• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 1183 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IWR1443BOOST: mmWave fluid level transmitter IWR1443

user5769462
Prodigy 50 points
Part Number: IWR1443BOOST


Hi,

I am doing one of the project in the industrial toolbox folder, path: ~\ti\mmwave_industrial_toolbox_2_5_1\labs\lab0004-high-accuracy-14xx, the mmWave fluid level transmitter. I followed the procedure and could measure liquid levels with decent precision, but I do not get what the graph means in the High Accuracy Visualizer website, https://dev.ti.com/gallery/view/4768107/mmWave_Fluid_Level_Transmitter_1_0_0/ver/1.0.1/, what are the x and y axis? It says the x-axis is Range (metres), but range of what? Is it the horizontal range the radar can measure? Or the distance from the board to the objects the reflects the em wave? I found that if i turn on the RangeLimitCfg in profile_2de.txt, and set some maximum, say 6m, then the distance the radar can measure can never go beyond 6m. But if I turn this option off, I can measure surface 20m+, and get peaks from 3.5m to 7m in the graph that I don't understand. So what is going on? 

Bang, Regards.

  • 0
    TI__Guru 50240 points
    Hello,

    This plot is known as the Range Profile. It is a relative measurement of received signal strength at a particular distance away from the sensor. The y-axis is unitless and is only meant to show relative signal strength. The x-axis is known as range (meters) and is a measurement of the distance from the board to a particular object.

    The Range Limit parameters will limit the Range that is highlighted by the High Accuracy Visualizer. To cut the range off at 6m means ignoring peaks in the range profile that may appear beyond that 6m.

    Let me know if you have more questions!


    Cheers,
    Akash
  • 0 in reply to Akash Gondalia
    Prodigy 50 points
    Dear Akash,

    Thanks for clearing things up for me, there is no confusion about the graph anymore. However, there is one thing that I found abnormal: when do a measurement, i.e. the distance from the board to some ojects or some liquids, we get the result "Fluid is Xmm away from sensor". This result X is always somehow always a little bit smaller than the x-coordinate of the highest peak on the graph, not by much, but noticable. And this margin tends to get bigger when measuring further away objects. I wonder why it's the case.

    Regards,
    Bang
  • 0 in reply to user5769462
    TI__Guru 50240 points
    Hi Bang,

    There could be some calibration issues going on here. Would you be able to post some more information about this discrepancy? Is it a matter of millimeters or centimeters? How much larger does this discrepancy become when distance increases?

    A table of this information would be very helpful.


    Cheers,
    Akash
  • 0 in reply to Akash Gondalia
    Prodigy 50 points

    Dear Akash,

    I did more tests, and find out some new properties that are worth mentioning. The reading of the radar, i.e. fluid is X mm away from the board, could even be varied with both the board and the object being measured fixed in their positions. For example I first measure the distance with just the radar perpendicular to the object. Then I restrict the the radar’s angle of receiving and transmitting em wave, by putting a rectangular or cylindrical carton tube (wrapped with tin foil) perpendicular to the antennae, so that the signal can only go through the tube. When i make a measurement set up like this, I get a different reading, sometimes more accurate than the raw radar reading, sometimes not. But regardless of the set up, as long as the radar and the object are fixed in their position, the value of the peak is always the same, only the readings can be vary. This picture below shows the difference in readings and noise, the graph on the right is the one with the tube. 

    the real distance being 901mm. Now its the peak value comparison:

    The readings are exactly the same. Here is one more example:

    As you asked, I made a table of three measurements with the tube:

    If there is more information needed, please ask, I can do more experiments.

    Thank you very much and regards, Bang.

  • 0 in reply to user5769462
    TI__Mastermind 37500 points
    Hello,
    In your example plots, using the shield tube results in a stronger reflected signal, and less other targets detected.

    The distance calculated is based on the radial distance. You may find that the specific use of Tx affects your measurement. Tx3 is closest to the
    Rx, however it has higher leakage to Rx4. If you are using Tx1, and the 4Rx, the measurement is based on Tx1 to the center of Rx2,Rx3. It is a radial distance, unless you have a redesigned card with one Tx and one Rx with a shield in the middle.

    Normally with level gauge there is a linear calibration after the range FFT to process your mm value. There is a (m) slope and intercept (b), if your measurements are repeatable, then the m and b can be used to adjust the measured value.

    There are also optimizations for specific chirp parameters, numadc samples, higher DFE output rate, and number of chirps/frame processed to have a higher integration time, which allows the detection to occur at a longer distance.

    Please capture the raw Rx data set, the non coherently combined FFT data set, and the output from a working and less-well working value, and send. In the code example you would capture the 1D FFT values for range from L3 memory, then the non coherent combined FFT. I will work with Akash on this next Mon / Tues.

    Regards,
    Joe Quintal