The GCARC WSPR Network
GCARC WSPR Network : Building, Beaconing, and Beyond
By Jonathan Pearce WB2MNF
Two months of planning paid off in a big way when about 15 Club members and guests gathered over two Saturday sessions on May 2nd and 9th to build and test their own TAPR-based WSPR transmitters. The assembly process was straightforward but satisfying : solder four capacitors and four small inductors onto the circuit board, connect it to a Raspberry Pi 3, install a pre-programmed SD card, and attach it to the QRP Watt meter we built as a 2025 Club project. With bias levels set and power output confirmed, each unit was connected to the Clubhouse antenna and the results spoke for themselves. All but one unit received spots within short order (the lone exception was a 10-meter unit that was unlucky enough to transmit during a period when the band was closed). Participants then took their beacons home, connected them to home Wi-Fi and antennas, and joined the network.
The GCARC WSPR Network has grown quickly. We now have 13 unique stations active across five bands - 40, 30, 20, 15, and 10 meters - and were pleased to welcome several friends from the South Jersey Radio Association who joined the project.
To make the most of all that beacon data, we've been building a dedicated reporting website that weaves together WSPR spot data, solar weather, and propagation indicators into a single picture of current band conditions. You can see it at : https://wp.w2mmd.org/wp/Club-activities/current-band-conditions - it includes reports from individual GCARC network stations, the WW0WWV stations in Fort Collins, CO, and a range of other propagation data. We hope it becomes a useful resource for HF operators and a practical planning tool as Field Day approaches.
At Dayton Hamvention this year I had the pleasure of attending a presentation by George Byrkit K9TRV - TAPR's store manager who handled our board purchase and, as anyone who has met him can attest, a genuinely entertaining speaker. George takes a fascinatingly different angle on WSPR : rather than transmitting, he's focused entirely on receiving, running four different antenna systems simultaneously and rigorously comparing their performance. His site near Ann Arbor, MI qualifies as ITU-R "Quiet Rural," but he's still had to hunt down a remarkable variety of noise sources - including, memorably, a noisy LED light fixture in his bathroom shower that was only discovered when his wife happened to turn it on while he was in the shack.
George's four antennas include the DXE RSEAV1 active whip and three variants of the N6GN Single Antenna System (SAS), a compact, balanced design that runs on ordinary CAT5 cable rather than coax - making it both less expensive and less prone to ground loop noise. His top performer turned out to be the Medium-Z SAS with double-gain preamp, followed by the RSEAV1, the High-Z SAS, and the standard Medium-Z SAS - though George is quick to note that your results will differ based on your own noise floor and what interference sources you're willing to tolerate. His practical advice on noise hunting is worth the price of admission alone : walk around your shack with a portable AM radio tuned to a dead frequency and listen to what it hears near your equipment - computers, monitors, test gear, and switching power supplies are all prime suspects.
Receiving WSPR is quite accessible using a Raspberry Pi and an SDR receiver with dedicated WSPRDAEMON software. If that sounds interesting, reach out and we'll put together some resources. George's slides from Dayton are available at : https://files.tapr.org/meetings/Hamvention2026/K9TRV-Xenia2026.pdf.
If you'd like to join the GCARC WSPR Network - whether with a TAPR board or another device - get in touch. We're continuing to develop comparative reporting tools to help network participants tune and optimize their stations, so the best is still ahead.
73, Jon WB2MNF
By Jonathan Pearce WB2MNF
Two months of planning paid off in a big way when about 15 Club members and guests gathered over two Saturday sessions on May 2nd and 9th to build and test their own TAPR-based WSPR transmitters. The assembly process was straightforward but satisfying : solder four capacitors and four small inductors onto the circuit board, connect it to a Raspberry Pi 3, install a pre-programmed SD card, and attach it to the QRP Watt meter we built as a 2025 Club project. With bias levels set and power output confirmed, each unit was connected to the Clubhouse antenna and the results spoke for themselves. All but one unit received spots within short order (the lone exception was a 10-meter unit that was unlucky enough to transmit during a period when the band was closed). Participants then took their beacons home, connected them to home Wi-Fi and antennas, and joined the network.
The GCARC WSPR Network has grown quickly. We now have 13 unique stations active across five bands - 40, 30, 20, 15, and 10 meters - and were pleased to welcome several friends from the South Jersey Radio Association who joined the project.
