LimeSDR signal generation distortion

#1

Hi. I am trying to evaluate the LimeSDR as a signal generator.
I used the LimeSuiteGUI and used the TSG module to generate a simple CW (sine wave) at the TX1_1 port. I connected a scope to this output port to see the output waveform.
I notice that for low frequencies (35MHz -100MHz) the LimeSDR generates a very distorted signal (images attached). As frequency increases the distortion decreases (or maybe because BW limitation of the scope I am not beening able to to see the distortion at higher frequencies )

Is there any known reason that could cause this distortion?
Is it possible to improve the signal by adjusting any parameters on the LimeSDR?

Thanks
Mauro

#2

Hi Mauro,

Most SDR use a Switching Mixer. As a result in the output there are components around LO, 3xLO,5xLO,… To complicate things more, if you use the NCO to make an up conversion then for the components around 3xLO, 7xLO, 11xLo ,… a down conversion is made. Without external filtering it is not possible to generate a pure sinewave.

Regards Cornie

#3

Hello @mauro,

I did a similar experiment not a long time ago and got same results: Using LimeSDR as a signal generator As I understand there are two possible solutions. 1) Use another signal generator for frequencies < 100 Mhz 2) Use external filters.

#4

Hi Cornie,

Thanks for the answer!
As you said, by doing a FFT of the signal we could see that the harmonics are 3xLO, 5xLO,…
(we are using the DC mode of the TSG).

I have read the lms7002M datasheet and I don’t remember if the document specifies how the TXMIX is implemented. Are you sure it is a switching mixer (where did you read from)?

If the Lime uses a switching mixer there will always be harmonics in the output signal (no mater if the frequency is low 50MHz or high 1GHz), correct?

Regard Mauro

#5

Hi Afiskon,

Thanks for the answer.
I read your post “Using LimeSDR as signal generator” . Yes, the results looks similar.
I used a 200MHz scope to analyze the signal.
Have you check if for higher frequencies (500MHz, 1GHz 2GHz…) the distortion is reduced? What instrument did you use to measure it?

Regards Mauro

#6

@mauro I used Rigol DS1054Z, an “updated” (read - keygened/cracked) one with 100 MHz bandwidth. Everything above ~90 Mhz had no distortion. Though with this bandwidth I can’t be quite sure regarding harmonics.

As a side note it’s quite disappointing that LimeSDR claims to support frequencies starting from 100 kHz but in fact can’t transmit a simple CW signal on 80 meters band.

#7

That is not true… but may be true for 160 m band :wink: