Connect with us

Water detection

Discussion in 'Electronic Design' started by Barry Lennox, Apr 12, 2007.

Scroll to continue with content
  1. Barry Lennox

    Barry Lennox Guest

    I have a requirement to detect a water level in a clear plastic tube
    of about 20-25 mm diameter. This must be done non-invasively, so I'm
    considering either optical or capacitive sensing would be best

    Has anybody done this? If so, what works best, and if optical, what
    part of the spectrum works best (UV, red, green, IR) ?

    The sensing can be considered "binary" ie, it's either above or below
    a given level in the tube.

    Thanks

    Barry
     
  2. Joerg

    Joerg Guest

    I've only done it with ultrasound but the cheapest method would probably
    be plates that are part of an oscillator. When water reaches that level
    the oscillator loses lock and quits oscillating. The latter is quite
    popular in industry (although the ones I've dealt with were all magnetic
    coupling). Power goes down, triggers a Schmitt gate, bingo. Some uC such
    as the MSP430F2xxx family have Schmitt action built into their ports.
     
  3. D from BC

    D from BC Guest

    Here's an idea..
    How about a trip to the local scrap yard and look at how cars detect
    fluid levels..
    Windshield, brake and oil fluid may have an electronic sensors.
    Pullm off and maybe getm for cheap.

    What do you mean by non-invasive?
    Some fluid level detectors have a magnet on a float.
    At low level the magnet reaches the hall sensor (or reed sw) and trips
    the "you're out" light..
    Do you mean no floats allowed?

    Clear water + clear plastic + optical level detection...mmmmmm :(
    Light conductance of air vs water .... ohh that's a toughy....
    Light reflection due to plastic/water refractive index....
    possibly.....

    Take care..optical methods may have to compete with ambient light..

    Capacitance method: A 25mm water dielectric.... ewwww
    D from BC
     
  4. How about a bubbler tube and a dye in the liquid?
     
  5. Jim Thompson

    Jim Thompson Guest

    Coil around tube at the "threshold" point. Coil is part of an AGC'd
    oscillator. AGC voltage reflects losses in water when it reaches
    coil.

    ...Jim Thompson
     
  6. How clean is the water and tube? If its murky, or the tube is likely to
    become contaminated over time, then optical is out. What is the
    thickness (and composition) of the plastic?. And what is the required
    accuracy of the level detection? This will have to be considered if a
    capacitive pickup is designed to be applied to the exterior of the tube.

    If optical sensing is possible, just aim a light source (cheap laser
    diode) through the tube offset w.r.t. the center of the tube. As the
    water level rises and crosses the beam, the change in index of
    refraction between the air and water will change the path of the beam. A
    photodiode placed appropriately will sense the shift in the beam path.
     
  7. Paul Burke

    Paul Burke Guest

    If you're happy with discrete levels, and the water's clean enough, ,
    just have an array of narrow angle LEDs on one side, and an array of
    phototransistors on the other side. Place the sensors at the focal point
    of the tube full of water, and use a comparator to dicriminate between
    the bright illumination of the full tube and the relatively dim
    illumination of the empty one. You could add a reference tube, or a pair
    above the maximum water level, to make the measurement ratiometric.

    Paul Burke
     
  8. mpm

    mpm Guest

    If it's a closed system and you can plug one end, you could try
    measuring the air pressure above the column of water and calibrate it
    for whatever set point you need.

    -mpm
     
  9. colin

    colin Guest

    Salvage a level sensor from an old washing machine.
    Im not sure what they use nowdays,
    but when I worked on the design of some of the first wasing machine
    controllers
    they were realy keen to use something cheaper than the pressure sensor types
    used for the cam wheel type controllers.

    reading the oscillator agc voltage suggestion seems good to me.
    I gues if you made it a long coil you could even make it linear :)

    Colin =^.^=
     
  10. Bob Eld

    Bob Eld Guest

    I designed a non-contact water level sensor for a large tank. It worked on
    the capacitive principle by changing the frequency of a multivibrator when
    water flooded the dielectric region of the capacitor. The detector was built
    inside a short piece of one inch PVC pipe arranged with an inner pipe and
    foil electrodes. The electrodes were on opposite sides of the pipes from the
    water. The PVC was machined to fit and accomodate the electrode foils. The
    water entered the region between the inner and outer pipes. The electrodes
    form the plates of the capacitor but do not touch the water. All electrical
    circuitry and electrodes weres sealed in the pipes and protected from the
    water and moisture. There is no electrical connection to the water.

    Water has a dielectric constant of 80 so it increases the capacitance of the
    capacitor when present and decreases the oscillation frequency or pulse
    timing. When water is not present, the oscillator frequency is higher. A
    second fixed frequency multivibrator operating in sychronism with the
    detector multivibrator formed the reference. The operating frequency was
    about one MHz . A D flip-flop compared the timing of the two multivibrators.
    It was connected so that when one time was longer than the other, the output
    was high and when shorter, the output was low. The high/low condition of the
    D-flop indicated the presence of water or not.

    I think you could do something like this for your application. Optical
    methods work but are more tricky and are easily affected by contamination,
    algae, or water deposits. The capacitive system is more fool proof and can
    be made very low power. Also avoid any system that makes contact with the
    water because corrosion and electrolosis can affect things and can
    contaminate the water.
    Bob
     
  11. Joerg

    Joerg Guest

    You could also hang the cap into the feedback path. Dimensioned right
    this would provide a digital signal because the oscillator cuts in and
    out depending on the water level. A sense resistor and transistor in the
    supply path could act as an OC digital port.
     
  12. Apart from pressure sensors and this method you describe, there is also 'echo',
    use a 555 timer and some piezo sensors and echo the open pipe from the top.
    This system is pretty good, but gets confused if there is foam or stuff floating
    on the water.
    Somebody mentioned to me once the following solution:
    Place mike and small speaker in opening pipe, amplifer feedback, thing will resonate
    at a frequency related to the area where there is no water.
    So sort of an organ pipe :)
     
  13. Joerg

    Joerg Guest


    And then the people wonder why their dog doesn't want to go outside
    anymore...
     
  14. LVMarc

    LVMarc Guest

    why not make the tube part of a resonator in an oscialltor. As the
    column proportionately fills; the air is replaced by water and a
    dielctric constant shift experienced. The shift is proportionate to
    column fill level. you may obtain a continous water lvel to Frequency
    transducer this way!

    Marc
     
  15. Barry Lennox

    Barry Lennox Guest

    Many thanks for all the replies thus far. To answer most of the
    questions:

    Non-invasive means that I cannot get to the inside of the plastic
    tube.
    I have not yet seen the tube to be used, but think it's PVC or
    polythene, and has a wall thickness of 2-3 mm.
    The water is generally dirty, and flows fairly freely, so pressure
    sensors will be tricky.
    The application is subject to a fair amount of movement and vibration,
    so float switches may be a problem.
    I can control the ambient light in the application.
    EMI is a real concern.

    Later today I can get to the site and determine the tube to be used,
    and hopefully get a sample piece. I'll then do some trials with
    "capacitor plates" either side of the tube; a laser diode or LED, and
    maybe a coil around the tube.

    I'll report back when it all works.

    Thanks
    Barry
     
  16. IME a bubbler will work best. You bleed compressed air into it slowly and
    the back pressure operates your device or system or indicator. This will
    even work with flammable or opaque liquids.
     
  17. jasen

    jasen Guest

    capacitative could be hard, how thick is the tube wall?
    Any colour will work, arrange it so that the beam is difracted away from
    (or towards) the sensor when the water is present,


    Bye.
    Jasen
     
  18. Ken Moffett

    Ken Moffett Guest

    Sorry to jump in down-thread, but I missed the original post.

    Single or two point capacitive level detection can be cheap
    and easy on non conductive containers and lines.

    Take a look at Quantam Research Groups capacitive sensor
    chips. For example:

    http://www.qprox.com/downloads/datasheets/qt114_104.pdf

    They have lots of good supporting technical data, and are
    available from Digikey.

    I've used some of their chips for fluid level detection and
    control circuits. One was to control the level alcohol in a
    1L ,6"daimeter plastic lab jar. Detection and comtrol was
    within fractions of a mm. In another forum someone wanted to
    detect the presence, and absence, of fuel in a plastic line
    on an aircraft. For fun I breadboarded a circuit that would
    do this:

    http://i7.tinypic.com/2rxgtgp.jpg

    This moght do the trick for you Barry.

    Ken
     
  19. Jamie

    Jamie Guest

    small low signal narrow R.F. TX, RX on the side of the tubes, R.F. does
    not pass well through fluids.
     
  20. Barry Lennox

    Barry Lennox Guest

    Ken, Thanks for the info, it looks quite promising.

    Barry
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-