To make the most of all that beacon data, we've been building a dedicated reporting website that weaves together WSPR spot data, solar weather, and propagation indicators into a single picture of current band conditions. You can see it at : https://wp.w2mmd.org/wp/Club-activities/current-band-conditions - it includes reports from individual GCARC network stations, the WW0WWV stations in Fort Collins, CO, and a range of other propagation data. We hope it becomes a useful resource for HF operators and a practical planning tool as Field Day approaches.
At Dayton Hamvention this year I had the pleasure of attending a presentation by George Byrkit K9TRV - TAPR's store manager who handled our board purchase and, as anyone who has met him can attest, a genuinely entertaining speaker. George takes a fascinatingly different angle on WSPR : rather than transmitting, he's focused entirely on receiving, running four different antenna systems simultaneously and rigorously comparing their performance. His site near Ann Arbor, MI qualifies as ITU-R "Quiet Rural," but he's still had to hunt down a remarkable variety of noise sources - including, memorably, a noisy LED light fixture in his bathroom shower that was only discovered when his wife happened to turn it on while he was in the shack.
George's four antennas include the DXE RSEAV1 active whip and three variants of the N6GN Single Antenna System (SAS), a compact, balanced design that runs on ordinary CAT5 cable rather than coax - making it both less expensive and less prone to ground loop noise. His top performer turned out to be the Medium-Z SAS with double-gain preamp, followed by the RSEAV1, the High-Z SAS, and the standard Medium-Z SAS - though George is quick to note that your results will differ based on your own noise floor and what interference sources you're willing to tolerate. His practical advice on noise hunting is worth the price of admission alone : walk around your shack with a portable AM radio tuned to a dead frequency and listen to what it hears near your equipment - computers, monitors, test gear, and switching power supplies are all prime suspects.
Receiving WSPR is quite accessible using a Raspberry Pi and an SDR receiver with dedicated WSPRDAEMON software. If that sounds interesting, reach out and we'll put together some resources. George's slides from Dayton are available at : https://files.tapr.org/meetings/Hamvention2026/K9TRV-Xenia2026.pdf.
If you'd like to join the GCARC WSPR Network - whether with a TAPR board or another device - get in touch. We're continuing to develop comparative reporting tools to help network participants tune and optimize their stations, so the best is still ahead.
73, Jon WB2MNF
Saturday, May 9, 2026
100 Milliwatts To Australia : The W2MMD WSPR Beacon :
From A Bag Of Parts To Six Continents In 24 Hours
By Jon Pearce WB2MNF
It arrived in a flat cardboard box from China - a couple of bare circuit boards, several small zip-lock bags of electronic components, and a set of instructions that assumed you already knew what you were doing. The QRP Labs Ultimate3S WSPR transmitter kit is not for the faint of heart. There is no enclosure, no pre-wound coils, no plug-and-play convenience. What you get is potential. What you make of it depends entirely on your skill, your patience, and how well you read a schematic.
Enter Chris Prioli AD2CS. Chris took the kit home on a Saturday afternoon and by Sunday night he had a working five-band WSPR transmitter. That might sound straightforward until you consider what was involved : populating two circuit boards with dozens of surface-mount and through-hole components, winding no fewer than fifteen toroid coils for the low-pass filter modules - five separate filter boards, three toroids each, every one wound by hand, turn by turn, to precise specifications. Each filter must pass its intended frequency cleanly while suppressing harmonics well enough to satisfy FCC Part 97 requirements. Chris wound them all at his kitchen table, checked each one, and plugged them into the relay-switched filter board that lets the transmitter hop between bands automatically.
He also installed the GPS receiver module that gives the transmitter its sense of time - critical for WSPR, where transmissions must begin within one second of an even UTC minute or they simply vanish into the noise. Chris added an SMA connector on the back panel for the GPS antenna, a nice professional touch that makes the external GPS antenna easy to connect and disconnect. When he brought the completed unit to the Clubhouse, it was a work of art.
The antenna took a different kind of effort. We ordered a 65-foot end-fed wire antenna from Amazon - nothing exotic, just a long wire with a 49:1 matching transformer on one end. We ran it from the Skunkworks tower across the parking lot to the light pole at the far end, feeding it with 50 feet of RG-8X coax. First measurements with the VNA showed it was too long; the resonant frequencies were all below where we needed them. So we shortened it. Then shortened it again. About four feet of trimming later, the VNA showed less than 2:1 SWR on 40, 20, 15, and 10 meters. Thirty meters was another story - 8:1 SWR, a significant mismatch - but at 100 milliwatts, even a badly matched antenna won't damage anything. We decided to try it anyway.
