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.

EVM430-FR6047: ADC Capture Shifted Amplitude

Part Number: EVM430-FR6047

Greetings,

Today I set up the EVM430-FR6047. After a while I figured out the right parameters to get a good ADC Capture. But my waveforms have really high peaks. I think the problem is that my ADC Capture is shifted as shown below.
Does anybody have some suggestions how I can fix my waveforms ?



Thank you for helping :)

  • Hi Sascha,

    Can you provide screen shots of your configuration and advanced configurations in the GUI?

    Yes, there does seem to be something fishy here. Typically you want to capture a small portion the signal leading up the received pulse where it is mostly flat, as shown below.

    I also see the UPS / DPS (red/blue) traces are not precisely overlapping.  This would suggest the two transducers may not be matched well.  What is the manufacturer and did you get these as samples or did you work with the manufacturer to customer design them?

    Picture from slaa959 Ultrasonic water flow calibration

    For more information about setting the parameters correctly and what an ideal waveform should look like, check out the MSP Academy - Ultrasonic water meter design and testing.

  • Hi Dennis,

    thanks for your answer Slight smile

    These are my parameters:

    We are working with the Audiowell US0014-001 Transducers (1MHz). 

    Kind regards, Sascha

  • Hi Sascha,

    Ok, the one parameter that seems out of line is the GUI Based Gain Control.  Normally the gain is +/- a couple of dB so this tells me your signals are weak.  This could be due to either the distance between the transducers is too far, or they are not properly aligned.  By chance is this a "clamp on" transducer configuration where the transducers are mounted on the outside of the pipe rather in the pipe with direct contact with the water?  If not proprietary can you send photo of your setup?

  • USS troubleshooting guide.pdf

    Hi Sascha,

    Do you have an oscilloscope?  If so, see the attached simple (preliminary) troubleshooting guide.  In there it will show where to connect your oscilloscope probes and what your transmit and receive waveforms should look like.  If you can, please take a picture of the oscilloscope and share.

  • Hi Dennis,

    with this setup my results are not that good :/

    I used this pipe: https://www.geberit.de/produkte/rohrleitungssysteme-versorgung/geberit-mepla/

    Do you think it will ever work with this pipe ?

    I also tried it with copper pipes with the same setup and had really good results.
    Unfortunately we cannot say which material we are going to have in this application.

    Do you have any suggestions ?

    Kind regards

    Sascha

  • Hi Sascha,

    Actually, I'm quite impressed with your setup.  Nice 3D print job :) 

    I didn't realize your setup was a "clamp-on" configuration instead of "in-line".  I'm assuming the pipe you are using is plastic, correct?

    The clamp-on configuration has some challenges above and beyond an in-line configuration.  Primarily having enough energy to propagate through the wall of the pipe, and having the proper alignment between the two sensors.  

    Here at TI we did not investigate the clamp-on configuration too much.  My understanding is one of the original USS engineers did some experiments, but I'm not sure what he did with the results.  In this case I will need to reach out to him for guidance.

    In the mean time, what would be extremely helpful is a picture from your oscilloscope showing the transmit and receive signals.  This will show me the quality and amplitude of the signals as well as the timing.  Can you do this?  Here is example I took with my cell phone.

  • Hi Dennis,

    thanks Slight smile

    yes we need to use a clamp-on configuration. What do you mean by in.line ?
    I thought  I could fix the alignment problems with my 3D-printed parts.

    Initially I found this (https://www.ti.com/lit/an/slaa949/slaa949.pdf?ts=1653981127999&ref_url=https%253A%252F%252Fwww.ti.com%252Ftool%252FEVM430-FR6043) and wanted to rebuild it.

     

    I see the sending signal perfectly fine. But the answer is missing.

    Tomorrow I'll try to improve my setting one more time.
    May be I can show some better results.

  • I tested my setup with copper, PVC and aluminum. Somehow aluminum is the problem. Have someone ever had that problem before? 
    If yes, how have you solved it ?

  • Hi Sascha,

    Regarding the alignment, the 3D pieces look like they are separate, so how to you align them.  Do you move them around until you see the strongest signal strength?

    Regarding "in-line", this is the typical flow tube.

    Yes, any metallic material will be more difficult to propagate a wave through compared to plastic.  Aluminum may have a characteristic that does not allow it to work as well as copper.  PVC will work the best.

    I reached out to our engineer who did some experiments and he says the following:

    The customer's signal looks pretty weak. They may also consider using the 1MHz transducer designed for clamp on types(45 degree angled) as I have used in the app note below.  I had great results on PVC.  It may also be that the material of the pipe is attenuating the signal greatly.  I couldn’t quite tell what material that pipe is.  

    https://www.ti.com/lit/an/slaa949/slaa949.pdf

    Regarding the scope capture, your scope ch1 looks normal, but the on your scope ch2 it appears the received signal occurs to soon. Based on the distance between the sensors it will take several 10's of usec for the signal to propagate through the walls and fluid.  You should see something like what I show in the lower right hand corner.  In my capture, my scope ch4 (green) looks similar to yours, and you can see 100usec later the received signal on scope ch3 (violet).

