Link performance metrics can look abstract until you’re on-console trying to decide whether a pass is healthy, a gateway is degrading, or a customer ticket is real. Four of the most common metrics you’ll see across satellite modems, TT&C receivers, and RF monitoring dashboards are C/N0, Eb/N0, BER, and PER. This guide explains what each metric means, how they relate, and how to use them operationally to troubleshoot interference, weather fade, pointing issues, and configuration problems.
Table of contents
- What These Metrics Measure
- C/N0: Carrier-to-Noise Density
- Eb/N0: Energy per Bit to Noise Density
- BER: Bit Error Rate
- PER: Packet Error Rate
- How They Relate: Converting and Correlating
- Operational Thresholds and What “Good” Looks Like
- Common Failure Patterns and What the Metrics Look Like
- Using Metrics During LEO Passes vs Steady Links
- Logging, Alerting, and Evidence for Escalation
- C/N0, Eb/N0, BER, PER FAQ
- Glossary
What These Metrics Measure
These metrics all describe signal quality, but at different layers:
C/N0 describes how strong the carrier is relative to noise per unit bandwidth (RF layer).
Eb/N0 describes how much signal energy you have per information bit (digital link layer).
BER describes how many bits are wrong after demodulation/decoding (bit layer).
PER describes how many packets/frames fail at the packet layer (network/service layer).
Operationally: C/N0 and Eb/N0 tell you the link is getting weaker; BER and PER tell you when that weakness is actually breaking data delivery.
C/N0: Carrier-to-Noise Density
C/N0 (carrier-to-noise density) is the carrier power divided by the noise power in a 1 Hz bandwidth. It’s usually reported in dB-Hz. Because it’s normalized per Hz, it’s a useful metric for comparing signal quality across different symbol rates and bandwidths.
Operational meaning: C/N0 is often one of the earliest indicators of a problem. If your antenna pointing drifts, a radome ices up, rain begins, or interference raises the noise floor, C/N0 will typically move first.
How to use it: track C/N0 against expected pass profiles (for LEO) or daily baselines (for steady links). Sudden step-changes often imply configuration or hardware issues; slow declines often match weather, obstruction growth, or component aging.
Eb/N0: Energy per Bit to Noise Density
Eb/N0 is the energy available per information bit relative to noise density. It’s usually reported in dB. Eb/N0 is closer to the “how decodable is this signal?” question than C/N0 because it accounts for data rate and coding.
Operational meaning: Eb/N0 is a direct predictor of whether your modem can hold lock at a given modulation and coding. If Eb/N0 drops below a threshold for your waveform, you’ll see increased errors, re-sync events, or a complete loss of lock.
How to use it: map Eb/N0 to your waveform’s expected performance. When using adaptive coding and modulation (ACM), Eb/N0 helps explain why the system is stepping down to more robust modes during fades.
BER: Bit Error Rate
BER is the fraction of bits received incorrectly. Depending on the system, you may see different BER points:
Pre-FEC BER: errors before forward error correction (FEC). This is a sensitive measure of raw link health.
Post-FEC BER: errors after FEC. This reflects what actually leaks through into delivered data.
Operational meaning: BER tells you when the modem is struggling. Rising pre-FEC BER with stable post-FEC BER usually means FEC is doing its job but you’re burning margin. If post-FEC BER rises, you’re entering a failure zone where delivered data integrity is at risk.
PER: Packet Error Rate
PER (or sometimes FER, frame error rate) is the fraction of packets/frames that fail to decode or fail integrity checks. It’s closer to what users feel: retransmissions, throughput collapse, dropped sessions, or corrupted data products.
Operational meaning: PER often remains low until you cross a quality threshold, then it can rise sharply. That “cliff” effect is why links can look fine until they suddenly don’t—especially when fading and interference combine.
How to use it: treat PER as a service-impact metric. If PER rises, investigate immediately—either the link is fading past its operating point or something changed (interference, mispointing, bad configuration).
How They Relate: Converting and Correlating
These metrics are connected, but not identical. A common operational relationship is:
C/N0 → Eb/N0 → BER → PER
As signal quality drops (C/N0), the energy per bit falls (Eb/N0). As Eb/N0 approaches the waveform threshold, bit errors increase (BER). Once BER overwhelms coding and framing, packet failures rise (PER) and service impact becomes visible.