We powered up the transmitter, confirmed roughly 100 milliwatts of output on each band with a wattmeter, verified that the signal looked clean on a spectrum analyzer, and waited for spots to appear on WSPRnet. And waited. And waited.
Nothing.
This began a troubleshooting session that stretched deep into the night. We set up WSJT-X on a laptop connected to an SDR receiver and confirmed that we could decode our own signal locally - proof that the WSPR encoding, timing, and modulation were all correct. The packets were properly formed. The callsign W2MMD was embedded correctly. The grid square FM29jr was right. But no station anywhere in the world was decoding us.
The first clue came from the SDR waterfall display. When we zoomed in on the transmitted signal, it wasn't where it was supposed to be. WSPR signals need to land within a very specific 200 Hertz-wide slice of each band.
Ours was more than a kilohertz off - completely outside the window where any receiving station would be listening. Every WSPR receiver in the world was tuned to the right frequency. We were transmitting to an empty room.
The root cause turned out to be subtle. Our Si5351A synthesizer module - the chip that generates the actual RF signal - had a 25 MHz temperature-compensated crystal oscillator installed instead of the standard 27 MHz crystal that the QRP Labs manual assumes. The reference frequency parameter in the firmware was set to the correct value for a 25 MHz oscillator, but the interaction between this non-standard reference and the GPS calibration routine was causing problems. The GPS calibration would measure the crystal frequency accurately, calculate a correction, and apply it - but the resulting output frequencies were systematically offset on every band. The calibration was doing exactly what it was designed to do, and it was making things worse.
The solution was counterintuitive : we disabled the GPS frequency calibration entirely and applied manual frequency corrections to each band individually. We kept the GPS for timing - it still synchronizes our transmissions to UTC with sub-second accuracy - but we told the firmware to stop trying to correct the oscillator frequency. Then, band by band, we measured the actual output frequency on the SDR, compared it to where it should be, calculated the offset, and adjusted the programmed frequency to compensate. Each band needed a different correction. We iterated until every band landed within the WSPR decode window.
At 1820 UTC on April 23rd, the first spot appeared on WSPRnet. K3FEF, 177 kilometers away in FN21, decoded W2MMD on 40 meters. Within minutes, more spots rolled in - N8GA in Ohio, N9EAT in Iowa, stations up and down the East Coast. We switched to monitoring 20 meters. More spots. Then 15 meters, and stations in Europe started appearing: DK4RW in Germany, HB9VQQ in Switzerland, OE9GHV in Austria, all decoding our 100-milliwatt signal at distances exceeding 6,500 kilometers. Then 10 meters opened, and PU1JSV in Brazil - 7,685 kilometers away - decoded us.
And then, on 20 meters, VK5ARG in South Australia pulled our signal out of the noise at 16,965 kilometers. Nearly halfway around the planet. One tenth of a watt into a wire strung across a parking lot.
Within the first 4 hours of operation, W2MMD accumulated 3,905 spots from 271 unique receiving stations across six continents. Forty meters proved to be the workhorse, producing 2,265 spots from 155 receivers with reliable coverage day and night. Twenty meters emerged as the surprise performer - 834 spots reaching as far as Australia on long-path propagation, making it the undisputed DX band for the station. Fifteen meters reached Brazil (PY4XC, 7,464 km) and deep into Europe. Ten meters, despite reduced output power at 28 MHz due to synthesizer rolloff, reached Brazil at 7,685 km. Even 30 meters, fighting that 8:1 antenna mismatch and radiating perhaps 30 effective milliwatts, reached Austria - OE9GHV at 6,558 km - proving that WSPR can do extraordinary things with almost nothing.
The station hit Europe on four of the five bands. It reached South America on three. Africa on two. And Australia - on a single 20-meter long-path spot that probably shouldn't have been possible but was.
The W2MMD beacon is now running continuously from the Clubhouse, transmitting two-minute beacons sequentially on each of the five bands. Our AI-driven analysis server monitors the spot data and will generate daily propagation reports published to the Skunkworks website at skunkworks.w2mmd.org. We're currently building out an analytics dashboard to visualize these results - propagation by band, by time of day, by world region - giving GCARC members real insight into what the ionosphere is doing from our geographic location.
