NOAA HRPT (1.7 GHz)

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.

Most people receive NOAA weather satellite images using an omnidirectional antenna at 137 MHz (APT mode), producing ~4 km/pixel images. The same satellites also transmit High Resolution Picture Transmission (HRPT) at 1698-1707 MHz, delivering 1.1 km/pixel imagery across five spectral channels — visible, near-IR, and three thermal IR bands. The catch is that HRPT requires a directional antenna that tracks the satellite throughout the pass. That is exactly what Birdcage provides.

The signal

NOAA POES satellites (NOAA-15, NOAA-18, NOAA-19) orbit at approximately 850 km altitude in sun-synchronous polar orbits. Each satellite transmits HRPT continuously on a fixed frequency:

Satellite	HRPT frequency	Orbit altitude	Status
NOAA-15	1702.5 MHz	807 km	Active (degraded AVHRR)
NOAA-18	1707.0 MHz	854 km	Active
NOAA-19	1698.0 MHz	870 km	Active

MetOp satellites (MetOp-B, MetOp-C) also transmit HRPT in this band at similar frequencies, using AHRPT (Advanced HRPT) with QPSK modulation at a higher data rate.

HRPT uses BPSK modulation at 665.4 kbps with a 3 MHz occupied bandwidth. The signal carries all five AVHRR (Advanced Very High Resolution Radiometer) channels simultaneously, plus telemetry.

HRPT vs. APT

Most amateur weather satellite enthusiasts start with APT (Automatic Picture Transmission) at 137 MHz, receivable with a V-dipole and an RTL-SDR. APT is analog FM, 2 channels only (one visible, one IR), and 4 km/pixel resolution. HRPT is the same satellite’s high-resolution digital downlink: 5 channels, 1.1 km/pixel, and full 10-bit radiometric depth per channel. The tradeoff is that HRPT requires a tracking antenna and more complex demodulation.

The table below summarizes the differences:

Property	APT (137 MHz)	HRPT (1.7 GHz)
Antenna	Omnidirectional (V-dipole, turnstile)	Tracking dish or helical
Polarization	RHCP (but linear works)	RHCP (polarization match matters)
Resolution	~4 km/pixel	~1.1 km/pixel
Channels	2 (A + B)	5 (all AVHRR)
Modulation	AM/FM (2400 Hz subcarrier)	BPSK (665.4 kbps)
Bandwidth	~40 kHz	~3 MHz
Decoding	wxtoimg, noaa-apt	SatDump

Link budget

Parameter	Value	Notes
Satellite EIRP	~37 dBm (5 W into ~6 dBi antenna)	Typical NOAA POES
Frequency	1698-1707 MHz	L-band
Estimated dish gain	~18 dBi	Better than hydrogen line due to shorter wavelength
Beamwidth (est.)	~12 degrees	Sufficient to track LEO passes without losing signal
Path loss (overhead)	~163 dB	~850 km range
Path loss (10 deg EL)	~171 dB	~2400 km slant range
LNA noise figure	0.5-0.7 dB	Nooelec SAWbird NOAA
System noise temp	~80 K	With filtered LNA
Required C/N	~6 dB	BPSK at 665 kbps
Link margin (overhead)	~20 dB	Comfortable
Link margin (10 deg EL)	~12 dB	Still workable

The 18 dBi dish gain provides substantial link margin. Most amateur HRPT stations use 60-90 cm dishes or helical antennas with 12-15 dBi gain and succeed. The extra margin means reception should be possible even at low elevation angles where the slant range is longest and atmospheric attenuation is greatest.

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	RHCP helical (5-7 turns) for 1.7 GHz	RHCP patch array
LNA	Nooelec SAWbird NOAA (1698 MHz filtered)	Wideband LNA + 1700 MHz bandpass
SDR	RTL-SDR V4 (2.56 MHz stable BW covers one satellite)	BladeRF (captures all three simultaneously)
Bandwidth needed	~3 MHz per satellite	RTL-SDR handles one at a time; BladeRF can do all three

Circular polarization matters

NOAA HRPT is transmitted with right-hand circular polarization (RHCP). A linearly polarized feed loses 3 dB from polarization mismatch. Use an RHCP helical feed to capture the full signal — at low elevation angles where margin is tight, that 3 dB matters.

Proposed procedure

1. Mount the 1.7 GHz RHCP feed at the dish focal point. A 5-7 turn helical antenna designed for 1700 MHz, centered on the reflector focal point. Secure the coax routing so it doesn’t bind during rapid AZ/EL moves.

