The Manastash Ridge Radar
Frank D. Lind
August 5, 1998
The Manastash Ridge Radar system is a passive radar system that utilizes common FM radio broadcasts as its signal source. The system is designed to observe ionospheric turbulence that occurs in the same region as the aurora, or northern lights. This is a short status report to indicate our current state of progress for those people who have indicated an interest.
The Manastash Ridge Radar is currently online and operational! We have been operating the current equipment for about 10 months now. One receiver is deployed at the University of Washington while the other is deployed at Manastash Ridge Observatory. The system is stable, both mechanically and electronically, although we have made several modifications in order to fix some minor problems. The system design and theory has been described in Radio Sci. Bull. 284, 4-7, 1998.
Our current system consists of two direct conversion receivers, matched PC based streaming digitizers, and two GPS clocks for time synchronization. The Manastash Ridge receiver has a log periodic antenna with about 6 dB of directional gain, and we use a simple folded dipole for the UW receiver. The Manastash Ridge side receiver also has an additional baseband gain block. The system can be operated remotely using the internet with the software still in a somewhat primitive state.
Our current bottleneck is data processing and management. We are still limited by available computational power to small amounts of data. (It takes 30 minutes to process 10 seconds of data). Typically we will take and process between 100 and 500 seconds of data a week but we expect a significant improvement in this over the next few months.
We are currently working with a group in UW CSE to produce a VLSI correlator chip that will be capable of real time data processing. Currently an FPGA prototype of this system has been funded and is under construction. For postprocessing of the data we intend to construct a small Beowulf class cluster computer sometime in the next year. We are presently converting to the use of NetCDF for data management and this process will be complete by the end of the summer. We also intend to upgrade the control software to be more capable and to extend our automated data aquisition and processing capabilities. At the present a somewhat clumsy script interface is being used to provide system automation. Ultimately we intend to produce a graphical user interface for the system using the java programing language. Future versions of the radar will be made avaiable to the public and scientific community for use in research and education.
In the data that we have processed so far we have observed: a weak signal getting over the Cascade mountains, a strong scattered signal that is definitely from Mt. Rainier (a large local mountain), and lots of aircraft (1-3 per night). We have also demonstrated ~1 km range resolution and ~1.5 m/s velocity resolution on real targets. We expect to see auroral echoes once we have increased our data coverage and made some signal processing improvements and corrections. We are also planning an improved antenna system for the next generation of the receivers. We hope to have some output from the radar available on a daily basis by early fall.
This is an overview of the system.
The UW Receiver will be located here in the new Electrical Engineering building.
The Manastash Ridge Observatory, where the remote receiver is located.
Looking approximately NE from MRO is Eastern Washington and its farming communities.
Here is a block diagram of the direct conversion receiver.
This is the UW receiver itself and the GPS unit. The computer is to the right.
Here is the IQ signal from the UW receiver taken with 12 bit data. Note an FM signal is constant modulus and should be a circle. The width of the circle is caused largely by ground clutter.
Here is a spectral plot of the same 12 bit data for 1 complete second. You can see the FM signal quite clearly.
This is a spectral plot from MRO of data taken using only the sign bit of the digitizers. Notice the signals from adjacent stations leaking into the band that contains a station on the UW side. This data is simultaneous with the figure below.
Here is the FM signal that was broadcast from the UW side of the mountains. Again we use only the sign bit.
The log plot of the cross ambiguity for the previous two signals shows what we believe to be a signal scattered by Mt. Rainier. We have seen the signature in most the data we have processed.
The zero doppler shift cut from the cross ambiguity of the two received signals. The precursor peak may be the direct signal leaking over the mountains, or perhaps a different scatterer.
This is the log plot of the self ambiguity of the signal that was transmitted. Note the sidelobes that show up here also show up in the cross ambiguity. The central peak is missing here due to a plotting problem.
This is the same self ambiguity but on a linear scale. Note the wonderful sharpness of the FM self ambiguity.
This is a cross ambiguity showing multiple aircraft as observed by the radar in 10 seconds of data. The processing was done for maximum doppler resolution and the aircraft is travelling at -21.5 +/- 1.5 m/s. The clutter at zero doppler is produced by signal propagating over the Cascade mountains and by scatter from Mt. Rainier. We have verified the scatter from Mt. Rainier by varying the antenna pointing.