Let’s be honest: most spectrum analyser screens I see in the field are set up wrong. Engineers stare at a blur of noise, tweak a knob, and call it a day. But if you’re responsible for a satellite link—whether it’s a teleport, a flyaway, or a permanent earth station—you need to know what that squiggly line is actually telling you. I’m Henrik Lindqvist, and I’ve spent more hours than I care to count chasing carriers and hunting down interference. Here’s how I do it, without the glossy brochure talk.
This isn’t a theoretical exercise. It’s a walkthrough of the real steps you take when you’re standing in front of a rack, coffee in hand, trying to figure out why the link margin just tanked. We’ll cover setup, the scans that matter, and how to read the traces so you can spot trouble before it becomes an outage.
Know Your Transponder Before You Touch the Analyser
Don’t just power up and start peeking at the IF. That’s a recipe for chasing ghosts. First, get the transponder data from the satellite operator. You need the centre frequency, the bandwidth, the saturation flux density for your beam, and the expected downlink EIRP. If you’re on a multi-carrier transponder, find out who your neighbours are and what their allocated bandwidths look like. Write this stuff down. A healthy spectrum only makes sense when you know what’s supposed to be there.
A typical C-band or Ku-band transponder gives you 36 or 72 MHz to play with. Remember, the transponder itself is a non-linear beast. Push it too hard and you get intermodulation products and spectral regrowth that splatters into the adjacent channels. Your carrier needs to sit neatly in its slot, not hogging power and not bleeding over the edges. Without the operator’s frequency plan, you’re just guessing.
Setting Up the Analyser So It Doesn’t Lie to You
Most measurement mistakes start with the analyser’s auto settings. They’re fine for a quick look, but for satellite work they can hide the very problems you’re trying to find. Take manual control.
Centre Frequency and Span
Set the centre frequency to the middle of the transponder you’re checking. For a 36 MHz transponder, a 50 MHz span gives you the full channel plus the guard bands on each side. That’s your starting point. If you’re hunting a specific interferer, tighten the span to 10 MHz or even 5 MHz to see the carrier structure. Don’t go too wide—the noise floor climbs and swallows the low-level signals you need to see.
Resolution Bandwidth and Video Bandwidth
RBW is your ability to separate signals. A DVB-S2 carrier at 27.5 Msps occupies roughly 33 MHz. For an overview, 100 kHz RBW works well. Want to inspect the roll-off or find a narrowband spike? Drop to 10 kHz or 1 kHz. Just remember: narrower RBW means slower sweeps. For VBW, set it to three times the RBW for a smooth trace, or one-tenth the RBW if you want to average out the noise and see the signal envelope clearly.
Reference Level and Attenuation
Keep the highest signal near the top of the screen, but don’t let it clip. An LNB amplifies the whole band, so the total power hitting your analyser can be surprisingly high. Start with 20 dB of internal attenuation. If you reduce attenuation and the noise floor jumps, you’re overdriving the front end. Back off until the mixer stays linear. A clean measurement depends on this.
The First Wideband Scan: Reality Check
Connect to the LNB output. If the LNB gets power elsewhere, use a DC block. If the analyser powers it, match the voltage and 22 kHz tone to the LNB spec—usually 13 V or 18 V. Sweep from 950 MHz to 2150 MHz, the standard L-band IF range. This shows every transponder on that polarisation. Compare what you see to the operator’s frequency plan.
Look for carriers that shouldn’t be there. Rogue signals come from misconfigured uplinks, faulty equipment, or deliberate interference. Note the frequency, bandwidth, and level. A carrier that’s way too hot can saturate the transponder and squash everyone else’s gain. That’s an urgent call to the satellite operator.
Zooming In on Your Carrier
Once you’ve found your carrier, centre on it and set a span about twice the occupied bandwidth. For a 33 MHz DVB-S2 carrier, 50 MHz is good. You should see a flat top with steep shoulders. The shape tells you a lot about the uplink and the transponder’s operating point.
Occupied Bandwidth and Shoulder Attenuation
Use the analyser’s occupied bandwidth measurement at 99% power. That’s the bandwidth containing 99% of the carrier’s total power. If it’s wider than your allocation, you’re causing adjacent satellite interference. Fix it by reducing the symbol rate or improving the transmit filter roll-off.
Shoulder attenuation is the drop in power spectral density from the carrier’s in-band level to the level at the edge of your allocation. A DVB-S2 carrier with a roll-off of 0.2 should show shoulders at least 30 dB down. If they’re higher, check the transmit filter, modulator linearity, and the HPA operating point. A saturated HPA fills in the shoulders with regrowth.