    Make sure you have your scope connected to the correct locations as shown here in the schematic diagram and see if you can capture the signals again.

  • Hi Sascha,

    My comment about the alignment of the transducers has to due with the Snell's law, the thickness and composition of the pipe.  As you see in this diagram below, the two transducers form an angle of 45deg to the surface of the pipe, but the angle of propagation of the energy changes as it travels through the different material boundaries. I put the red arrows to show the original path of the wave so you can see the alignment that is required.  Check the internet for tons of information related to this.  Here is an interesting article from which the diagram below came from.

    As you read about Snell's law, note that our engineer used a 50% incident angle to reduce the amount of reflective wave energy, so you might not be detecting any signal because all the energy is being reflected off the surface of the pipe.

  • Hi Dennis,

    we decided to first focus on copper pipes. This way it is easier to get everything up and running. 

    With the copper pipe we have good results. Both with the oscilloscope and ADC Capture. Also the waveforms are looking better. 



    Now I am wondering how to calibrate the flow rate. Where can I say How big the pipe is ?

    This is the guide I followed. But I get a illegal meter constant.

    Thank you

    Sascha

  • Hi Sascha,

    As far as I know there isn't an easy way to directly input the cross sectional area of the pipe. You will have to go through the calibration procedure and get the meter constant. 

    Can you provide a screenshot when the GUI says "illegal meter constant". 

    Regards,

    Evan

  • Hi Sasha,

    What is the actual flow rate during this calibration?  Do you have a "reference" flow meter and if so, what flow rate is it showing?  Reason I ask is the screen shot of the Volume Flow Rate you provided shows 0 lph flow and has a couple of large disturbances.  Make sure when you perform this test the flow is constant (not changing rapidly).

    Refer to this other E2E posting.  Do the calculations and tell me what you calculate.  Then enter that value under the "Parameters" tab for the Meter Constant, then repeat your flow measurement for ~1 minute.  Then when measurement is complete and you have entered the actual flow rate in the dialog box, take a screen shot showing the new meter constant (or if it shows an error) as Evan suggested.

  • Hi,

    We never had a illegal meter constant again.
    I calibrated my setup with a reference and it works fine for now. 

    Thank you for all the help, very good support Slight smile

    If you want you can close this thread.

    Kind regards,

    Sascha 

  • Hi Sascha - glad to hear you got everything working.

  • Hi,

    I still have the problem and now I have to resolve it.

    Do you think TIDA-01486 can help me?

    Kind regards,

    Sascha

  • Hi Sascha,

    Ok, so back to the original issue where the received signal is offset, can you try again to take a picture of your oscilloscope showing the received signal as shown below?  I wasn't certain I see the correct receive signal in your previous scope capture.  There must be a delay between the transmit and receive signals.  If you are having trouble seeing or locating the receive signal relative to the transmit pulses, you can calculate approximately where (in time) the receive signal should appear based on the distance between your transducers and temperature of the water.  You can see the delay between the transmit pulses and received signal in my capture below.  My sensors are 16cm apart so there is ~106usec TOF between sensors as shown.

    Once you have a good capture of the receive signal, what is important is the DC bias voltage level the received pulse (violet in my picture) is at.  In my capture it is +800mV from ground, as expected.  See my earlier post -  I show where on the PCB to take these measurements.  You should see both the transmit and receive signals.  If you are having trouble seeing or locating the receive signal relative to the transmit pulses, you can calculate approximately where (in time) the receive signal should appear based on the distance between your transducers and temperature of the water.

  • Hi Dennis,

    We know where the signal should be (We tested it with PVC pipea and calculated it).
    As you can see we have the right offset (+800mV from ground)

    But if we work with metal pipes we dont have the receive signal.


    I think we need to amplify the signal somehow. Because with PVC pipes it is no problem. But for metal (copper, aluminum, steel) it is to week. Did you ever had this problem before.
    Can TI provide us an amplification board or do we have to build one ?
    I hope you can help us.
    Kind regards,
    Sascha 
  • Hi Sascha,

    I would recommend looking at https://www.ti.com/tool/TIDA-01486

    This reference design incorporates amplifiers on the TX and RX sides. Unfortunately this design is only available for reference, not for distribution. Let me know if you have any further questions!

    Regards,

    Evan

  • Hi Sascha,


    It's been a while since we have heard from you, so I'm going to assume you were able to move forward with your project.
    I will mark this posting as RESOLVED, but if this isn’t the case, please click the "This did NOT resolve my issue" button and reply to this thread with more information.
    If this thread is locked, please click the "Ask a related question" button, and in the new thread describe the current status of your issue and any additional details you may have to assist us in helping to solve your issues.

**Attention** This is a public forum