Divider opacity to prevent crosstalk with dirty cover glass

[AWill.17]

I'm working on a design for a PCBA with enclosure, using the VL53L0X. There's a chance the cover glass will get dirty, so we took some tips from this youtube video regarding the use of an opaque divider: https://www.youtube.com/watch?v=BHM2CZVXZGc

Ideally, we'd like to be able to build the divider into our enclosure instead of ordering an additional component. We'd also like to mold the enclosure from one type of plastic (no double-shot molding, to keep costs down). We can achieve clear cover "glass" no problem, and then texture the mold to make the rest of the enclosure semi-opaque to the human eye. My question is whether we can use texture in a similar way on the divider between emitter and receiver. Is that opaque enough to function like the divider in the video? We wouldn't expect photons to pass through unimpeded on a straight path, but I'd call it closer to translucent than fully opaque.

Is this a scenario you've encountered before? Or can you dive a little more into the sensor's theory of operation to take an educated guess at how well this might work?

Thanks!

[John E KVAM]

It is my experience that if light gets into plastic, it will bounce all over the place internally until it goes exactly where you don't want it.

Try this experiment. Take a big sheet of plastic some quarter-inch thick. In a dark room point a visible laser perpendicular to the place. Look at the edges of the plastic. No light will show. Then place your hand on the far side of the plastic, simulating dirt.

Notice that every edge of the plastic lights up.

It is my conclusion that the photons will hit the plastic at odd angles and start bouncing rather than leaving. So it's my guess that it wont work the way you want it to.

But try it. That's what the EVK is for.

Take your plastic and cover it with a couple of tiny bits of tape. And then sand blast it.

Then do it again on the other side.

That should be a close approximation of your design.

Put the plastic over your sensor in the EVK and run the crosstalk calibration.

If you are right, the number will be very close to zero.

Unfortunately I'm betting against you.

If your crosstalk number looks good, then take some 'dust' and sprinkle it on. (I used coffee creamer.)

You can re-run the crosstalk calibration, or you can try ranging on your hand. The distance you can range will be way down. But the distance should be accurate.

Good luck

• john

[AWill.17]

Thanks John! We'll give it a try. We were pretty skeptical ourselves but I figured it's worth getting an informed 2nd opinion and running some tests.

Fun fact, I used baby formula for prior tests. :)