Estimating C/N from the Trace
A spectrum analyser doesn’t give you Es/N0 directly—that’s a demodulator metric. But you can get a rough C/N. Put a marker on the carrier peak and a delta marker on the noise floor in a clear spot. The difference is the C/N in the analyser’s RBW. To convert to the signal bandwidth, add 10 * log10 (signal bandwidth / RBW). For example, a 15 dB delta with 100 kHz RBW on a 33 MHz carrier gives you 15 + 10 * log10(33e6 / 100e3) = 40.2 dB. It’s a rough number, but it’s a quick health check.
Hunting Interference
Interference hunting is part measurement, part detective work. The usual suspects: CW tones, adjacent carrier spill-over, and cross-polarisation leakage.
A CW tone looks like a narrow spike, often drifting a bit. It could be a bad local oscillator, a terrestrial microwave link, or a noisy power supply in your own rack. To confirm it’s coming from the satellite, terminate the LNB input with a matched load. If the spike vanishes, it’s external. If it stays, start checking your own gear.
Cross-polarisation interference is sneakier. It happens when a carrier on the opposite polarisation leaks into your chain—bad antenna alignment, radome issues, or heavy rain. Switch the LNB to the opposite polarisation and look for a carrier at the same frequency. Compare its level to your desired carrier. You want at least 25 dB of cross-polarisation discrimination in clear sky. Less than that, and your feed alignment needs work.
Testing Transponder Linearity (If You Control the Uplink)
This test needs coordination with the satellite operator. Start with low uplink power and increase it in small steps, measuring the downlink carrier level each time. Plot uplink power against downlink power. The curve starts linear, then bends as the transponder saturates. The 1 dB compression point is where the downlink power falls 1 dB below the linear trend. Operating past that wastes power and creates interference. For multi-carrier transponders, back off 4 to 7 dB below the single-carrier saturation point to keep intermodulation products in check.
Documenting What You See
A measurement without a record is just a memory. Save the trace with all settings: centre frequency, span, RBW, VBW, reference level, attenuation, date, and time. Annotate it with the satellite name, transponder, polarisation, and carrier. If your analyser can export CSV data, do it. Screenshots are handy, but raw data lets you re-plot and compare over time. That’s how you catch slow degradations in antenna pointing or LNB performance before they kill the link.
Mistakes That Even Experienced Engineers Make
One classic blunder: forgetting the LNB’s local oscillator drift. An LNB with ±1 MHz drift can shift the whole spectrum, making a perfectly centred carrier look off-frequency. Always check the LO stability, especially outdoors where temperature swings are brutal. A GPS-locked reference on the analyser helps, but it won’t fix a drifting LNB.
Another trap: measuring a multi-carrier transponder with a wide RBW. The analyser shows the envelope of all carriers combined, which can look like one fat signal. Narrow the RBW until you can resolve the individual carriers. Even on a fully loaded transponder, you should see the gaps between them if the RBW is tight enough.
And never trust a single measurement. Switch polarisations, check other transponders on the same bird, and if you can, compare your spectrum with a reference from another earth station. A lot of interference problems only become obvious when you have a clean trace to compare against.
FAQ
What’s the difference between a spectrum analyser and a satellite signal meter?
A signal meter is built for quick alignment and basic checks—total power, C/N, BER for known signals. A spectrum analyser gives you the full frequency-domain picture: carrier shape, adjacent signals, interference. For serious troubleshooting and commissioning, you need the analyser. The meter is for the installer; the analyser is for the engineer.
How do I know if my spectrum analyser is overloading?
Add 10 dB of internal attenuation. If the displayed signal level drops exactly 10 dB, the front end is linear. If it drops less, or the noise floor changes shape, you were overdriving the mixer. Keep adding attenuation until the level change matches. Also watch for spurious signals that appear or disappear when you change attenuation—those are internally generated and mean overload.
Why does my carrier look tilted on the screen?
A tilted carrier usually points to a frequency response problem. It could be a poorly equalised cable, a failing LNB, or the transponder’s own gain slope. To isolate it, inject a known flat noise source at the LNB input. If the tilt stays, the problem is in your receive chain. If it goes away, the transponder or uplink is the culprit. A slight tilt is sometimes normal and gets cleaned up by the demodulator’s adaptive equaliser.
What’s the right RBW for a DVB-S2 carrier?
Depends on what you’re looking for. For an overview and shoulders, 100 kHz to 300 kHz works. For carrier flatness, use 30 kHz to 100 kHz. For narrowband interference, drop to 1 kHz or 3 kHz. Just remember the trade-off: narrower RBW means a lower noise floor but a much slower sweep. Find the balance that shows what you need without putting you to sleep.