issue during VL53L1X crosstalk calibration

[DSai.1]

Dear Sir,

We are currently in the process of conducting crosstalk calibration using the VL53L1X ultra-lite driver, but unfortunately, the results we're obtaining are not as expected. Our approach involves utilizing an A4-sized gray chart board as the designated target.

Presently, the xcd reading registers at approximately 840mm, with the chart positioned at a distance of 1000mm from the VL53L1X sensor. The corresponding signal strength is indicated as 11064. After implementing the calibration procedure through the VL53L1X_calibrateXtalk() function, we retrieve an xtalk value of 0x99C6. This value is subsequently applied in our crosstalk compensation calculations for ranging.

Our observation involves ranging a stationary target. Curiously, despite maintaining the target's immobility, the resulting range status reads as 7. As an experiment, we moved the chart closer to the VL53L1X by 100mm. This adjustment led to a range status of 0 and a corresponding range distance of 2366mm. Notably, this measured distance significantly surpasses the values recorded prior to the calibration process. We would greatly appreciate any insights or guidance you could provide concerning this perplexing issue.

In our procedure, we have not conducted offset calibration prior to engaging in crosstalk calibration. The rationale behind this decision is rooted in the fact that our application does not necessitate an extremely high level of accuracy.

Additionally, we are keen to ascertain the optimal chart size for conducting crosstalk calibration. Would it be advisable to opt for a larger or smaller chart size in this context?

Thank you in advance for your assistance.

Best regards,

Dai Sai

[Zhiyuan.Han]

Hi DaiSa

From your description, seems you are facing xtalk overcompensation issue. 

• For xtalk calibration, A4 paper size is not big enough to cover all the FOV. you need at least 500*500 chart size to cover all the FOV.
• From the return signal rate, seems your chart reflectance is very high. can you try a bit lower reflectance chart. 
• For xtalk calibration, from you data log, I did a manual calculation, the xtalk level should be around 4500, but you got 0x99C6=39366, which is much higher than expect value. 
• Below is the test code I am testing. please have a try.

uint16_t xtalk;

status = VL53L1X_CalibrateXtalk(dev, 1000, &xtalk); 
printf("xtalk= %u\n", xtalk);

Br

Zhiyuan.Han

 

 

 

 

[DSai.1]

Dear Mr. Han,

thank you very much for your helpful information.

I will try to make calibration based on your guide for a better performance.

best regards

Dai Sai.

 

[John E KVAM]

You almost got there with your question on the optimal chart size. Assume the diameter of the circle at some distance D is 1/2D. So at 1M the area of the circle is 500mm. At that distance you are going to need 4 A-size sheets. But at least you know you have a pretty good cover glass.

In your calibration - and your testing you are picking up what is behind the target. And that's your issue.

One thing to do is try a much darker target. Then you will under-range at a much shorter distance.

I'd invest in a can of flat (matte finish) black spray paint and paint a bit of cardboard with it. Use several coats and do a good job of coverage. Then use that target. Repeat your ranging test, and decide where to put the target. It will be a lot closer than your 1M distance. That means you can use a smaller target and you will make your manufacturing guys happier. If black does not work for you, see if you can find a very dark gray matte finish can of paint and try again. Each reflectivity will have it's own distance, but once you find a paint and a distance you like you are set. 

[DSai.1]

Dear John E KVAM

thank you very much for your advise about selecting cardboard.

We will prepare the cardboard according to your information.

 Best regards

Dai Sai