Advice on VL53L7CX and very high reflective fabric

[S1M]

Hello everyone,

We recently tested the VL53L7CX and ran across an issue with very high reflective fabrics. Here are two images of such fabric. The first is a picture taken with flash light off, second picture is taken with flash light on. From the moment I wear this and come in front of the sensor I "disappear", the sensor also outputs status 255. Is there a way to handle this? Is there past experience with such items?

  

Here is an overview of our setup:

The VL53L7CX module from POLULU with a cover glass from GILISYMO is connected to a Raspberry Pi Pico 2040. Currently we're testing without the cover glass. We're using the SparkFun VL53L5CX Arduino Library

Thank you for your help.

[John E KVAM]

That fabric is called retro-reflective tape and it is VERY interesting. It returns all the photons that hit it exactly from where they came. It's why stop signs are so bright - but only to the car approaching them. 

That material causes all sorts of havoc. One issue is that the sensor has no trouble seeing the photons that are beyond the 4M max range. Between 4 and 8 meters the sensor will correctly give you an error status (Wrap around). But beyond 8 meters it might return your distance minus 8 meters. (If you are at 9 meters it might return 1M)

I'm going to guess you get the 255 error because you have saturated the sensor. The sensor is getting so many photons that it does not have enough power to reset the SPADs (called quenching.)

You could lower the integration time. This would keep the senor from saturating, but it will also lower the max distance at which you can range when you are not wearing that vest. 

it would be interesting to know how close you have to get before you start getting the 255 errors.

I've not seen an issue, but then I don't wear a safety vest in my lab. (And I have to admit to not doing a test with retro-reflective tape in quite a while.)

Tell me how close you have to get before you get the errors and I'll see if I can duplicate your issue. 

- john

[S1M]

Hello John,

Thank you for your message. I think it started around 2,5 to 3 meters but we will test it next week and measure the distance with a tape measure. We will also check how the sensor reacts on 10 meters. Once I have the results I'll come back to you.

[S1M]

Hello John,

We used the sensor in continuous mode because we thought this would give us better ambient immunity. We tested with autonomous mode and various integration times but the behavior is the same. Different values for the sharpener also does not change the results. As long as the sensor is on, we can see trigger 255 statuses when the sensor is exactly parallel to the reflective fabric. The maximum distance we used in our lab was 1.5 meters with a resolution of 8x8 and ranging frequency of 15 Hz. Is there anything else we can do? The max ranging distance required for our use case is 70 cm. Objects farther away is not important. The 90 degrees diagonal FOV of the VL53L7CX is important for our use case. 

Next time I'll change the setup so we can do a test on 10 meters.

[John E KVAM]

My thinking is that with that retro-reflective material, you are getting way too much signal. And for every photon we detect we have to reset the SPAD (Single Photon Avalanche Diode). That reset takes power, and one idea is that you are simply drawing too much current. 

Try changing from continuous to autonomous. Then drop the integration time. As the sensor is rated for 3.5M and you want to go only 70cm, you can go to the lowest integration time and still have plenty of signal. 

Do you need the 15 Hz. You can go lower. It will allow the power boost to recover and I think you would avoid the issue.

One thing at a time though. Drop the integration time. See if it fixes the one issue, then make sure you can see normal objects at 70cm. Then try dropping the frame rate if you have to.

Could you do me a favor? get the GUI going and display the signal along with the range. Then bring the retro-reflective material closer. What happens to the signal intensity just prior to you starting to get all the errors? Does it go through the roof? That way we know if my theory is correct.

 

- john