Hydrogen Line (1420 MHz)

The hydrogen line at 1420.405 MHz is the most accessible target in radio astronomy. Neutral hydrogen atoms throughout the Milky Way emit radiation at this precise frequency when their electron spin flips. The emission is always present, covers the entire galactic plane, and the Doppler shift of the line encodes the radial velocity of the gas — making it possible to map the rotation curve of the galaxy with amateur equipment.

This is the experiment with the deepest community support. Dozens of amateur radio astronomers have detected the hydrogen line with dishes smaller than ours, and the processing pipeline is well-documented.

The signal

Galactic hydrogen emission produces a spectral line centered at 1420.405 MHz (21.106 cm wavelength). The line is not infinitely narrow — thermal motion of the gas and bulk galactic rotation broaden it. Pointing toward the galactic plane, you see multiple velocity components from different spiral arms, spread across roughly 1420.405 MHz +/- 1.4 MHz (corresponding to radial velocities of +/- 300 km/s).

The signal strength varies dramatically with pointing direction. The galactic plane (especially toward the galactic center in Sagittarius) produces the strongest emission, with antenna temperatures of 50-200 K above the background. At high galactic latitudes, the line is still present but much weaker (5-20 K). This contrast is what makes the experiment satisfying — you see the galaxy’s structure in your data.

Link budget

Parameter	Value	Notes
Frequency	1420.405 MHz	21 cm hydrogen line
Wavelength	21.1 cm
Dish diameter	84 cm (major axis)	33” x 23” elliptical
Estimated dish gain	~15 dBi	Assuming 50% aperture efficiency
Beamwidth (est.)	~15 degrees	Coarse resolution, but sufficient for large-scale structure
Source antenna temp	50-200 K	Galactic plane; 5-20 K off-plane
LNA noise figure	0.7 dB	Nooelec SAWbird H1
System noise temp	~80-120 K	LNA + spillover + sky background
Integration time	30-60 seconds	Per pointing, FFT accumulation

The key ratio is the source temperature vs. the system noise temperature. With T_source / T_sys of roughly 1-2 on the galactic plane, the line should be clearly visible after 30-60 seconds of spectral integration per pointing. Off the plane, longer integration (2-5 minutes) may be needed.

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 at 1420 MHz	Patch antenna, can antenna (lower gain)
LNA	Nooelec SAWbird H1 (1420 MHz filtered, 0.7 dB NF)	Any filtered 1420 MHz LNA
SDR	RTL-SDR V4	BladeRF (12-bit helps but isn’t necessary)
Bandwidth needed	~3 MHz centered on 1420.405 MHz	Captures full Doppler range

Even an 8-bit RTL-SDR works well for hydrogen line work. The signal is a spectral line, not a wideband signal — the FFT bins average out quantization noise effectively, and 3 MHz of bandwidth is well within the RTL-SDR’s stable range.

Feed design considerations

The helical antenna is the most common feed choice for amateur hydrogen line work. A 5-turn RHCP helix designed for 1420 MHz has the following approximate dimensions:

Parameter	Value
Turn spacing	0.25 lambda = 5.3 cm
Circumference	1.0 lambda = 21.1 cm
Diameter	~6.7 cm
Total length	~26.5 cm (5 turns)
Ground plane	~0.75 lambda = 15.8 cm diameter
Gain	~12-14 dBi (as standalone antenna)
Beamwidth	~40-50 degrees

The helical feed should illuminate the dish reflector evenly. If the feed beamwidth is too narrow, the dish edges are under-illuminated and effective aperture decreases. If too wide, energy spills past the dish edges and picks up ground noise. The optimal feed beamwidth depends on the dish f/D ratio, which hasn’t been measured on this reflector. Start with 5 turns and adjust — 7 turns narrows the beam (better for deep dishes with short focal length), fewer turns widens it.

Hydrogen is unpolarized, so the RHCP helical receives half the total power (the LHCP half is lost). A linearly polarized feed would also receive half. There is no polarization advantage either way — use whichever feed is easier to build.

