I have a small idea/project I am working on in which I want to use a water fiber optic sensor that can detect whether the water stream is broken and if so shut off the water.

Having looked into fiber optic sensor systems in general, all I have managed to work out is I would need an intrinsic sensing region system. I do not know what type of Optical Modulation mechanism I would need so have hit a wall.

Any guidance into which of these systems, Intensity modulated, Phase modulated, Wavelength modulated or Polarization modulated, could be useful would be massively appreciated as I'd love to be able to look into this further.

  • $\begingroup$ Also does water flow have to be laminar for light to travel down it? If so at what Re does the turbulence become severe enough to disrupt the lights path? $\endgroup$ Commented Feb 20, 2016 at 14:26

3 Answers 3


As long as you are not sending data, any light input in one end should be detectable with a color filter covered light sensor at the other end. Example: green film for a green led/laser. Light at the sensor means the beam is not blocked, no light means it is blocked. No fancy modulation should be necessary unless you are trying to filter out similar background colors from the sensor. A simple on-off pulse at 1000Hz would do well for this purpose if background light noise is a problem. If lots of these are used in close proximity you may want to pick a different frequency for each. For a small beam a LED should be fine, but depending on water impurities, diameter of the stream, and length of the stream a laser may be required.

Another thing to consider is stream diameter and bend radius. There is probably some complex math involving a lot of variables we don't know, but basically as your stream(fiber) gets larger diameter or the bend radius gets smaller, more and more light escapes the stream(fiber). Here is a good article on it. In the image below you can see how light escapes the "macrobend". You will want to reduce this by decreasing the stream diameter or increasing your bend radius so more light makes it to your receiver. enter image description here

This 940nm infrared receiver and transmitter may also be an option. This has the added benefit of not being visible, but will potentially have less range because water is more absorptive at that frequency. If needed they do make powerful IR laser emitters such as this 808nm IR laser diode

You will want to choose light source that is in the visible or near infrared regions. Other frequencies are absorbed much more quickly. See full wiki article and graph below. Water Electromagnetic Absorption Graph

  • $\begingroup$ Massive thank you for your guidance, this have given me a lot to look into! I understand it can be simple to create the system with the LED emitting light through one end of the 'fiber' with a light sensor at the other, but is it possible to incorporate a sensor into the same end as the LED, which could detect that the lights path is uninterrupted/broken along the stream? I understand how this will be limited to length of stream but can it be done? $\endgroup$ Commented Feb 23, 2016 at 12:24
  • $\begingroup$ While possible it would be very very tricky. You would have to do a time of flight system with very very good noise filtering and geometric fudge factors. Here is a time of flight distance measuring system: fluke.com/fluke/uken/Laser-Distance-Meters/… So possible, but the beam path in the fiber is not straight and many other variables will come into play that will cause issues. I wouldn't recommend attempting it. If possible, share your desired end result for this application and we can likely find you a much simpler solution. $\endgroup$
    – ericnutsch
    Commented Feb 24, 2016 at 0:43

Fiber optics depend on the surface reflecting light by total internal reflection. Therefore the properties of the outer surface are crucial, and reflection depends on what your understanding of a "stream" is.

If the water is in direct contact with anything solid (or liquid) (most likely anything but air), it will not work. Total internal reflection depends on the outer material being transparent (for the wavelength considered) with the refractive index being smaller than that for water (1.333).

If your surface is not smooth (having air bubbles or the like), you will most likely get a significant intensity drop of transmitted light. Detection of a broken stream in such cases will depend on the power of your light source, the sensitivity of your detector and the length of the stream.

As long as the surface is smooth, it should work. However, my everyday life experience with water tells me that freestream liquid jets tend to become unstable and separate into drops. I presume there are a lot of publications out there on this phenomenon. This is something that would need further investigation.

Also, a curly surface might cause troubles. I don't know if there are any investigations on this, but you should have a look into this.

Regarding your comment on turbulence: Turbulence alone does not influence transmisison of light in water. However, it influences the outer water surface of a free stream. Nice pictures of this process are here: http://dx.doi.org/10.1016/j.ijmultiphaseflow.2013.11.013 (open access article). They show that turbulent free streams are unstable and disperse into small drops.


Is it possible?

Yes, it is certainly possible. It may even be easier than you are assuming.

Water stream light pipes

Using a water stream as a light pipe is a somewhat common science experiment. You can look up various videos and how-tos, but the general principal is that the light will follow the water stream as it flows.

An example project shows this image of the light in the stream:

enter image description here

Once you have the light going through the stream, all you need is a detection method which may be a simple as a light sensor. Depending on the amount of ambient light, special methods might be required to filter out unwanted light. The modulation might not be required.

I'm not sure how long of a length this will work for though.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.