This is what development looks like!
Open Research Institute organized and executed the RF Village at BSides San Diego on 4 April 2026. This highly anticipated sold-out annual event has a focus on cybersecurity and DIY problem solving. Held at Montezuma Hall at San Diego State University, the one-day event had multiple speaking tracks, at least three Capture the Flag contests (CTFs), an electronic hackable badge, a variety of food and drink available throughout the day, extensive volunteer support, relevant and timely workshops, an After Hours party with more food and drink at Aztec Lanes bowling alley, and a vibrant Village Square that combined Villages and Vendors.
ORI served as the organizer for RF Village, bringing four staff members and multiple exhibits. We debuted our Lunar Lander CTF, announced the W6ORI amateur radio club, and gave away a large box of Amateur Satellite handbooks. We had RFBitBanger kits available for donation (5 were sold), hosted a live Meshcore node, and exhibited a real live FPGA development station with lab equipment for Opulent Voice. Our poster session included Authentication and Authorization and the technical side of RFBitBanger. Thank you to BSides San Diego for the excellent support, with rotating village volunteer staffers, volunteer green room, excellent communications before during and after the event, and genuine care for positive participant experience. Organizations like this make demonstrating open source digital radio work a real joy instead of a daunting chore.
Demonstrations give you deadlines, documentations, and get things “done”. The BSides Opulent Voice demonstration revealed some immediate problems with the 24 dB transmitter fix. The signal was clearly being transmitted at sufficient power, but the symbol and frame lock were not happening. We were seeing “garbage” frames where we were expecting to see actual live data.
Since the over-the-air voice call had worked so well just a few days prior, what was going on? The demonstration was still very successful, as plenty could be shown to the steady stream of people at the RF Village. As the day progressed, more information was gathered. It was very clear we’d have to go back to the lab to figure out how a badly needed transmitter fix had broken the receiver. Why was it working over the air, and not in RF loopback on the bench? First, we went back to the VHDL-only test bench. This sends 10 frames into the transmitter, routes the transmit I and Q streams right back to the receiver, and then demodulates, decodes, and displays the frames. The data in should match the data out. And, frames were coming out, but they were completely scrambled! Something had gone wrong in the RF loopback.
The difference between simulations and real life is usually a lot, and Opulent Voice is no different. A real hardware RF loopback has the radio chip in the loop. The VHDL test bench has only the FPGA contents. We don’t have analog to digital converters (ADCs), digital to analog converters (DACs), or anything else that is in the radio chip. Since the transmitter fix had a lot to do with how the transmitter DAC dealt with the data, it seemed reasonable to assume that the transmitter fix had upset the receiver.
The missing gain was in the transmitter, but the fix applied some math changes to both the transmiter and the receiver, and investigating this took up part of the next day back in the lab. WIth two minor changes, the test bench started working flawlessly again. A new version of the firmware was created, and… it didn’t work in hardware at all! Symbol lock and frame lock were totally non-functional. The plot had certainly thickened.
This is one of the many reasons why demonstrations are so valuable. We find things that we might not go looking for, and it forces us to regularly show things working end-to-end. Work continues this week in the lab to separate the transmitter fix from inadvetently affecting the receiver, and bring us back to working perfectly over the air as well as in simulation.