Posting of lms2fits

[Nathan]

I posted a new spectroscopic code, lms2fits, to SourceForge for capturing dynamic spectra.  It uses the LimeSDR-USB board, which is a dual-channel transceiver, having a USB3 chip that allows capture of up to 61 MSps of complex samples in two channels.  These channels are intended to be X and Y polarization, which are processed to the four Stokes parameters and streamed to three-axis FITS files.  The analog signal chain performs down conversion, allowing reception of frequencies up to 3.8 GHz.  So it is suited to reception of Jovian decametric emission below 40 MHz, and solar emission over a wide range of frequencies.  I have been running the receiver for several months at frequencies below 85 MHz, capturing the entire 60 MHz span for solar observing during the day, and from 14 to 38 MHz for Jovian emission during the night while Jupiter is currently high.

The file viewer, fitsview, is written in python, able to open both two-axis files of rx2fits, and the three-axis files of lms2fits.  As well as displaying the four Stokes parameters, it displays intensities of the six  basis vectors X, Y, A, B, RCP, and LCP, and some combinations of Stokes parameters.  It performs background subtraction, as well as dividing out model response functions for gain flattening.  It has a lot of features that take time to learn to use fluently.  The repository has documentation written in markdown.

There are caveats regarding this board.

One is that it is pricey. 

A second is that the computing power required to run at the highest sample rates is more than is typical in laptops and perhaps many or most desktops.  Arm processors cannot handle the load, and a processor with a working dual-channel memory is needed.  These requirements also make a system handling the widest bandwidths more pricey.

Another detail about the board is that X and Y sampling, while synchronized, are not simultaneous.  This occurs in part because the sample clocks of the X and Y channels are generated by separate synthesizers, and that offsets occur elsewhere in the digital path.  These offsets vary from startup to startup.  This means that a calibration noise source would be needed for absolute calibration of Stokes parameters, switched on at the start of operation or periodically during operation.  Without calibration, absolute polarization cannot be measured; but different polarizations can be distinguished, such as in this solar burst, and in these teepees. 

More data generated by lms2fits can be viewed at this link.  I am using an LWA antenna with the system located at a remote site for low EMI.

The receiver has not shown the sensitivity at HF and VHF frequencies of the SDRPlay and the rx888mk2 receivers that I have worked with.  Even with the front-end modification, it cannot see scintillation of Cas A and Cyg A, or the variation of the galactic background as the galactic plane crosses the sky.  It is best for seeing solar and Jovian emission.

Intermodulation has been a severe problem, in my case from the intense line sources at HF frequencies.  Front-end filters would surely help, as well as better adjustment of the input signal strength and gain setting.  Better characterization of the gain, noise figure, and IP3 as a function of settings would also help.

I would not suggest using this receiver for more narrow-band work, such as for HI surveys.  There are other instruments cheaper and better suited to such work.

Radio Jove's Spectrograph Users Group is a nice resource for this kind of dynamic observing.