Dish performance at 21 cm

The Carryout G2 reflector is roughly 4 wavelengths across at 1420 MHz. This is small by radio astronomy standards — purpose-built hydrogen line dishes are typically 2-3 meters — but it is enough to detect the line. The limiting factor is angular resolution, not sensitivity.

With a ~15 degree beamwidth, you cannot resolve individual molecular clouds or HII regions. What you can resolve is the large-scale structure of the galactic plane: the broad emission from the inner galaxy vs. the narrower emission at the anticenter, and the velocity structure that maps spiral arms. Professional surveys from the 1950s-1960s that first mapped the Milky Way’s spiral structure used dishes with comparable angular resolution.

The elliptical aperture (84 cm x 58 cm) means the beam is not circular. The major axis of the dish produces a narrower beam (~15 degrees) and the minor axis a wider beam (~22 degrees). Orienting the dish so the major axis aligns with the galactic plane scan direction gives the best angular resolution along the scan.

Proposed procedure

Mount the 1420 MHz feed at the dish focal point. Remove the stock Ku-band LNB and mount a 5-7 turn RHCP helical antenna in its place. The feed should be centered on the reflector’s focal point — start at the same position as the stock LNB and adjust.
Connect the signal chain. Feed output to SAWbird H1 LNA, LNA output via coax to SDR. Power the LNA via the SDR’s bias tee (4.5V) or an external injector. Do not use the Winegard’s internal coax path without a DC block — the 12-18V LNB bias will damage the LNA and SDR.
Verify the signal chain in SDR++. Tune to 1420.405 MHz with ~3 MHz bandwidth. You should see a clean noise floor with the dish pointed away from the galactic plane. A noticeable bump at 1420.405 MHz when pointed toward the galactic plane confirms the chain is working.
Configure step-and-integrate tracking. Define a grid of AZ/EL pointings covering the galactic plane’s path across your local sky. At each grid point, the positioner holds while the SDR records for 30-60 seconds.
Record spectra. At each pointing, capture an FFT-averaged spectrum (1024 or 2048 bins across 3 MHz). Save the spectrum along with the AZ/EL coordinates and timestamp.
Step to the next grid position and repeat. The 15-degree beamwidth means grid spacing of 10-15 degrees is appropriate — finer spacing oversamples but won’t hurt.

Software pipeline

Software	Role
SDR++	Quick-look spectrum visualization, initial chain verification
GNU Radio	FFT accumulation flowgraph, IQ recording
Virgo	Dedicated hydrogen line observation and analysis tool
PICTOR	Web-based radio telescope interface (if automating)

A minimal GNU Radio flowgraph: RTL-SDR source (1420.405 MHz, 3 MHz sample rate) into an FFT sink with 1024 bins and 30-second averaging. Export the averaged spectrum as CSV for each pointing.

Virgo is purpose-built for hydrogen line observations with RTL-SDR hardware. It handles spectrum calibration, RFI flagging, and velocity conversion out of the box.

Calibration

The raw spectrum from the SDR contains the hydrogen emission on top of the receiver’s bandpass shape (which is not flat). To extract the true emission profile, you need a reference spectrum taken with the dish pointed away from any emission (high galactic latitude). Dividing the on-source spectrum by this off-source reference removes the instrumental bandpass and leaves only the astronomical signal.

This ON/OFF calibration technique is standard in radio astronomy. The procedure:

Point to a blank patch of sky at high galactic latitude. Record a reference spectrum (same integration time as your science observations).
Point to the target position on the galactic plane. Record the science spectrum.
Compute (ON - OFF) / OFF to get the fractional excess due to hydrogen emission.
Convert the frequency axis to velocity using v = c * (f_0 - f) / f_0, where f_0 = 1420.405 MHz.

Repeat the reference measurement periodically (every 30-60 minutes) to account for gain drift in the LNA and SDR.

