Is it possible to use a capacitive proximity sensor to detect fluid levels (empty/not-empty) in a polyethylene bottle?

[kimmccavit]

I'm looking for a non-contact method to detect when the fluid in a polyethylene bottle is below a predetermined level.  The fluid is dispensed from the bottle in small increments over a period of many days or even months, so the meniscus will drop very slowly.  There will be a capacitance change if the fluid is above or below the sensing plates, but it will occur at a very slow rate compared to a hand or finger in a more typical application. 

[bryce · Accepted solution]

A cap sensor (touch or proximity) can definitely be used to detect water level.  Since you only need a limited set of discrete levels, you can do this very well using a cap sensor device like the MPR121 or using Freescale Touch Sensing Software with a MCU.

Here's an example showing a simple implementation of the discrete method with one of our older devices, but it has some good images to show you how it was done: http://cache.freescale.com/files/sensors/doc/ref_manual/APPEFIELDRM.pdf

Jump down to page 10 and take a look at the picture to see.

If you want to try it out, I would recommend the MPR121 eval kit: Freescale Sensor Toolbox MPR121 Evaluation Ki|Freescale

You can then mock up a system with metal contacts around the container and wire them directly to the eval board.  The provided GUI will allow you to see the signal on the screen so you can determine if the setup will have enough signal to differentiate when the fluid passes the level that you want to detect.

[jimmiller]

IF the fluid is opaque or significantly effects light use a photo cell.

[kimmccavit]

thanks for the input….

We are pursuing a photo sensor option also, but the fluids we need to detect range from almost completely clear to completely opaque and the type of fluid being used is not known in advance. This makes the code to analyze the optical data more complicated, but it is probably workable.

I don't have much experience with capacitive proximity detectors, but if they could be made to work it would simplify the design for this product. Also, not having to penetrate the enclosure to use an optical approach would have an advantage since any penetration would have to sealed with a clear lens which would need to be periodically cleaned.