AWR1243: AWR devices for level sensing

MIKHAIL SALOV

Part Number: AWR1243

Good day Colleagues,

Could you tell me if it's possible to measure a liquid level in a metal tank? Or not because of multiple reflections?

If yes, could you explain me a few questions:

1. What is a minimum measurable distance for AWR family? I found only information about max distance (up to 160 meters).

2. Do you have any proposals for a transparent plastic for a such high frequency?

3. How a dirt or fog or rain affect? Because of a really high frequency it does, but how much?

4. What if the dirt will on a plastic which covers antennas? How this will affect?

Thank you, Mikhail

over 7 years ago

0 Francois Charlot over 7 years ago

TI__Mastermind 28287 points

Hello Mikhail,

To answer your questions:

- A metal tank will indeed cause signal reflections. It could be necessary to use a horn lens antenna or a dielectric lens to reduce the field of view of the antennas so to avoid these reflections.

- The IWR family of mmWave sensors for industrial applications has a bandwidth of 4 GHz in the 77–81 GHz band. This translates to a minimum sensing distance of 3.75 cm.

- Most materials with a low relative permittivity (e.g., in the range of 2 to 4) will transmit most of the radar energy. There is a wide choice there, like polyethylene, PTFE, PVC... Choosing among them could be related not to the radar performance, but to mechanical stability (temperature, humidity...), manufacturability, etc. The material thickness also has some importance as a greater thickness will cause more attenuation.

- Relative humidity (RH) has a negligible impact on the mmwave transmission: 0.02% between 0 and 100% RH. You may refer to this paper (link), table 3.

- For rain, you may refer to table 10 in this study from the FCC (link). As an example, a quite heavy rain of 25 mm/hour translates to only 1 dB of attenuation for 100 meters of mmwave signal path.

- Dirt affects the propagation of mmwave, but it is impossible to give a figure as dirt could be anything: dust would cause little attenuation, very humid mud would cause a lot because of the water it contains... So one needs to experiment there.

Thanks,
François.

0 MIKHAIL SALOV over 7 years ago in reply to Francois Charlot

TI__Expert 6950 points

Good day Francois,

Thank you very much! Glad to read from you!
RF is cool but requires a lot of precise manufacturing with expensive material.

Just to double check, the minimum sensing distance if 3.75cm means the step also will be 3.75cm? I.e. between 3.75 and 7.5cm will be a blind zone? correct?

Thank you, Mike

0 Francois Charlot over 7 years ago in reply to MIKHAIL SALOV

TI__Mastermind 28287 points

Hi Mikhail,

Indeed RF requires precise manufacturing and some more expensive material than FR4; but I would say the cost adder for such PCB antennas is reasonable, and one gets a very compact solution.

To give a bit of background on the range FFT processing: the radar signal goes into an FFT that will divide the radar range (e.g., 0 to 50 meters) into a number of "range bins" that correspond to the number of points in the FFT. So a 512-point FFT would divide the range bins of about 50/512 ≈ 10 cm. Within one bin, the radar cannot separate 2 objects. One may increase the FFT length, but there is a fundamental limit which is given by resolution = c / (2 x bandwidth). 3.75 cm corresponds to a 4 GHz bandwidth. One may oversample the FFT but will be limited by this fundamental limit anyway.

So the radar could not tell if there are 1 or more objects in the same range bin, unless it could discriminate them through other means, such as velocity (e.g., 2 objects within the same range bin but one does not move and the other moves) or by tracking over time.

Assuming only one object is within a range bin, it is possible to use the phase information of the signal to increase the accuracy of the distance measurement. For example, we have demonstrated an accuracy in the range of 10 μm at a distance of about 2 meters. There is a plan to release this month some example code for such high accuracy measurements, either through the TI Resource Explorer (link) or a new TI Design. Keep monitoring it!

By the way, TI have just released some mmWave experiments, one of them being to show how objects may be detected through different materials. This experiment is accessible here.

Thanks,
François.

0 MIKHAIL SALOV over 7 years ago in reply to Francois Charlot

TI__Expert 6950 points

Thank you, Francois, for a such detailed answer! And link above.

Radars are very familiar to me. Accuracy of 10um is a very great.
So 10um means the step of the distance detection is 10um or radar finds an object with 10um accuracy (i.e. 1meter +/-10um)?
I am asking because in a tank the level can change just in a few mm. It's important to see this level difference.

Thank you, Mikhail

0 Francois Charlot over 7 years ago in reply to MIKHAIL SALOV

TI__Mastermind 28287 points

Hi Mikhail,

Yes, you could detect level changes in the order of 10 μm at distances of ~1-2 meters. Again, stay tuned for the release of this new software I mentioned above.

Thanks,
François.

0 MIKHAIL SALOV over 7 years ago in reply to Francois Charlot

TI__Expert 6950 points

Wow, thank you, Francois,

Mike

0 Parisa Esmaili over 7 years ago in reply to Francois Charlot

Prodigy 10 points

Dear François,

Could you please let me know is it possible to detect level changes in order of μm for tank in range of 10 cm up to 50 cm (small size)?

0 Akash Gondalia over 7 years ago in reply to Parisa Esmaili

TI__Guru 50240 points

Hello Parisa,

We can confirm sub-mm accuracy although the full extent of μm has yet to be determined. I'd recommend you look at the High Accuracy Lab on the TI Resource Explorer here: dev.ti.com/.../

If you have questions on this matter then feel free to open up a new thread!

Cheers,
Akash

Sensors

Sensors forum

AWR1243: AWR devices for level sensing