Expected results

Pointing along the galactic plane should produce a clear emission line at 1420.405 MHz with broadening and multiple peaks corresponding to different spiral arm velocities. A full galactic plane survey (sweeping through accessible AZ/EL positions as the galactic plane transits) should show:

Strong, broad emission toward the galactic center (Sagittarius, if visible from your latitude)
Narrower emission toward the galactic anticenter (Auriga/Taurus)
Distinct velocity components at different longitudes, corresponding to spiral arm crossings
Weak or absent emission at high galactic latitudes

The Doppler velocity of detected hydrogen (v = c * delta_f / f_0) directly maps to the radial velocity of the gas along each line of sight. Plotting velocity vs. galactic longitude reproduces the galaxy’s rotation curve — a result first obtained by professional radio astronomers in the 1950s and now routinely replicated by amateurs.

What a single spectrum looks like

A single 60-second integration pointed at galactic longitude ~30 degrees (toward the inner galaxy) typically shows:

A strong peak near 0 km/s from local hydrogen in the solar neighborhood
A broader component shifted to positive velocities (receding gas from the inner galaxy, galactic rotation)
Possibly a second peak at higher positive velocities from a spiral arm crossing

The total line width spans roughly -100 to +150 km/s at this longitude. The peaks are 0.5-2 K above the baseline (before calibration, the raw power spectral density will show this as a bump of a few percent above the noise floor).

Automation with Birdcage

The step-and-integrate workflow is a natural fit for scripting. The Birdcage CLI accepts AZ/EL positioning commands, and the SDR recording can be triggered via command-line tools (SDR++ has a remote control interface, or use rx_sdr for headless capture). A complete galactic plane survey could be automated as:

Read a list of target AZ/EL positions from a file (pre-computed based on galactic plane coordinates and current sidereal time)
For each position: move the dish via birdcage move, wait for settle, start IQ recording, wait for integration period, stop recording, tag the file with coordinates
After the survey: run the calibration pipeline, convert to velocity spectra, plot the galactic longitude-velocity diagram

The galactic plane transits the local sky over several hours. A full survey of accessible longitudes takes one night of automated operation.

Open questions

Surface accuracy. At 21 cm wavelength, surface irregularities up to ~2 cm are tolerable (lambda/10 rule). The Carryout G2’s stamped metal reflector should be fine, but this is unconfirmed.
Focal length. The reflector’s exact f/D ratio hasn’t been measured. Feed placement will need empirical optimization using a known source (the Sun is a broadband calibrator, or use a strong geostationary satellite at Ku-band with the stock LNB to find the focal point first).
RFI environment. The 1420 MHz band is protected for radio astronomy, but illegal or spurious emissions near this frequency exist. The SAWbird H1’s bandpass filter helps, but local RFI surveys may be needed.
Integration time vs. pointing stability. The motors hold position well once stopped, but any drift during a 60-second integration would smear the beam. Characterize pointing stability before committing to long integrations.
Continuum vs. line emission. The 1420 MHz band also contains continuum (broadband) emission from the galactic plane. With sufficient sensitivity, the dish could map continuum brightness temperature in addition to the spectral line. This requires more careful bandpass calibration but uses the same hardware.

Community and references

The amateur hydrogen line community is one of the most active in radio astronomy. Several resources are worth reviewing before starting:

SARA (Society of Amateur Radio Astronomers) publishes tutorials and maintains a mailing list where members share results from dishes similar to or smaller than ours
Virgo documentation includes detailed build guides for helical feeds and SAWbird integration
Open Astronomy Catalog projects collect amateur H-line observations for comparison
The landmark paper by van de Hulst (1945) predicted the 21 cm line; Ewen and Purcell detected it in 1951 with a horn antenna smaller than our dish

The most rewarding aspect of this experiment is that the result — a galactic rotation curve derived from your own backyard measurements — is the same measurement that demonstrated the existence of dark matter in galaxies. The rotation curve doesn’t flatten at the expected rate based on visible matter alone. While our resolution is too coarse to make that argument rigorously, the basic shape of the curve is visible even in amateur data.

EME Monitoring (1296 MHz) Similar L-band feed and LNA chain -- a 1420 MHz helical can often cover 1296 MHz with reduced gain.

Antenna Pattern Measurement Characterize the dish beam at Ku-band first, then adapt the technique to verify your 1420 MHz feed illumination.