Limit field of view for a narrow level sensor

[Samot No]

Hello,

I am building an IoT level sensor for a vessel. The vessel is emptied at the bottom, and I would like to sense how much stuff is still in there from the top lid. The tank is approximately 40cm deep, and 10x10cm wide and it is dark inside.

Considering the wide viewing angle of the proximity sensors (25 degrees) and the narrow vessel, I'm worried that I will get into trouble with false readings, so I was wondering if I could limit the view of the sensor, or that you might know another way of reasonably accurately (+/- 0.5cm) measure the amount of salt in the vessel.

Kind regards,

Samot

[Dominique LOYER]

​Dear Samot,

If you use the VL53L1X sensor, you can reduce virtually the field of view on the receiver by changing the ROI size, minimum is 4x4. As you want to be very accurate, use a long timing budget like 100ms. and use the SHORT distance mode.

If it's still too large and you see the wall of the vessel, you can see if this contribution in the range error can be consider as an offset.

Best regards

[John E KVAM]

The L1X Regions of Interest (ROI) are a great idea, but the L1X is primarily a long-distance chip and might give you the wrong answer at the very, very short distances. So test that. If you can back the sensor into the body of your housing by a couple of centimeters, you will be fine.

Other solutions:

Calibrate the container. Light hitting the sides and returning will lead to a wrong answer, but that wrong answer should be consistent. So with a little calibration of your contain you should be able to match the ranging result to the actual fullness.

Salt is really reflective. If you paint the walls of the container black, the number of photons from the walls will be overwhelmed by those from the salt and give you the right answer. Might not work if the salt sticks to the sides of the container though.

I have also found that building a cap over the Transmit side of the sensor and putting a small hole in it will block the light from spreading. The trick is to do the geometry to calculate the size of the hole and the thickness of the cap. But the cap has to be optically black in the IR region - which takes some work. (Black rubber works well.) And of course you have to completely block any stray light from hitting the cap and bouncing back immediately.

But one has to admit that using the VL53L1 and ROIs is a pretty easy solution.

[Samot No]

Thanks for your fast responses. I do not have the option of painting the walls, I have to work with how it is. Due to the fact that the L1X is a long-distance chip, I just glanced over it, but the ROI settings might be just what I need, 15 degrees FoV is just small enough. The cap might indeed also work, I will consider the option too.

I was wondering, does the sensor has a voltage dependence? All measurements in the datasheet are done at 2.8V, but does it get better or worse with higher or lower voltages?

[Dominique LOYER]

Hello,

Good, let us know if the ROI fullfill your needs.

About the power supply range, no impact in term of performance.​ If you use a power supply between 2.8 to 3.3V, we recommend toenable the compiler option power 2V8 so the I2C standard will be perfectly compliant.

Best regards

ST support

[Samot No]

Hi,

It took some time to try it out, and I forgot to answer you. But for the record, setting a smaller ROI works great!

Best regards!