The received signal from sdrrx keeps unchanged even if I connect or disconnect the loopback cable to trainsmitter port TxA and receiver port RxA

Hi,

I am trying to reproduce this example: Receive Tone Signal Using Analog Devices AD9361/AD9364from:

https://ww2.mathworks.cn/help/supportpkg/xilinxzynqbasedradio/examples/receive-tone-signal-using-analog-devices-ad9361-ad9364.html

I have the boards: AD-FMCOMMS4-EBZ(which is AD9364) and Zedboard I only modified 1 line in the upper example: prmToneRx.SDRDeviceName = 'ZedBoard and FMCOMMS2/3/4'; alternatively: prmToneRx.SDRDeviceName = 'AD936x';

If I connect transmitter port TXA to receiver port RXA with a loopback cable, the received signal on figure is exactly the same as that in the upper page.

I used a physical Spectrum Analyzer to monitor transmitter port TxA, and the Spectrum Analyzer kept on changing if I change the center frequency of sdrtx.

The received signal from sdrrx keeps unchanged even if I connect or disconnect the loopback cable to transmitter port TxA and receiver port RxA

This is a weird issue. Do you have any suggestion to this issue?

Thanks

Answer 1

[Paul]

Hi Ardeal,

There is enough leakage between the unconnected RF connectors to lead to the results you are seeing on the example.

Since you also mentioned having replaced the SDRDeviceName property from 'AD936x' to 'ZedBoard and FMCOMMS2/3/4', I suppose you are using the R2018b release of the product. This is unrelated to your problem but you could give a look in the documentation for the " Compatibility with Previous Releases " page. Keeping the device name to 'AD936x' should work if you have the right SD card image loaded on the board using 'downloadImage'. Maybe that helps you in your next experimentation.

Paul

[Ardeal Liang]

Hi Paul,

I am using 2018a but not 2018b. I correct the comment about 'AD936x'. In fact, after double confirmation, I couldn't use 'AD936x'

according to your comments, the problem I am experiencing is unrelated with the version of 2018a and 2018b.

The SD card image I am using is burnt to SD card by Matlab command targetupdater. I believe the image in SD card is absolutely good.

I cannot understand the leakage is so big that the signal with/without loopback cable is the same.

Do you have any example to prove that the issue is caused by leakage of RF signal?

Best Regards,

Ardeal

[Paul]

If you are using R2018a then ignore my previous comment about the device name. You can access archived documentation for R2018a online: Receive Tone Signal Using Analog Devices AD9361/AD9364

Regarding the leakage, you can convince yourself this is happening by turning off the automatic gain control of the receiver.

From:

 'GainSource', 'AGC Fast Attack', ...

to:

 'GainSource', 'Manual', ...

This way, the gain will remain constant instead of letting the chip automatically adjust it. When you unplug the loopback cable you should see a clear difference in the plotted signals.

[Ardeal Liang]

Hi Paul,

Thanks for your quick and professional answer!

I made the 'Manual' modification, and then did the experiment you mentioned: experiment1: keep the system running, unplug the loopback cable. no difference could be seen on the figure.

experiment2: unplug the loopback cable, and then run the system. Obvious difference happens. This is figure when I unplug the loopback cable.

what I saw is the same as your experience or observation?

Best regards, Ardeal

[Ardeal Liang]

Hi Paul,

Your conclusion is proved to be correct. But I still have a question: The following figure is the signal when a loopback cable is connected between TXA and RXA ports.

during the upper figure is running, I unplug the loopback cable, but the signal didn't get smaller immediately.

If I re-run the script, the figure without loopback cable is correct and followed:

My question is:

If I unplug the loopback cable, the signal should become smaller immediately.

is my understanding correct?

Best Regards,

Ardeal

[Paul]

That's because the data you are seeing is not in real time. Unplugging the loopback cable has to happen when the device is acquiring the data, which happens after the first call to :

 [data, len, overflow] = sdrReceiver();

The example is set up to work in burst mode, meaning it will retrieve all your data at once for the amount of time required ('StopTime' in the example) and it will be displayed back to you frame by frame, slower than real time. Burst mode guarantees you have continuous data with no drop of frames.

If you give a look at the bottom of the scope while it is running, you will see the time that is currently being displayed. It will actually take longer than 10s to display 10s worth of data.

What you can do is set 'StopTime' to 10s and as soon as you run the example, plug / unplug the cable every 5 seconds for 20 - 30 seconds. That should be enough for the device to be configured to transmit the tone and receive 10s worth of data while you play with the cable. You will see your scope displaying changes in the signal after a while.

[Ardeal Liang]

Hi Paul,

I have just now tried my ideal/question on ADI IIO-Oscilloscope. I confirm the following result:

Set the receiver Gain to be Manual.

If I run the system with loopback cable plugged in, the received signal is almost the same as that in Matlab.

During the running, if I unplugged the loopback cable, the signal got smaller immediately.

This means that there is still something wrong

Best Regards,

Ardeal

[Ardeal Liang]

Hi Paul,

Thanks for your answer!

Change the mode to non-BurstMode solve the problem.

Thank again!

Ardeal

[Paul]

Great! Happy to know this helped you.