2. Connect the signal chain. Feed to SAWbird NOAA LNA, LNA via coax to RTL-SDR. Power the LNA via bias tee. Apply a DC block if using the dish’s internal coax path — the 12-18V LNB bias is present on the center conductor.

3. Predict a pass in Gpredict. NOAA satellites have multiple passes per day. Look for a pass with maximum elevation above 30 degrees for best results. Note the start time, duration, and AOS/LOS azimuth.

4. Configure Gpredict to drive Birdcage. Set up a rotator in Gpredict pointing to 127.0.0.1:4533 (the Birdcage rotctld interface). Select the target satellite and enable tracking. Gpredict will push AZ/EL updates throughout the pass.

5. Start the SDR recording before AOS. Tune the SDR to the satellite’s HRPT frequency with at least 2.5 MHz bandwidth. Begin recording raw IQ samples to disk. A 12-minute pass at 2.56 Msps (8-bit) produces roughly 3.7 GB of data.

6. Let the pass run. Gpredict drives the positioner, the SDR captures continuously. Monitor the waterfall display in SDR++ — the HRPT signal should appear as a ~3 MHz wide carrier that strengthens as the satellite rises, peaks near overhead, and fades as it sets.

7. Stop recording after LOS. The full pass IQ file is ready for offline decoding.

Software pipeline

Software	Role
Gpredict	Pass prediction, real-time rotator control via rotctld
SDR++	SDR tuning, waterfall monitoring, IQ recording
SatDump	HRPT demodulation + image decoding (recommended)
GNU Radio	Custom demodulation flowgraphs (if needed)

SatDump is the recommended decoder. It handles NOAA HRPT natively — feed it the IQ recording and it demodulates the BPSK signal, extracts all five AVHRR channels, applies calibration, and produces georeferenced imagery. It can also operate in real-time mode, demodulating as the SDR captures.

For live decoding, SatDump can connect directly to the SDR while Gpredict drives the positioner. The two programs run independently — SatDump on the USB SDR, Gpredict on the serial positioner.

SatDump pipeline

SatDump’s NOAA HRPT pipeline processes the signal in several stages:

1. Demodulation — BPSK carrier recovery, symbol timing, bit sync
2. Frame sync — locates HRPT minor frame boundaries (11090 words per frame)
3. AVHRR extraction — demultiplexes the five AVHRR channels from the data stream
4. Calibration — applies per-channel calibration coefficients from telemetry (converts raw counts to radiance or brightness temperature)
5. Georeferencing — maps each scanline to Earth coordinates using satellite ephemeris and scan geometry
6. Compositing — generates false-color and enhanced images from channel combinations

The entire pipeline runs in a few seconds on recorded data, or in real-time during a pass. SatDump outputs images in multiple projections (equirectangular, Mercator) and formats (PNG, GeoTIFF).

Expected results

A successful HRPT reception produces five-channel imagery of the satellite’s ground track:

Channel	Band	Wavelength	Content
1	Visible	0.58-0.68 um	Daytime cloud/surface imagery
2	Near-IR	0.725-1.0 um	Vegetation, land/water boundaries
3A/3B	Mid-IR / Thermal	1.58-1.64 um / 3.55-3.93 um	Snow/ice (3A) or nighttime thermal (3B)
4	Thermal IR	10.3-11.3 um	Sea surface temperature, cloud tops
5	Thermal IR	11.5-12.5 um	Sea surface temperature, moisture

A single overhead pass covers a ground swath approximately 2,800 km wide. At 1.1 km/pixel, coastlines, large rivers, mountain ranges, and cloud structures are clearly resolved. False-color composites (combining channels 1, 2, and 4) produce striking images showing vegetation in green, water in dark blue, and clouds in white.

Pass duration depends on maximum elevation: approximately 8 minutes for a 30-degree pass, up to 14 minutes for a near-overhead pass.

Beyond NOAA: MetOp and Meteor

The same hardware chain receives other weather satellites in this band:

• MetOp-B and MetOp-C (EUMETSAT) transmit AHRPT (Advanced HRPT) at 1701.3 MHz using QPSK at 2.33 Mbps. SatDump handles AHRPT decoding. MetOp provides the same AVHRR channels as NOAA plus additional instruments (IASI infrared sounder, ASCAT scatterometer telemetry). AHRPT requires a wider bandwidth (~4 MHz) and higher SNR due to the faster data rate, but the dish gain should provide sufficient margin.