This is where the W2MMD beacon becomes more than a novelty. Because the transmitter location, antenna, output power, and transmission schedule are completely fixed, every variation in the spot data reflects propagation conditions alone - not operator choices, not antenna differences, not power levels. When W2MMD reaches Europe on 15 meters on a Tuesday afternoon but not on a Wednesday, that's the ionosphere talking, not equipment. That consistency is what makes the data scientifically useful, and it's exactly the kind of controlled baseline that we need for meaningful propagation research.
We already have about 40 stations signed up to participate in the GCARC WSPR Network, each running their own beacon. The Clubhouse station complements the network by providing that fixed reference point - a known quantity against which individual member stations can compare their own results. If your home station isn't reaching Europe on 15 meters but W2MMD is, the problem isn't the band - it's your antenna.
Chris built something remarkable from a bag of parts and a stack of toroids. The rest of us figured out how to make it talk to the world. And now it sits quietly in the Clubhouse, transmitting its 100-milliwatt whisper into the ether every two minutes, and stations from New Jersey to South Australia are listening.
From A Bag Of Parts To Six Continents In 24 Hours
By Jon Pearce WB2MNF
It arrived in a flat cardboard box from China - a couple of bare circuit boards, several small zip-lock bags of electronic components, and a set of instructions that assumed you already knew what you were doing. The QRP Labs Ultimate3S WSPR transmitter kit is not for the faint of heart. There is no enclosure, no pre-wound coils, no plug-and-play convenience. What you get is potential. What you make of it depends entirely on your skill, your patience, and how well you read a schematic.
Enter Chris Prioli AD2CS. Chris took the kit home on a Saturday afternoon and by Sunday night he had a working five-band WSPR transmitter. That might sound straightforward until you consider what was involved : populating two circuit boards with dozens of surface-mount and through-hole components, winding no fewer than fifteen toroid coils for the low-pass filter modules - five separate filter boards, three toroids each, every one wound by hand, turn by turn, to precise specifications. Each filter must pass its intended frequency cleanly while suppressing harmonics well enough to satisfy FCC Part 97 requirements. Chris wound them all at his kitchen table, checked each one, and plugged them into the relay-switched filter board that lets the transmitter hop between bands automatically.
He also installed the GPS receiver module that gives the transmitter its sense of time - critical for WSPR, where transmissions must begin within one second of an even UTC minute or they simply vanish into the noise. Chris added an SMA connector on the back panel for the GPS antenna, a nice professional touch that makes the external GPS antenna easy to connect and disconnect. When he brought the completed unit to the Clubhouse, it was a work of art.
The antenna took a different kind of effort. We ordered a 65-foot end-fed wire antenna from Amazon - nothing exotic, just a long wire with a 49:1 matching transformer on one end. We ran it from the Skunkworks tower across the parking lot to the light pole at the far end, feeding it with 50 feet of RG-8X coax. First measurements with the VNA showed it was too long; the resonant frequencies were all below where we needed them. So we shortened it. Then shortened it again. About four feet of trimming later, the VNA showed less than 2:1 SWR on 40, 20, 15, and 10 meters. Thirty meters was another story - 8:1 SWR, a significant mismatch - but at 100 milliwatts, even a badly matched antenna won't damage anything. We decided to try it anyway.
We powered up the transmitter, confirmed roughly 100 milliwatts of output on each band with a wattmeter, verified that the signal looked clean on a spectrum analyzer, and waited for spots to appear on WSPRnet. And waited. And waited.
Nothing.
This began a troubleshooting session that stretched deep into the night. We set up WSJT-X on a laptop connected to an SDR receiver and confirmed that we could decode our own signal locally - proof that the WSPR encoding, timing, and modulation were all correct. The packets were properly formed. The callsign W2MMD was embedded correctly. The grid square FM29jr was right. But no station anywhere in the world was decoding us.
The first clue came from the SDR waterfall display. When we zoomed in on the transmitted signal, it wasn't where it was supposed to be. WSPR signals need to land within a very specific 200 Hertz-wide slice of each band.
Ours was more than a kilohertz off - completely outside the window where any receiving station would be listening. Every WSPR receiver in the world was tuned to the right frequency. We were transmitting to an empty room.
The root cause turned out to be subtle. Our Si5351A synthesizer module - the chip that generates the actual RF signal - had a 25 MHz temperature-compensated crystal oscillator installed instead of the standard 27 MHz crystal that the QRP Labs manual assumes. The reference frequency parameter in the firmware was set to the correct value for a 25 MHz oscillator, but the interaction between this non-standard reference and the GPS calibration routine was causing problems. The GPS calibration would measure the crystal frequency accurately, calculate a correction, and apply it - but the resulting output frequencies were systematically offset on every band. The calibration was doing exactly what it was designed to do, and it was making things worse.
