VL53L1X-SATEL - Low signal and unreliable readings - any ideas?

[JPark.28]

Hi all

I've built an oil tank level measurement system following https://www.instructables.com/WiFi-Oil-Tank-Monitor/

The system works very well on the bench, but when installed on my oil tank I'm seeing very low signal levels, especially at low ambient temperatures.

For example:

"rangeStatus":"range valid","peak_signal_count_rate_MCPS":"1.57","ambient_count_rate_MCPS":"0.13",

"rangeStatus":"range valid","peak_signal_count_rate_MCPS":"1.51","ambient_count_rate_MCPS":"0.13"

"rangeStatus":"signal fail","peak_signal_count_rate_MCPS":"1.49","ambient_count_rate_MCPS":"0.13"

"rangeStatus":"signal fail","peak_signal_count_rate_MCPS":"1.38","ambient_count_rate_MCPS":"0.39"

At the moment I have the sensor behind an IR window that came with another VL53L1X sensor.

Has anyone else had success using these TOF sensors with kerosene?

Do you have the sensor directly exposed to the oil vapour?

Many thanks

Kind regards

John

[John E KVAM]

you didn't say what your distances are, but I'm guessing it works close up, but you have issues as the tank level drops.

Sometimes increasing the TimingBudget helps. More time to get more photons. But it's possible the distance and reflectivity are too much for this sensor.

I would try using the VL53l1CB. This sensor is identical in shape, so no layout changes are required. But it processes the data differently. It uploads the raw data to your MCU and with your processing power we dig out a better signal. Doesn't come for free though. The code is 9K larger and uses 80K cycles to process the data.

As a test, you can try using the L1CB software on your L1X (download it from ST.com) You can try it. But please don't go into production like this. The L1X is not guaranteed to run the L1CB for very long, but for a short test it should be OK.

[JPark.28]

Hi John

The distances are 0 - 2m.

Timing budget is set to 500ms and Distance Mode is set to long.

Are there other settings I should try adjusting to try to deal with the reflectance of the oil surface?

Do you have any data on using these sensors with kerosene - I note that it is mentioned in some of the applications data but there are no examples that I've been able to find.

Many thanks

Kind regards

John

[JPark.28]

Also I was not aware of the L1CB version - I'm taking a look at that too.

[John E KVAM]

I've done fuel oil and water. Fuel oil is dark, but reflective. It's also really smelly.

The customer was filling a tank and taking measurements. I suggested using a tripod and changing the sensor location rather than the fuel depth. But do in somewhere there is very little sunlight. And use LED lighting to simulate the internals of a tank.

We actually settled on some reflective paper that had the same characteristics as the oil.

Way less smelly.

At 2M the diameter of the circle of illumination is 1M. If you tank is not that wide, you will get photons hitting the wall and bouncing all over.

There is a web site.

Ray tracing simulator: https://ricktu288.github.io/ray-optics/simulator/

The L1X and the L1CB have the ability to narrow the FoV to give a better figure. But it comes with a loss of some signal. But the L1CB should get you a lot more signal to start with.

So I'd go with the the L1CB and narrow the field of view.

[JPark.28]

Very interesting!

The sensor is quite close to the wall of the tank so I guess I will be getting loads of reflections from that. I'll try changing the location first, and maybe reduce the FOV.

Do you think the dark plastic "IR" window is worth removing?

Many thanks

Kind regards

John