Practical note: vendors may compute these differently, and some report “effective” Eb/N0 based on internal estimators. Use trends and baselines more than absolute numbers, and validate thresholds per modem/waveform.
Operational Thresholds and What “Good” Looks Like
“Good” is specific to your modulation, coding rate, symbol rate, and implementation margin. Operationally, teams often define:
Green zone: Eb/N0 comfortably above threshold; post-FEC BER ~ zero; PER negligible; stable lock.
Yellow zone: Eb/N0 close to threshold; pre-FEC BER rising; ACM step-down events; occasional retries; reduced throughput.
Red zone: loss-of-lock events; post-FEC BER non-zero; PER spikes; dropped sessions or incomplete downlinks.
The actionable step is to document your own thresholds from commissioning data: what C/N0 and Eb/N0 look like in clean conditions, and what values correlate with real user impact on your system.
Common Failure Patterns and What the Metrics Look Like
Rain fade (Ku/Ka): C/N0 and Eb/N0 drift downward; BER rises; PER spikes near the cliff; ACM may step down repeatedly.
Interference: C/N0 drops (noise floor rises) while pointing looks normal; BER may spike intermittently; spectral monitoring often shows a new carrier
or elevated noise.
Mispointing / tracking error: C/N0 and Eb/N0 degrade with az/el correlation; LEO passes show a “flattened” peak; reacquisitions may occur near max
elevation.
Hardware degradation (LNA, cabling, filters): persistent baseline shift down in C/N0; worse across all passes; little correlation with weather.
Configuration mismatch (symbol rate, coding, polarization): sudden inability to lock or abrupt BER/PER deterioration right after a change window.
Using Metrics During LEO Passes vs Steady Links
LEO passes: metrics naturally vary as the satellite rises, peaks, and sets. You should compare against an expected “pass profile” rather than a single static threshold. Look for anomalies: lower-than-normal peak C/N0, early BER onset, or unexpected dropouts at certain azimuths (possible obstruction or interference).
Steady links (often GEO): trends matter more than shapes. Daily baselines, weather correlation, and step-change detection are powerful. Sudden drops are often configuration or equipment events; gradual declines often suggest seasonal weather patterns or component drift.
Logging, Alerting, and Evidence for Escalation
Operational maturity comes from turning these metrics into evidence:
Log continuously: C/N0, Eb/N0, pre/post-FEC BER, PER, lock state, modulation/coding mode, and power/pointing state.
Time-sync everything: consistent timestamps (ideally UTC) across modem logs, spectrum monitoring, and antenna control.
Alert on patterns: rapid C/N0 drops, rising BER trend, repeated ACM step-downs, PER spikes, and loss-of-lock frequency.
Capture before/after: baseline windows make troubleshooting and regulatory escalation far more credible.
When you need to escalate an issue (interference, chronic fades, suspect equipment), these logs are the difference between “we think” and “we can show.”
C/N0, Eb/N0, BER, PER FAQ
Why does PER jump suddenly while C/N0 looks only slightly worse?
Because many links have a threshold “cliff.” Error correction can hide deterioration until margin is exhausted, then packets begin failing rapidly.
Should I alert on BER or PER?
If you want early warning, alert on pre-FEC BER and Eb/N0 trends. If you want service-impact alerts, alert on PER and loss-of-lock. Most operations teams use both: BER for early diagnosis and PER for customer-visible impact.
What’s the difference between C/N and C/N0?
C/N is measured over a specific bandwidth; C/N0 is normalized per 1 Hz and is more comparable across different bandwidths and symbol rates.
Can I compare Eb/N0 numbers across different modem vendors?
Be careful. Different implementations estimate Eb/N0 differently. Use trends and in-system baselines, and validate thresholds with real performance data from your own waveform and hardware.
Glossary
C/N0: Carrier-to-noise density (dB-Hz), a measure of carrier strength relative to noise per Hz.
Eb/N0: Energy per bit to noise density (dB), a predictor of digital link decodability.
BER: Bit error rate, fraction of bits received incorrectly (often reported pre- and post-FEC).
PER: Packet error rate, fraction of packets/frames that fail decoding or integrity checks.
FEC: Forward error correction, coding that allows the receiver to correct errors without retransmission.
ACM: Adaptive coding and modulation, dynamic adjustment of waveform robustness based on link conditions.
Loss of lock: A modem/receiver state where the signal can no longer be tracked or decoded reliably.
Link margin: Extra performance headroom beyond the minimum required to maintain target quality.
More
