EME Monitoring (1296 MHz)

Not yet tested

This experiment describes an established amateur radio technique adapted for the Birdcage positioner. The hardware chain and procedures are proposed but not yet validated on this specific platform.

Earth-Moon-Earth (EME) communication works by bouncing signals off the lunar surface. Amateurs transmit on 1296 MHz (the 23 cm band), the signal travels ~384,000 km to the Moon, reflects off the regolith, and returns another 384,000 km to Earth. The round-trip path loss is enormous — roughly 271 dB — so the signals that arrive back are extraordinarily weak. Digital modes like Q65 (part of WSJT-X) are designed to decode signals buried well below the noise floor, making receive-only EME monitoring feasible with smaller antennas than you would need for a two-way contact.

This is a listen-only experiment. We are not transmitting — just pointing the dish at the Moon and listening for other stations’ reflected signals.

The signal

EME activity on 1296 MHz concentrates during scheduled contests and activity weekends. Stations typically run 200-1500 watts into 2-6 meter dishes or large Yagi arrays, producing EIRP values of 60-75 dBm. The Moon reflects roughly 7% of incident power (geometric albedo at microwave frequencies), distributed across its entire disk.

After the round-trip, the signal arriving at Earth is staggeringly faint. A typical 1 kW EME station with a 3-meter dish produces a flux density at our antenna that is 15-25 dB below the thermal noise floor in any reasonable bandwidth. This is where digital signal processing earns its keep: Q65 and JT65 use long integration times (60-second or 120-second periods) and forward error correction to pull signals out of noise.

EME signals exhibit several distinctive characteristics:

• Libration fading — the Moon’s slight rocking causes multipath between different reflecting regions, producing deep fades every few seconds
• Doppler spread — the Moon’s rotation smears the reflected signal by ~30 Hz at 1296 MHz
• Polarization rotation — Faraday rotation in the ionosphere rotates the polarization plane unpredictably

Link budget

Parameter	Value	Notes
Frequency	1296 MHz	23 cm amateur band
Wavelength	23.1 cm
Moon distance	~384,400 km	Average
Path loss (round-trip)	~271 dB	Free-space, 2x Earth-Moon distance
Moon reflection loss	~12 dB	Geometric albedo ~7% at microwave
Typical TX station EIRP	+65 dBm	1 kW into 25 dBi antenna
Signal at Earth	~ -218 dBm	Spread across ~30 Hz Doppler
Our dish gain (est.)	~15 dBi	84 cm dish at 1296 MHz, 50% efficiency
LNA noise figure	0.5-0.8 dB	Wideband LNA + bandpass filter
System noise temp	~60-100 K	Depends on sky temp, ground spillover
Noise power (30 Hz BW)	~ -199 dBm	kTB at 100 K
Expected SNR	~ -34 dB	In 30 Hz — well below noise
Q65 decode threshold	~ -27 dB	In 2500 Hz reference bandwidth

The numbers are marginal. With 15 dBi receive gain, we are roughly 7-10 dB below the typical Q65 decode threshold for a single 60-second period. This means:

• During contests, when many stations run high power and large antennas, the strongest signals may be decodable
• On quiet days, detection is unlikely
• Stacking multiple periods (averaging across several transmit cycles) can recover a few dB, bringing marginal signals over the threshold

Honest assessment: this is at the edge

Most successful EME receive stations use 25-35 dBi of antenna gain. Our 15 dBi puts us 10-20 dB below typical. This experiment will work during EME contests with the strongest stations, but don’t expect casual all-day monitoring to produce decodes. It’s a challenge experiment — success depends on timing, pointing accuracy, and luck.

Hardware requirements

SDR Hardware Setup Feed options, LNA selection, bias tee safety, and SDR configuration shared across all external SDR experiments.

Component	Recommended	Alternatives
Feed	5-7 turn RHCP helical for 1296 MHz	Short Yagi as feed, patch
LNA	Wideband LNA (0.5 dB NF) + 1296 MHz cavity bandpass filter	Any low-NF LNA with adequate filtering
SDR	BladeRF 2.0 micro (12-bit dynamic range helps)	RTL-SDR V4 (8-bit, marginal for signals this weak)
Bandwidth needed	~5 kHz (Q65 occupies ~65 Hz)	Record wider for monitoring multiple stations

There is no off-the-shelf SAWbird for 1296 MHz specifically. The recommended approach is a wideband LNA (Mini-Circuits, Qorvo, or similar) followed by a separate 1296 MHz bandpass filter. The filter is critical — cell tower transmissions in the 1700-2100 MHz range can desensitize or saturate the LNA if unfiltered.

