How to perform liquid level sensing through the bottom of a metal tank

I'm posting this set by step setup for folks who need to perform level sensing from the bottom of a metal container. I'll turn this into an App Note eventually.

1, Read my app note on transducers to get an understanding about how transducers work, how to select them and how to mount them externally.It can be found found at: www/ti.com/lit/pdf/snaa266 

2. Read my app note on how to configure the TDC1000 in a level sensing application from the bottom. I go over how transducer choices effect system design and how to configure the TDC1000. It can be found found at: www/ti.com/lit/pdf/snaa270   

Here’s pics of the assembly process:

 

Gluing transducer to container bottom. Notes:

  1. just enough glue to cover the transducer
  2. No need for electric isolation (glue isolates)

 

Apply light pressure to transducer while waiting for glue to set.

 

Side view showing amount of glue.

Top View showing amount of glue.

Hot gluing stage: provides transducer acoustic dampening, transducer protection, and wire strain relief

 

Adding 4 rubber feet to container bottom to protect the transducer

Test setup: Pot, and TDC1000-TDC7200 EVM, transducer connected to TX1/RX2 (pin 10, 11) on J5

Scope shot with level above 15 cm. Note: it is ~1V p-p, meaning higher levels are easily achievable.

TDC1000 register settings (here’s a good starting point):

Here is a graph of the TOF measurement using the demo while I’m filling it from empty to full (~16cm).

/cfs-file/__key/communityserver-discussions-components-files/991/TDC1000_5F00_7200_5F00_metal_5F00_pot_5F00_Setup_5F00_final_5F00_fill_5F00_test.txt

5 Replies

  • Thanks!

    Regards,

    Paul Grohe

    Low Power Amplifiers and Comparators (LPAC) Applications

  • I should note this works as long as the tank is moderately level (<10 deg) and the surface is relatively still.
    For containers with wider level variances or more heavily disturbed surfaces (fluid pouring in etc.) a level tube can be used to allow the TDC1000 to work in those circumstances.
  • Hi Mathew,

    Thanks for sharing this post. It is very interesting and I have some questions.

    I would like to measure the fluid level of a closed metal tank from the bottom and it is about 1,5m high (500L). My first guess is the fluid level is too much and the signal may fade before reaching the surface. Since the major application of this is to detecting when the tank gets empty, I also guess I coud have some signal back when the fluid surface is reaching the bottom.

    - What is the estimated maximum height for a metal tank, which inner fluid level can be measured with a TDC1000?

    - What is the estimated maximum height for detecting an empty tank?

    - Is the metal thickness important?

    - Real industrial stainless steel containers are not usually so flat as that pot. I guess if the bottom is tilted or irregular this may affect the measurement, right?

    - Industrial applications usually require more robust fixing systems than crazyglue and hot glue. Any recommendaton when it comes to industrialize this type of sensors?

    Regards, Ramon.

  • In reply to Jose Ramon hernandez:

    - What is the estimated maximum height for a metal tank, which inner fluid level can be measured with a TDC1000?

    This is something to be found experimentally as ultrasonic energy has to go through multiple mediums (transducer to gel-couplant to steel to liquid) before reflected back from the liquid level surface. While you use a gel-couplant to ensure all the ultrasonic wave from transducer is transmitted to steel without loss, steel is not a good couplant of the ultrasonic energy. The energy transmitted to the liquid is attenuated quite a bit and so higher the level you want to detect higher the excitation energy you need to apply. You can read about acoustic impedance and some basics of ultrasonic transmission here
    www.ni.com/.../

    - What is the estimated maximum height for detecting an empty tank?

    Transducers have a dead zone and cannot receive while ringing. For 1Mhz transducers, minimum height you can measure is about 1 to 2cms. You can read about it in the app notes I referred in the other posting. By using a delay or stub element (which also is a good couplant) between the transducer and the metal tank you can measure even more to empty.

    - Is the metal thickness important?
    Yes, there's more attenuation and hence less ultrasonic energy transmitted in to the liquid.

    - Real industrial stainless steel containers are not usually so flat as that pot. I guess if the bottom is tilted or irregular this may affect the measurement, right?
    You can use a cylinder of gel (or something shaped to the contour) to ensure there's good uniform coupling without air gap.

    - Industrial applications usually require more robust fixing systems than crazyglue and hot glue. Any recommendaton when it comes to industrialize this type of sensors?
    I agree. FYI, in app note SNA266, CYA is to ensure strong connection between the face of the transducer and the tank wall. Also, its low compressibility ensures there is no acoustic dampening. Whereas the glue, that is applied around the transducer circumference and backside, is for acoustic dampening (due to its flexibility) and wire strain relief. So the glue, if applied on the transducer frontside, can cause performance degradation.

    Thanks,
    Vishy
  • Hi Matthew,

    I am interested in your approach for measuring water level in tanks using ultrasonic sensing by a piezoelectric transducer. Please advise if I can use this method to measure water level in rain-fed containers (e.g. jerrycans, pails, tires) and at outdoors? How much this would cost?

    Best,
    Mekki