LimeSDR EMI compatibility issues and radiated interference

General Discussion
#1

Hi folks,
We are designing a GNSS receiver using the LimeSDR as the front-end and we faced a serious problem with the radiated interferences coming from the LimeSDR PCB.
The LimeSDR PCB is radiating a lot of electromagnetic noise specially between 1 GHz to 2 GHz. This interference prevents the placement of GNSS antennas nearby the LimeSDR PCB, and thus, they can not be integrated in a small device.

Please check the three videos I have recorded showing the problem:

https://youtu.be/2h7TVVrfxLs

https://youtu.be/pIsLOmMKQ2Y

https://youtu.be/JGieg1Mjds8

It is curious that the LimeSDR is not actually being used by the host computer, the interference appears just plugging the USB. I have tested it with two different LimeSDR boards just to discard a defective component, but I obtained the same result.

Any ideas of how to minimize the radiated interference? As far as I know there is no official or unofficial metal cage to solder in the LimeSDR board…
Any ideas which is the signal / component that is radiating such amount of RF noise??

Thank you and best regards,

Javier.

Dr. Javier Arribas Lázaro
Senior Researcher - Communications Systems Division
Statistical Inference for Communications and Positioning
Centre Tecnològic de Telecomunicacions de Catalunya (CTTC)
Parc Mediterrani de la Tecnologia (PMT)
Av. Carl Friedrich Gauss, 7
Box 08860 - Castelldefels - Barcelona
Voice: +34 93 645 29 00
Fax: +34 93.645.29.01
E-mail: javier.arribas@cttc.es
http://www.cttc.es/people/jarribas/

#2

@Zack, do you have any suggestions?

#3

Hi @jarribas,

This suggests that there is no configuration loaded. Hence PLLs will be un-locked, RF outputs are on.

#4

Thank you Zack for the suggestion. I already checked this condition. The radiated interference does not change at all if I configure the LimeSDR. It also does not change even if I start the sample stream with GNURadio companion. I can record a video if required.

My thought is that it should come from a clock signal either from the FX3 chip or from the FPGA. What is curious is that it is apparently not a narrow band interference coming from , let say, harmonics from the 30.72 MHz clock… it appear as a wideband noise.

I measured it with a hackRF (shown in the video) and also with a shielded RTLSDR (nooelec one with metallic case and 0.5 ppm TCXO) with similar results = wideband interference present at least in between 1 GHz to 2 GHz