Simple sinewave strangeness

I have noticed that I get some strange tones when I do a simple sinewave generator flowgraph using the TX/RX chain with a cable connected directly between TX1_1 and RX1_H on my LimeSDR. At the receiver end I get almost a double tone with some strange garble. I have done the same flowgraph with a USRP1 with the Basic RX/TX and get a much cleaner tone. I get similar results when I use 220Mhz on the LimeSDR, but I have used 2.245Ghz in the graph because the sound artifacts are more pronounced.

I would like to know what is causing this as I believe it is related to some other problems I have been attempting to troubleshoot.

LimeSDR Graph

LimeSDR Tone

USRP Graph


Is it produced in the tx chain or rx?

The tones I linked were when I added a wave file sink in parallel to the Audio Sink.

You make an extreme violation of the sample theorem here. Since the LimeSDR has no brickwall filters I would increase the LimeSDR sample rate to 3-4 times the BW and use interpolation - decimation blocks to adapt the 48k to the LimeSDR sample rate. You may have to reduce the BW to avoid CPU overload.

I have found 5M on the TX and 1.5M on the RX is the lowest I can go without getting some sort of error like

RuntimeError: RxLPF frequency out of range, available range from 1.4e+06 to 1.3e+08 MHz

Any clue on how I can significantly reduce the BW?

From the “LMS7002M DataSheet” -
“RX analog filters are tunable from 0.7 MHz to 80 MHz.”
“The TX filtering chain pass band is tunable from 2 MHz to 80 MHz”

So 48kHz samplerates in your flowgraphs is not good.

Those error messages are coming from here at a guess, and the limits look to be slightly different than those defined in the LMS7002M datasheet:

Another thing you may want to look into if the noise is still present using sensible sample rates is to try a USB cable with better shielding, noise could be coming from USB 3.0. Or from the PC, try moving the LimeSDR further away.

The LMS7002M has internal interpolation - decimation filters up to a factor of 32. Don’t know if you can use them with the osmocom source - sink blocks. Fact is that it would be very usefull to have additional interpolation filtering in the FPGA. For low HF frequencies we are forced to use filter TX filter BW of 30 MHz. For those BW there are only second order lowpass filters available. So for avoiding aliasing sample rates of at least 100 -200 msps would be needed. Even with the internal interpolation filters the USB and the PC must provide sample rates that are significantly higher than the normal 48 or 192 ksps TX sample rates of hamradio SDR programs. That is probably the reason that it takes so long the get some TX SDR programs adapted for the LimeSDR.