measure target through a tube

[anotherandrew]

I can see from the datasheets (VL53L4CD for example) that the emitter and sensor volume is represented as two cones that start at the sensor face and grow bigger as the distance from the sensor increases.

If I am measuring the distance to a target that is visible through a "tube", can the sensor figure out that the walls of the tube are not the target and ignore them, or do I need to ensure that the tube diameter is larger (+ some margin) than the diameter of the sensing element's "cone" as shown in the datasheet?

[John E KVAM]

This query is more common than one might guess. I call it the piston problem.

If the sides of the "tube" are specular (mirror-like) then the photons will

bounce off the sides and continue on their way. Might add a bit of distance to

the measurement, but it will be small.

If the walls are Lambertian (matte finish) then the photons will scatter and some

will definitely make their way back.

So you could make the tube wider than the Field of View, or you could paint the

walls a very black black. Is optically black anodized aluminum a possibility?

You could also simply brighten up the piston. With zillions of photons coming from

the piston, and only a few from the walls, the 'bad' ones won't affect the

average very much.

Another thing you can do is calibrate the system. If you know the right answer, and the

range return, you could construct a look-up table to account for the

differences.

the VL53L4CD has a FoV of 18 degrees. To get that, the full intensity of the light

covers that 18degrees. But at the edges, the light rolls off. So if you want to

insure the light at the edges does not reflect, your cylinder has to be wider

than the 18degrees. I'd go with 22-25 to be sure. (Datasheet says 25 degree exclusion area, but you can cheat a little, depending on how many 'wrong' photons you can live with.)

In the end, a combination of all of these ideas would be best.

A cylinder as wide as is reasonable. Polished, dark walls. Bright piston. And if you need it, a calibration table.

[GWerth]

Hi John - I have a similar question/application. We'd like to mount a sensor at the top of a silo to do level measurement of the contents of the silo. The contents could be sand, salt, grain, etc., but not liquid, so a simple single depth point result will not be sufficient. The goal is to determine the volume of the material sitting in the silo and the challenge is to determine the surface contour of the contents, i.e., multiple depth points. The silo is about 2 to 2.5 meters in diameter and about 8 meters high. We're thinking about using the VL53L1 sensor since it has a range of 8 meters and we can use multizone to get multiple depth points. Would very much appreciate your input on the best approach. Thanks in advance.

[anotherandrew]

Thank you for the detailed answer!

It does seem counterintuitive that shiny walls are preferred over matte ones, and I'm curious why the colour of the shiny walls is important in this case?

[John E KVAM]

GWerth -

Simply put, your silo is too large for our sensor. The max you can get is 8M, but that assumes you have a full field of view, and a reflective target. With a dull target getting 8M is unrealistic, and especially if you are not using the full field of view.

This set up works to focus a projector, but they are generally pointed at a very reflective target.

I'm not saying you cannot get some data, but I'm not optomistic you can see all the way to the bottom.

Andrew-st -

If you have a mirror finish on your walls, the photons will hit the wall, and bounce toward your target. A matte finish means that the photons hit the wall, and scatter everywhere.

The best answer would probably be a smooth finish with black walls. Most of the photons would be absorbed by the black, and those that reflect would go the right direction.

But it would take some experiment to prove all that.

• john