Sixmile Run · bench sheet 05
Hydrogen-line first light troubleshooter
A blank or strange 1420 MHz spectrum is usually a measurement problem you can isolate. Start with what the screen shows. Change one thing at a time, and keep both the raw and reference spectra. A broad feature in the right frequency range is only a candidate until it responds to the sky.
Before diagnosing the shape
- 01 · PowerConfirm the feed-side LNA is powered. Check the bias tee, any power indicator, connectors, and whether the SDR input changes when the LNA is disconnected. A long cable before an unpowered LNA can erase the sensitivity you expected.
- 02 · PointingUse the transit planner and try the inner plane around Galactic longitude 30° to 60° when it is well above the horizon. A low target puts trees, buildings and ground into a wide beam.
- 03 · ComparisonRecord an on-plane spectrum and an off-plane reference with the same gain, centre frequency, channel width and integration time. Take them close together in time. Divide on by off, then subtract one.
Choose what you see
01 / no broad line
Nothing rises above the noise
Check the sky before buying parts. The first useful observation is a drift across the inner Galactic plane. Confirm that your pointing and time put that plane in the antenna beam, then take a reference off the plane. The l = 30° target sheet gives a specific pointing; the transit planner gives tonight's meridian crossing for your site.
If the plane was in view, check the LNA power and gain. Move the LNA to the feed if cable sits in front of it. Use the receive-chain calculator with your actual cable lengths to see how much noise that placement adds. Then lower the SDR gain and repeat: a compressed receiver can look quiet and flat. Record one to five minutes per spectrum and repeat the on/off pair. The radiometer equation assumes stable, random noise; real bandpass drift can demand longer.
Finally, verify the frequency window. The rest frequency is 1420.405751768 MHz, but a Galactic line can be displaced by motion along the sightline. The Doppler tool predicts where zero LSR velocity falls for your date and pointing. Leave room on both sides of it.
02 / narrow feature
One narrow spike
Hydrogen emission from the Galactic plane spans many frequency channels. A feature only one or two channels wide is more likely to be interference or a receiver artifact. First, retune the SDR by a small amount. A spike that follows the centre of the display is a receiver artefact; one that stays at a fixed sky frequency needs another check.
Point away from the plane without changing the receiver settings. If the spike remains, investigate local RFI: disconnect nearby USB hubs, switching supplies and LED drivers one at a time, or run from a battery for a comparison. Keep the raw spectra. Filtering away a spike before you identify it hides the evidence.
03 / broad baseline
A broad slope or hill
A rounded hill in the raw spectrum is expected. It is the filter and SDR bandpass, not by itself a detection. Divide the on-plane spectrum by an off-plane reference taken with the same settings. If the divided result slopes across the band, your reference probably no longer matches the receiver. Take a fresh pair close together, or alternate on and off pointings.
Fit a straight line or gentle curve only to channels outside the possible hydrogen feature. Do not fit through the candidate line: that subtracts the thing you are trying to measure. Ignore the rolled-off outer edges of the SDR passband. LAB-012 walks through the reduction in order.
04 / periodic baseline
Repeating waves across the band
Regularly spaced waves can be a standing wave from reflections along a cable or between a feed and dish. They are not a chain of hydrogen clouds. Compare on-plane and off-plane spectra, then alter one cable length or connector at a time. A ripple that changes with that hardware change points back to the receive path.
Check loose connectors and keep the feed-to-LNA jumper short. If the pattern changes mainly as the equipment warms, take references closer to each on-plane capture. The LNA placement paper explains why even short loss before the amplifier matters.
05 / no pointing response
The feature stays put when you move the antenna
A candidate sky line should change when the beam leaves the Galactic plane. Take matched integrations on the plane and toward high Galactic latitude, with gain and tuning untouched. If the feature is unchanged, suspect the receiver, local interference, or a reference that contains the same line. A high-latitude reference can still carry some neutral hydrogen, so choose the clearest contrast you can and repeat.
A convincing detection survives more than one on/off pair and strengthens as the plane moves through the beam. Compare its broad velocity profile with the expected range in the l = 30° target sheet for that pointing; LAB-012 explains how the profile changes with direction. Use the Doppler calculator to convert measured frequency to LSR velocity and correct for observer motion. Do not use a single pretty plot as the test. LAB-009 lays out the full first observing session.
Keep a record of the failed run
Write down UTC time, site coordinates, pointing, centre frequency, sample rate, channel width, gain, integration time, LNA power state, and how you made the reference. Save both raw spectra. Change one variable, take the pair again, and compare. That gives you a diagnosis even if the second run is still blank.
If you want us to look at a stubborn result, send the two spectra and that log to contact@aliencouncil.com. We would rather see the failed runs than a cropped image of the one that looked promising.