Vl53L1X Calibration Issues

[Brettjohnson7191]

We have two VL53L1X sensors side by side. Both have gilizmo cover glass and are placed 140mm from a 8.5x11 sheet of mat grey paper. Running the calibrate offset function returns 2mm for sensor one and 6mm for sensor two. Following this the Xtalk calibration function is called this returns 37333cps for sensor1 and 52562cps for sensor2. 

Measuring distance after calibration shows the sensors are now way off and differ 2-3 inches from each other. 

 

My question is what could be going on to get such high values for Xtalk when we have ideal cover glass and placement?

Do we want the SPAD ROI to be 16x16 when we run the calibration? currently we have it at 4x4, 200ms of timing budget, and 200ms of intermeasurement period. The sensors are in a dark room away from an external sources of IR light. 

 

[John E KVAM]

You might have missed a step. After doing the offset cal, remove the white paper and replace it with something dark gray. Then find that point where the senor under-ranges by 10 or 20 percent. This is your crosstalk cal distance. 

 

Then do the crosstalk cal at this distance, giving it your exact distance. By knowing the real distance, the measured distance and the number of photons, it works out the crosstalk.

Note: The diameter of the target must be at least 1/2 the distance you are ranging. Otherwise, the sensor will 'see' photons from what is behind your target, and we don't want that.

- john

[Brettjohnson7191]

two follow-up questions

You mentioned using a white paper for offset cal. Shouldn't we use a dark grey target for this? U2510 recommends a 17% grey target. 

What approximate range would we expect the actual distance to begin deviating? I did run this test from 5" to 30" taking new measurements every 0.5". The resulting curve was linear but had offset issues. This test was run with an 8.5"x11 sheet of dark grey paper though so there may have been some bias from not meeting your size criteria. 

[John E KVAM]

Complex question. The offset affects the crosstalk, and the crosstalk affects the offset. 

Solution is to use a bright, near target for offset. With 20M photons from your target and 200K from the coverglass, you will not see the crosstalk effect. 

But for offset you really want a dull target where the 200K photons approach the 500K from your target. This will lead to a large difference between the real distance and the measured one, and we can derive the crosstalk from the % short and the total photons.

If you error is linear, you have an offset issue. If it starts out small and gets progressively larger, you have a crosstalk issue. If you cannot find a place that under ranges by 10% or more, you don't have a crosstalk issue.

- john