non-invasive level detection in plastic bottle


Thread Starter

Mike Fino

I'm looking for a non-invasive/non-contact level sensor with switching PNP output, DC excitation that can detect a low dielectric liquid within a polypropylene (0.25 in wall) 5-L bottle.

I've tried a capacitive sensor but the liquid appears to have too low of a dielectic. These bottles are used as a vacuum trap - an intermediate liquid storage before it can be pumped off to a final container. I need the sensor to tell the pump when to turn on so the liquid doesn't get too high and go into the vacuum
system. Other constraints: there are many of these bottles and only a few machines that they are used on. I'd like to spend, at most, US$500 per machine for some type of sensor. So the bottles themselves would just be place into a "nest" that contains this sensor. I'm currently wired for +/- 24VDC and a PNP output to be recognized by a PLC. This was for a capacitive sensor that worked with water runs but not on the
final media. Ideally, I'd like to plug an alternate sensor in place, but changing the PLC code is an option.
An ultrasonic sensor might be your answer. I think SICK sensors carry these. You also might want to try Pepprel & Fuchs(?) Give some of these manufacturers a call, they are usually straigh up about whether or not they have something for your

david mertens

Why not use a weight sensor, they come fairly cheap if you don't need good accuracy. This could even be a simple binary control with two set weights, one where the pump is switched on, the other when it has to be switched off. So two digital inputs could be all you need, or it could even be directly wired to a relay that operates
the pump, without need for a PLC.

Steve Myres, PE

The weight was the first idea that came to my mind too. If the discharge line to the drain pump penetrates the wall of the bottle rather than through the top, you could also use a very low range pressure switch in the feed line to the pump. This won't work with the line through the top, as the line is empty when the pump is turned off. I had a similar application on a wet bench project, and I made a modified bottle lid with a pair of probes for conductivity level measurement.

Jimmy Saldivias

Hit me if I am wrong, but I think ultrasound will only sense the presence of the bottle, not the level inside it. Or maybe I did not understand the
question. If I did understand correctly, and if the bottle is clear, and the liquid is colored, in that case you can use light to detect level.
If the liquid is conductive you can insert a couple of wires into the bottle and check for conductivity. Whenever there is liquid, you'll have conduction among the wires.
Hope this helps.

MBA Ing. Jimmy Saldivias
Phone: 591-4-4523438
Fax: 591-4-4523413
There are ultrasonics with a teach function. Simply put you would be adjusting the sensitivity level that would distinguish between an empty and a full bottle.
Is there a significant temperature difference between the liquid in the bottle and the ambient temperature? Maybe an infrared sensor could work.
Another approach is to angle a narrow angle optical pair so that diffraction at the air-liquid interface deflects the beam. This works with various liquids. An angled pair that makes from surface reflection can work also. A laser and detector can be very precise used this way.


It is possible to use bottle weight as a process parameter to control a drain pump. To address earlier questions:

a) the bottle is polypropylene and is translucent but not clear,
b) the media I'd like to detect is pink/red and photoelectrics look to be an option and one I plan to test,
c) the discharge line is through the top of the
d) there can be no detection methods that actually go inside the bottle or physically contact the media,
e) the capacitive sensor I used was from Carlo Gavazzi and was able to detect water and saline thru the bottle (0.25 inch wall) but not the media I plan to use.

The mechanics of a weight system are just a touch more complicated that just strapping a sensor to the side and using a switching output to control the drain pump, which is my preference.

All the comments so far are much appreciated -