The BladeRF’s 12-bit ADC provides roughly 24 dB more dynamic range than the RTL-SDR’s 8-bit ADC. For signals buried in noise, this extra dynamic range helps the digital correlator in WSJT-X distinguish signal from quantization artifacts.

Why 1296 MHz and not 144 MHz or 432 MHz

EME is active on several bands. The choice of 1296 MHz for this experiment is driven by the dish geometry:

Band	Frequency	Dish gain (est.)	Beamwidth	EME activity
2 m	144 MHz	~0 dBi	> 90 degrees	Highest (most stations)
70 cm	432 MHz	~5 dBi	~50 degrees	Moderate
23 cm	1296 MHz	~15 dBi	~15 degrees	Lower (but growing)

At 144 MHz, the dish is less than half a wavelength across and provides no useful gain — a Yagi would outperform it. At 432 MHz, the gain is marginal. At 1296 MHz, the dish finally provides meaningful gain, and the narrower beam better rejects ground noise when pointed at the Moon. The tradeoff is fewer active EME stations on 23 cm compared to 2 meters, which is why contest weekends matter.

Proposed procedure

1. Mount the 1296 MHz feed. A 5-7 turn RHCP helical antenna mounted at the dish focal point. RHCP is standard for 1296 MHz EME, though Faraday rotation randomizes the arriving polarization — a circularly polarized feed loses less on average than a linear one.

2. Connect the signal chain. Feed to wideband LNA, through 1296 MHz bandpass filter, via coax to BladeRF. Power the LNA via bias tee or separate supply. DC block required if using the dish’s internal coax.

3. Check the EME contest calendar. The best times to attempt this experiment are during major EME contests:

   • ARRL EME Contest (typically October and November weekends)
   • Dubus EME Contest (spring)
   • ARI EME Trophy (various dates)
   • Regular activity time windows listed on the N0UK EME page

4. Calculate the Moon’s position. Gpredict can track the Moon as a celestial target. Configure the rotator interface and verify the dish tracks the Moon’s ~0.5 degree/minute drift rate smoothly.

5. Tune the SDR to 1296.000-1296.100 MHz. Most EME activity concentrates in the first 100 kHz of the band. Record IQ at a sample rate that covers this window (~200 kHz is sufficient).

6. Run WSJT-X in Q65-60C mode. Configure WSJT-X with the SDR as its audio source (via virtual audio cable or direct SDR integration). Set the mode to Q65-60C (60-second periods, C submodes for deeper decoding). Let it run continuously during the contest period.

7. Monitor for decodes. WSJT-X will display decoded callsigns, signal reports, and grid squares. Even partial decodes (callsign fragments) count as successful detection.

Software pipeline

Software	Role
Gpredict	Moon tracking, rotator control
SDR++	Spectrum monitoring, IQ recording
WSJT-X	Q65 / JT65 decoding (purpose-built for EME)
MAP65	Wideband panoramic EME decoder (monitors entire band segment)

WSJT-X is the standard tool for EME digital modes. MAP65 is its companion for wideband monitoring — it decodes all Q65 signals in a ~90 kHz passband simultaneously, which is ideal for contest monitoring where many stations are active across the band.

Audio routing

WSJT-X expects an audio input, not raw IQ. The SDR must be demodulated to baseband audio before feeding WSJT-X. Two approaches:

1. SDR++ as demodulator. Run SDR++ with the BladeRF, tune to 1296.050 MHz (center of EME activity), select USB mode with 2.5 kHz filter bandwidth, and route the demodulated audio to WSJT-X via a virtual audio cable (PulseAudio, JACK, or PipeWire loopback).

2. GNU Radio flowgraph. Build a minimal flowgraph: BladeRF source at 1296.050 MHz, 200 kHz sample rate, complex-to-real conversion, decimation to 48 kHz audio, output to audio sink. This gives you more control over filtering and can feed MAP65 for wideband panoramic decoding.

For MAP65 specifically, the input must be wideband I/Q audio (not USB-demodulated) — MAP65 does its own channelization internally. The GNU Radio approach is more flexible for this.

Tracking: lunar rate

The Moon moves across the sky at approximately 0.5 degrees per minute (combining sidereal motion and the Moon’s own orbital velocity). This is far slower than LEO satellite tracking but much faster than sidereal rate for stars.

Gpredict handles lunar tracking natively. The positioner receives AZ/EL updates every ~1 second. At 0.5 degrees/minute, the Moon moves less than 0.01 degrees between updates — well within the dish beamwidth. Smooth tracking should not be a problem.

One consideration: the Moon is above the horizon for roughly 12 hours per day, but EME signals are strongest when the Moon is high in the sky (shorter atmospheric path, lower ground noise in the beam). Sessions when the Moon is above 30 degrees elevation are preferred.

Elevation constraint

The firmware enforces an 18-degree minimum elevation. The Moon spends a significant fraction of its above-horizon time below 18 degrees, especially at higher latitudes. This clips the usable monitoring window — you lose the first and last ~30 minutes of each moonrise/moonset period. During high lunar declination months, the Moon can reach 60+ degrees elevation from mid-latitudes, giving several hours of usable tracking time.

The 65-degree maximum elevation is less of a concern since the Moon only exceeds this angle for brief periods from most locations, and those high-elevation windows are the ones with the best SNR anyway.

Expected results

During a major EME contest weekend, with the dish pointed at the Moon and WSJT-X running in Q65 mode:

• Optimistic: 2-5 decoded callsigns per hour from the strongest stations (1 kW+ into 3+ meter dishes). These will appear in WSJT-X’s decode pane with signal reports around -25 to -30 dB.
• Realistic: 0-2 decodes per hour, with many partial decodes that don’t meet the error-correction threshold. Stations running very high power from favorable locations will be the most likely detections.
• Minimum success criterion: detecting any EME-reflected signal at all — even a consistent spectral trace at the expected Doppler offset without a full decode — would validate the hardware chain.

Each successful decode gives you the distant station’s callsign, grid square, and signal report, confirming that the round-trip Moon reflection was captured by your dish.

Passive radar bonus

Even without decoding callsigns, the dish can detect the Moon as a radar target. Point the dish at the Moon and compare the noise power in the 1296 MHz band with the noise power when pointed at blank sky at the same elevation. The Moon is a ~230 K thermal emitter at microwave frequencies. With a 15-degree beam and the Moon subtending ~0.5 degrees, the Moon’s thermal emission fills about 0.1% of the beam area, contributing roughly 0.2 K to the antenna temperature. This is probably too small to measure with our system noise temperature of 60-100 K.

However, the Sun is a much stronger thermal source (tens of thousands of K at 1296 MHz) and is easily detectable. A quick ON/OFF measurement of the Sun at 1296 MHz would validate the signal chain and LNA performance before attempting the more challenging EME monitoring.

Open questions

• Is 15 dBi enough to decode anything? The link budget says we are 7-10 dB below typical. During contests, the strongest stations run 70+ dBm EIRP — those extra dB may close the gap. A test during the ARRL EME Contest would answer this definitively.
• Feed reuse with hydrogen line. A 1296 MHz helical feed is close enough to 1420 MHz that it might provide usable (if reduced) gain for hydrogen line observations. Worth measuring the feed’s return loss at both frequencies.
• Polarization strategy. Faraday rotation randomizes arriving polarization. A dual-polarization feed (two orthogonal helicals or a septum feed) with coherent combining could recover up to 3 dB, but adds complexity. Start with single RHCP.
• Ground noise contribution. At low elevation angles, the dish’s sidelobes pick up warm ground (290 K), raising system noise temperature. Pointing the dish at the Moon when it’s low on the horizon may produce worse SNR than the link budget predicts.
• Stacking multiple periods. WSJT-X can average across multiple Q65 periods in some configurations. If single-period decoding fails, multi-period stacking might push marginal signals over the threshold.
• CW and SSB EME. Before digital modes, EME was done in CW (Morse code) and SSB (voice). CW EME requires about 15 dB less SNR than SSB but more than Q65. Some operators still use CW during contests. With 15 dBi of gain, even CW EME reception would be marginal, but not impossible from the strongest stations.
• 432 MHz EME. The 70 cm band (432 MHz) is the other major EME frequency. At 432 MHz, the dish would have only ~5 dBi gain (less than 2 wavelengths across) — too low for practical EME monitoring. The 1296 MHz choice is driven by the dish size.

What makes this experiment worth attempting

The link budget is unfavorable. The dish is undersized. The signals are below the noise floor. So why try?

Because EME reception with a sub-1-meter dish is a genuine challenge that would be a notable result in the amateur EME community. Most 1296 MHz EME receive stations use 2-6 meter dishes or large Yagi arrays. A confirmed decode from an 84 cm dish — even if it only happens during the ARRL EME Contest with the strongest stations — demonstrates what digital signal processing can do at the edge of feasibility. The attempt itself exercises the entire Birdcage signal chain at its limits: precise lunar tracking, low system noise temperature, clean audio routing to WSJT-X, and patience.

And if it doesn’t work on 1296 MHz, the same hardware chain with a 1420 MHz feed immediately pivots to the hydrogen line experiment, where success is virtually guaranteed.

Related experiments

Hydrogen Line (1420 MHz) Similar frequency, similar feed -- a 1296 MHz helical may cover 1420 MHz with reduced gain, enabling both experiments with one feed.

Antenna Pattern Measurement Characterize the dish beam pattern to understand sidelobe levels and actual gain at your operating frequency.