I have a LimeSDR with recent (a month old) firmware, and SoapySDR and LimeSuite code from GIT master.
I’m using it with an interferometer application, and cannot see any useful two-channel coherence, despite having no problem with other types of hardware (USRP B210, for example).
I had heard that sensitivity on the LNAW port below 700MHz was poor, so I placed about 15dB of tighly-filtered gain (6Mhz passband) ahead of the LimeSDR to see if I could get useful results. Still, nothing.
I’m using the osmocom source for everything, and specify:
soapy,driver=lime,nchan=2
This appears to produce the expected start-up messages from the LimeSDR drivers, but I’m not seeing any useful coherence.
I use a simple conjugate-multiply between the two channels and low-pass filter that. Works like a charm with other technology, so I’m doubtful that it’s, per se, my application.
I remember that David Lonard published his results at 1420 MHz doing interferometry using 2 channels of a LimeSDR.
He used SoapySDR + python and CUDA wrappers to perform FFTs in a GPU.
I understand you are working at a different frequency but I guess you could probably compare the configurations and see if something surfaces.
There’s a thread here where he describes his setup and shares the python sources: