Very high spurious emission during the LimeSDR mini calibration

Projects LibreCellular
#1

Hi everybody,

I have tested the LimeSDR mini (the old and new version) in the LTE eNB from srsRAN and openairinterface5G. In figure 1, there is the spectrum when running the LTE eNB from User Guide — LibreCellular 21.04 documentation srsRAN after the LimeSDR calibration is finished. As shown, the spectrum seems OK after the calibration is finished. However, during the calibration, a lot of spurious appear in the spectrum as shown in figure 2 (span of 3 GHz) and figure 3 (span of 54 MHz). In those tests, I used tx_gain = 53 and an attenuator of 30 dB at the output of the LimeSDR mini. So, the power of those spurious is high, around 6-8 dBm. The results are similar when I use soapy or eNB from openairinterface5G, with LimeSDR v.1 or LimeSDR v.2. I am concerned that these spurious can damage the power amplifiers like ZHL-1000-3W+ from minicircuits (https://www.minicircuits.com/WebStore/dashboard.html?model=ZHL-1000-3WX%2B) whose absolute maximum rating input RF power is 0 dBm. So, I would be very grateful if someone can give me a direction how to minimize those spurious during the LimeSDR mini calibration process. Best regards, Sindi.

Figure 1-Spectrum after calibarion is finished
Figure 1-Spectrum after calibarion is finished1920×1440 224 KB

Figure 2-Spectrum with 3GHz span during the calibration
Figure 2-Spectrum with 3GHz span during the calibration1920×1440 219 KB

Figure 3-Spectrum with 54.78 MHz span during the calibration
Figure 3-Spectrum with 54.78 MHz span during the calibration1920×1440 234 KB

#2

I have found with most of the SDR transmitters. When you are driving it with floating point data - if the input voltage to the SDR transmitter is even slightly greater than ± 1.0 volts they generate spurious emissions. I am not sure if that applies during calibration.

#3

I’d like to know the answer to this as well. OP, have you considered a limiter?

#6

Normally you’d have some kind of software control of the PA so you’d only enable it once the calibration procedure is done. But if the calibration occurs outside of normal operations then I’d use a limiter.

#7

OK. Thank you.

#8

Hi,
I think not, because the power level of those spurious does not change if I change the tx_gain. Even with a low tx_gain such as 40, the power level of the spurious is aproximately the same as with tx_gain = 53. It decreases in the operation frequency only…
Best regards, Sindi.

#9

Hi,
I had not considered a limiter, but I will consider it now. Thank you.

#10

Also be careful of providing a signal input to amplifiers which are not powered up, as this has the potential to damage them. Ideally you would switch the SDR output between a termination and amplifier input.

#11

Hi Andrew,

Ok, thank you very much for the tips. Best regards, Sindi,