• Meteor-M N2-3 (Roscosmos) transmits HRPT at 1700.0 MHz with similar parameters to NOAA HRPT. SatDump supports Meteor HRPT natively. The MSU-MR instrument provides 6 channels at 1 km resolution.

All of these are in sun-synchronous polar orbits with similar pass geometries. The same feed, LNA, and tracking configuration works for all of them — just retune the SDR by a few MHz between satellites.

Tracking considerations

LEO tracking is the most demanding use case for the Birdcage positioner. A few things to verify:

• Angular velocity. Near-overhead passes produce angular rates up to 2 degrees/second. The AZ motor maxes out at 65 degrees/second and EL at 45 degrees/second — far more than enough. The concern is not top speed but command latency through the rotctld pipeline.
• Rotctld update rate. Gpredict updates the rotator at roughly 1-second intervals. At 2 degrees/second, a 1-second update lag means the beam could be ~2 degrees off target. With a 12-degree beamwidth, this is still within the main lobe — but just barely. Higher-elevation passes have higher angular rates.
• Meridian flip. Passes that cross directly overhead may require a rapid AZ reversal (the satellite goes from, say, AZ 180 to AZ 0 instantaneously in the Gpredict model). The 0-to-180-to-360 rotator mode should handle this, but confirm the positioner doesn’t stall or wrap.
• Minimum elevation. The firmware enforces an 18-degree minimum elevation. Satellite passes below 18 degrees EL will be clipped. For HRPT, this is acceptable — the link margin at 18 degrees is still ~14 dB.

Open questions

• Tracking latency end-to-end. The full chain is Gpredict TLE prediction, TCP to rotctld, serial to firmware, motor response. Each step adds latency. Measuring the total lag (command to physical movement) is critical for determining whether smooth LEO tracking is achievable.
• IQ recording vs. live decode. Recording the full-bandwidth IQ stream and decoding offline avoids real-time processing pressure but requires disk space. Live decode with SatDump avoids storage issues but adds CPU load during the pass.
• Feed cable management. During a full LEO pass, the dish may sweep 180+ degrees in AZ and 18-65 degrees in EL. The feed cable must not bind, kink, or snag. A coiled service loop and cable clips along the arm are essential.
• Multiple satellites per day. NOAA-15/18/19 are all active, producing 6-12 visible passes daily from most locations. Automating the track-record-decode cycle would allow routine daily imagery collection.
• Doppler compensation. The satellite’s radial velocity causes up to +/- 35 kHz Doppler shift at 1.7 GHz. BPSK demodulation in SatDump handles this natively (the PLL tracks the carrier), but if recording raw IQ, make sure the SDR center frequency is close enough to the nominal frequency that the Doppler-shifted signal stays within the passband.
• Ground station antenna temperature. At low elevation angles, the dish sidelobes illuminate warm ground (290 K), increasing system noise temperature. Reception at 18 degrees elevation (the firmware minimum) will show measurably lower SNR than an overhead pass — the link budget accounts for this via increased path loss, but ground noise adds another 1-2 dB of degradation not captured in free-space calculations.

Why not just use APT?

A fair question. APT at 137 MHz needs no tracking dish, costs almost nothing, and produces recognizable weather images. Three reasons to pursue HRPT:

1. Resolution. 1.1 km/pixel vs. 4 km/pixel is the difference between seeing individual cities and seeing vague blobs. Coastlines are crisp, cloud streets are resolved, and tropical cyclone eye structure is visible.
2. Five channels. APT carries only 2 of the 5 AVHRR channels per pass (and the pairing rotates). HRPT provides all 5 simultaneously, enabling quantitative sea surface temperature mapping, vegetation index computation, and snow/ice classification.
3. Tracking validation. HRPT reception is a binary pass/fail test of the entire Birdcage LEO tracking pipeline. If you can decode a complete HRPT pass, the positioner is validated for all LEO satellite work — Iridium, Meteor, MetOp, and future missions.

The dish and tracking capability are already there. HRPT reception extracts dramatically more value from each satellite pass.

Related experiments

Iridium (1626 MHz) Similar LEO tracking requirements and L-band feed -- test tracking latency with Iridium bursts before committing to a full HRPT decode.

GPS/GNSS (1575 MHz) An L-band feed for GPS can also cover NOAA HRPT with minor retuning, if a wideband helical is used.