Why my demodulated pulse train look like superimposed on a low frequency sinusoid?

I generated a pulse train from USRP device and transmited it at 3.5GHz the TX was coupled through a cable with Lecroy SDA 18000 device. I captured the whole signal down converted the signal but the recieved pulse train is looking like a alow frequency sinusoid where the pulse train is superimposed on it. How to solve this issue?

[Mathieu NOE]

hello

maybe a main supply (50 or 60 Hz) interference (EMI) - check grounding of your equipment

or impedance mismatch (do you use the right cable ?)

if any hardware solution solves the problem , we can try to solve it by software , but then we would need your data

[Muhammad]

Hi,

Thank you for the kind response.

Apparently the setup is OK. I have tested the cable as well. I have attached the data file.

Best Regards,

[Mathieu NOE]

I don't have the same plot as you

your mat file contains one variable : sig_amp

[Muhammad]

Yes, sig_amp is the whole captured signal. while this is the downconverted signal

[Mathieu NOE]

seems sig_amp signal is not very good - or ? if I zoom inside the signal I see mostly quantization effects - why is the signal amplitude so low ? I suspect the "bad" demodulation could come from a bad input (modulated) signal ?

these are the first 2000 samples of sig_amp

[Muhammad]

Thanks for pointing out towards this aspect. I attenuated the input signal before feeding to the oscilloscope. This could be the reason. I will try without attenuation.

Thanks a lot.