J/S and J/N compare interference power with two different references: the wanted signal and the noise. dBm expresses an absolute power level relative to 1 mW, while dBc expresses a power relative to a specified carrier reference. None of these labels, by itself, defines a receiver's anti-jamming performance. A meaningful comparison also needs the measurement point, frequency range, time averaging and operating criterion.
The first task is therefore to identify the quantity behind the number. Only then can two reports be converted, compared or used to interpret a GNSS evaluation.
Start with power and its reference
For power quantities, a ratio in decibels is 10 log₁₀(P₁ / P₂). The ratio is dimensionless; its meaning comes from the two powers being compared. NIST's guide to logarithmic quantities explains both the power-ratio definition and the role of a common reference when subtracting levels.
Absolute power in dBm uses a fixed reference:
P[dBm] = 10 log₁₀(P / 1 mW)
Keysight's carrier-power documentation distinguishes average power in dBm from power spectral density and specifies the measurement bandwidth. A negative dBm value is a small positive power, not a negative physical power.
| Reported quantity | What it compares or expresses | What must be identified |
|---|---|---|
| dBm | Power relative to 1 mW | Measurement point, integrated frequency range and time statistic |
| J/S, in dB | Interference power divided by wanted-signal power | Which interference components and which wanted signal define J and S |
| J/N, in dB | Interference power divided by noise power | Noise reference, bandwidth and whether N excludes the interference |
| dBc | Power relative to the stated carrier reference | The reference carrier or signal, its power definition and the measured component |
For GNSS, “wanted-signal power” should be unambiguous: for example, the defined power of one specified satellite signal rather than an unspecified sum across a constellation. The presence of the word “carrier” in dBc does not establish that the reference is a single unmodulated spectral line. Read the document's convention.
J/S and J/N cannot be exchanged without the missing reference
When J, S and N are defined at the same physical point, using compatible power measurements:
(J/S)[dB] = Pⱼ[dBm] − Pₛ[dBm]
(J/N)[dB] = Pⱼ[dBm] − Pₙ[dBm]
Here N is a total noise power over a specified bandwidth. It is not the noise density N₀ in watts per hertz. Their difference follows directly:
(J/S)[dB] − (J/N)[dB] = Pₙ[dBm] − Pₛ[dBm]
Consider an invented power-accounting example, not a receiver specification. At one reference plane, assume J = −90 dBm, one defined wanted signal S = −130 dBm, and noise N = −111 dBm integrated over a stated 2 MHz bandwidth. J/S is 40 dB; J/N is 21 dB. They describe the same interference power with different denominators.
Now reduce only the wanted signal to −136 dBm. J/S becomes 46 dB while J/N remains 21 dB. Nothing in this arithmetic demonstrates a 6 dB improvement in the receiver. The ratio changed because its reference changed. If the report provides J/N but does not provide a compatible wanted-signal-to-noise ratio, the missing information prevents a numerical conversion to J/S.
The distinction between noise power and noise density is explained further in C/N₀, SNR and signal-quality evidence.
dBc is a relative label, not a universal anti-jamming definition
A document can express interference relative to a defined carrier reference. If that reference is exactly the S used in its J/S definition, and both measurements use the same boundaries, the numerical relative level can coincide with J/S. The label alone does not establish those conditions.
An offset-channel power, a spur level, an input interference ratio and a before/after suppression ratio are different measurements. In particular, subtracting an output interference power from an input interference power describes a change across processing or hardware; it does not automatically describe the input J/S at which a receiver meets a positioning criterion.
Keysight's spectrum-analyser documentation provides a concrete instrument convention: relative offset-channel power is expressed against an automatic or manually entered carrier reference. That convention illustrates why the reference has to accompany “dBc”; it is not a definition imposed on every GNSS product specification.
For a ResiNav model, retain the published specification's actual wording. This article does not assign a new test definition, conversion or performance guarantee to a published dBc or dB figure. The anti-jamming datasheet review guide covers the separate procurement task of comparing stated capability with test evidence.
Equal total power can produce different in-band power
Power spectral density describes how power is distributed in frequency. Integrated power describes the total over a stated frequency range. An analyser's displayed span, resolution bandwidth and channel integration bandwidth are not interchangeable. Keysight's spectrum-analyser guide distinguishes integrated channel power from density normalized to 1 Hz and describes the equivalent-noise-bandwidth correction used for noise measurements.
For a hypothetical flat spectrum, total power P spread uniformly over B hertz has density:
PSD[dBm/Hz] = P[dBm] − 10 log₁₀(B / 1 Hz)
Assume −90 dBm is uniformly distributed across either 2 MHz or 20 MHz. The respective densities are approximately −153.01 dBm/Hz and −163.01 dBm/Hz. If an ideal, unit-gain 2 MHz rectangular passband lies fully inside either occupied band, it admits −90 dBm from the first spectrum but −100 dBm from the second.
This 10 dB difference is simple integration under the stated ideal model. It is not a universal conversion between broadband interference and a continuous-wave tone, nor a prediction of tracking or positioning performance. A real front end has a frequency response rather than a perfect rectangular boundary. For the distinction between that response and a tracking-loop setting, see RF bandwidth vs tracking-loop bandwidth.
Frequency and time structure still matter after power is stated
A continuous-wave tone concentrates energy near one frequency. A swept tone changes frequency over time. Broadband interference occupies a wider instantaneous frequency range. A pulsed signal changes its power with time. These characteristics can overlap: a pulsed signal can also be narrowband or swept.
An early Stanford GNSS radio-frequency-interference validation study treated continuous-wave, broadband and pulsed interference as distinct test cases. Its receiver-specific results are not universal thresholds for modern equipment; the relevant lesson here is that waveform and test conditions belong with the power ratio.
A swept tone can be narrowband at an instant while covering a broad band during a longer observation. GNSS-SDR's input-filter documentation describes a notch implementation for instantaneously narrowband interference and explains that its frequency-update rate has to follow the interference's frequency variation. A long-term averaged spectrum therefore does not fully specify the time behaviour relevant to that implementation. This is a GNSS-SDR example, not a statement of ResiNav algorithms.
For a second invented example, assume rectangular pulses with constant on-power, zero off-power and a 10% duty cycle. Average power is 0.1 times on-power, so an average of −90 dBm corresponds to an on-power of −80 dBm. That calculation does not specify pulse duration or repetition rate and cannot establish a receiver's response. Report whether a level is on-pulse, peak or averaged, and over what interval. Strong-input behaviour is a further question addressed in receiver dynamic range, compression and clipping.
Turn a number into a reproducible comparison
A useful interference result can be read as a complete statement: a defined wanted signal and interference waveform were applied at a stated point, measured over stated frequency and time boundaries, while a documented receiver configuration met a specified criterion.
Keep the physical reference plane explicit. A generator setting, antenna-terminal level and receiver-input level are not interchangeable without the intervening gain or loss being accounted for. Preserve the distinction between one interference source and an aggregate, instantaneous and averaged levels, and an input ratio versus a measured suppression amount.
Finally, name the result being tested: acquisition, continued tracking, measurement behaviour or position output. A correctly defined power ratio makes the exposure understandable; it does not choose the pass criterion. The GNSS anti-jamming evaluation guide provides the accompanying evaluation framework.
NEXT STEP
Turn available platform information into an engineering review.
Use the Technology centre to frame the discussion, review relevant application scenarios, then send the available platform and receiver details for confirmation.