The solution was counterintuitive : we disabled the GPS frequency calibration entirely and applied manual frequency corrections to each band individually. We kept the GPS for timing - it still synchronizes our transmissions to UTC with sub-second accuracy - but we told the firmware to stop trying to correct the oscillator frequency. Then, band by band, we measured the actual output frequency on the SDR, compared it to where it should be, calculated the offset, and adjusted the programmed frequency to compensate. Each band needed a different correction. We iterated until every band landed within the WSPR decode window.
At 1820 UTC on April 23rd, the first spot appeared on WSPRnet. K3FEF, 177 kilometers away in FN21, decoded W2MMD on 40 meters. Within minutes, more spots rolled in - N8GA in Ohio, N9EAT in Iowa, stations up and down the East Coast. We switched to monitoring 20 meters. More spots. Then 15 meters, and stations in Europe started appearing: DK4RW in Germany, HB9VQQ in Switzerland, OE9GHV in Austria, all decoding our 100-milliwatt signal at distances exceeding 6,500 kilometers. Then 10 meters opened, and PU1JSV in Brazil - 7,685 kilometers away - decoded us.
And then, on 20 meters, VK5ARG in South Australia pulled our signal out of the noise at 16,965 kilometers. Nearly halfway around the planet. One tenth of a watt into a wire strung across a parking lot.
Within the first 4 hours of operation, W2MMD accumulated 3,905 spots from 271 unique receiving stations across six continents. Forty meters proved to be the workhorse, producing 2,265 spots from 155 receivers with reliable coverage day and night. Twenty meters emerged as the surprise performer - 834 spots reaching as far as Australia on long-path propagation, making it the undisputed DX band for the station. Fifteen meters reached Brazil (PY4XC, 7,464 km) and deep into Europe. Ten meters, despite reduced output power at 28 MHz due to synthesizer rolloff, reached Brazil at 7,685 km. Even 30 meters, fighting that 8:1 antenna mismatch and radiating perhaps 30 effective milliwatts, reached Austria - OE9GHV at 6,558 km - proving that WSPR can do extraordinary things with almost nothing.
The station hit Europe on four of the five bands. It reached South America on three. Africa on two. And Australia - on a single 20-meter long-path spot that probably shouldn't have been possible but was.
The W2MMD beacon is now running continuously from the Clubhouse, transmitting two-minute beacons sequentially on each of the five bands. Our AI-driven analysis server monitors the spot data and will generate daily propagation reports published to the Skunkworks website at skunkworks.w2mmd.org. We're currently building out an analytics dashboard to visualize these results - propagation by band, by time of day, by world region - giving GCARC members real insight into what the ionosphere is doing from our geographic location.
This is where the W2MMD beacon becomes more than a novelty. Because the transmitter location, antenna, output power, and transmission schedule are completely fixed, every variation in the spot data reflects propagation conditions alone - not operator choices, not antenna differences, not power levels. When W2MMD reaches Europe on 15 meters on a Tuesday afternoon but not on a Wednesday, that's the ionosphere talking, not equipment. That consistency is what makes the data scientifically useful, and it's exactly the kind of controlled baseline that we need for meaningful propagation research.
We already have about 40 stations signed up to participate in the GCARC WSPR Network, each running their own beacon. The Clubhouse station complements the network by providing that fixed reference point - a known quantity against which individual member stations can compare their own results. If your home station isn't reaching Europe on 15 meters but W2MMD is, the problem isn't the band - it's your antenna.
Chris built something remarkable from a bag of parts and a stack of toroids. The rest of us figured out how to make it talk to the world. And now it sits quietly in the Clubhouse, transmitting its 100-milliwatt whisper into the ether every two minutes, and stations from New Jersey to South Australia are listening.
TAPR.ORG : Tapr Hat for Raspberry Pi :
· https://tapr.org/product/wspr
Raspberry Pi Zero Kit : Canakit.com :
· https://www.canakit.com/raspberry-pi-zero-2-w.html
Discord Group Chat Network : · https://discord.com
· https://tapr.org/product/wspr
Raspberry Pi Zero Kit : Canakit.com :
· https://www.canakit.com/raspberry-pi-zero-2-w.html
Discord Group Chat Network : · https://discord.com
