Forth-Based Radio Direction Finder Andy Korsak KR6DD kr6dd@yahoo.com A Gforth program running in a Raspberry Pi 2B (donated by K9WS) will be demonstrated for capturing bearings from a Ramsey Electronics model DDF-1 Doppler radio direction finder (RDF) receiving audio from a ham radio receiver. The Gforth SW saves bearings data to a thumb drive in two formats: (a) date and time stamped raw data (8-15 nibble value defining which of 16 LED's the unit turns on 500 times/sec), and (b) date and time stamped bearings along with quality factors and bearing duration selected by a smoothing and qualifying process to pick out from noise the direction to sources of very short signal bursts and longer but strong transmissions. Reporting of raw data code via TCP/IP for remote processing needs completion and it would require access to a participating radio ham's local network, which may not be permitted in many cases. Reporting the processed bearings data at a much slower rate may be over AX.25 packet radio or cell/text/email messaging upon request, for which the SW needs completion. Since both raw and processed data are being flushed out to two thumb drive files, that data could be concurrently read "on the fly" via concurrent Raspberry Pi code written in e.g. Python or other languages. As a last resort, recorded bearings data at various RDF monitoring stations can be shared by physical copying of thumb drives and then triangulation can be accomplished at least the hard way "post mortem". A parallel project undertaken by two of my RDF team associates is developing low cost broad band TDOA receivers that could form a large network of time-of-arrival recording stations dumping data to flash memories that could be regularly swapped out and bearings would be triangulated therefrom. If enough recording stations could be sprinkled around the area of interest, they wouldn't have to be at high locations. Future plans: Preferably we can come up with a real time monitoring network after we deploy a number of Doppler DF remotely reporting stations at high elevations scattered around the greater SF Bay Area. Such a network was proposed by Bob Simmons WB6EYV in southern California who was the originator of the PicoDopp DF units. He apparently sold the rights to his design to a seller of the DDF2020T product for which there is available a remote data collection and triangulation SW package. Several other similar proposals and implementations have been made. We are also contemplating a more up to date HW solution to replace Doppler DF by fast HW RDF using real time matrix processing of digitized signal data such as was available only using very expensive military grade hardware a couple of decades ago in the FIRESTORM project of Dr Steve Stearns K6OIK. In my presentation I will show how Steve described FIRESTORM in a recent email. Recently a related hardware and software implementation like that has appeared, the Kerberos RTL board, but it still pales compared to Steve's system -- it's very slow because it has four separate synchronized local oscillator receivers but it lacks fast real time matrix processing HW. I saw it demonstrated the September Foothill Amateur Radio Society (FARS) meeting. The following references are probably more appropriate for my talk rather than being in an abstract. DDF2020T Software http://www.kn2c.us/ddf2020-software Review of the Global TSCM DDF2020T Doppler Set http://www.homingin.com/tscmfix.html New Apps for Hidden Transmitter Hunting http://www.homingin.com/apps.html SigTrax - The Evolution of Signal Tracking http://www.amcept.com/sigtrax Reviews for KN2C Radio Direction Finder DF2020T https://www.eham.net/reviews/view-product?id=10837 Pseudo Doppler Direction Finding System for Localizing Non-Cooperative VHF Transmitters with a Hybrid UAS - William E. Gerhard III https://pdfs.semanticscholar.org/ed48/0b9797cef11c7c7a6bd3f4e8a5f3ab9ecd30.pdf Here is a related project under development for an improved approach to Doppler RDF using better modern hardware. https://hackaday.io/project/163957-yet-another-doppler-direction-finder-w-teensy-dsp/details Summary of Steve K6OIK's alternative to Doppler RDF extracted from his email: "It was a 4-channel SDR system that did coherent array signal processing. Up to four simultaneous cochannel signals could be detected, separated, and recovered. For the demo three or four people would walk around randomly with GMRS HTs talking on the same frequency, walking in circles or at close separations.The algorithm was exceptionally good. We demonstrated recovery of a weak signal masked by a jammer 100 dB stronger, i.e. input SINR = -100 dB and full recovery. This was done in the mid to late 1990s. Pacific Division Director Brad Wyatt and SCV Section Manager Glenn Smith saw an early demo. I presented the Jam Resistant Repeater concept at ARRL Pacificon in 2000. The repeater would prevent doubling by separating signals. The repeater system included automatic signal recognition (radio fingerprinting), DOA direction estimation, and auto logging functions (to make an OO's job easy). Direction finding was done by a superresolution algorithm superior to MUSIC. A control operator could program how the repeater would handle signals – either store and forward (like DMR does today) or route a recognized jammer to the bit bucket (meaning no retransmission). The idea never caught on. Perhaps because nobody wanted the job of writing FPGA code. It would be easy to do today. Everything has improved from ADCs and SDR receivers to processors and programming languages."