Noise Figure on Active Receive Antennas at HF
Updated: January 5, 2026
Noise Figure on Active Receive Antennas at HF
Why “low NF” is mostly a meaningless spec on 600 m, 160 m, 80 m, and 40 m... and still mostly marketing on 20 m and 10 m
People argue about noise figure (NF) on active receive antennas because NF is a real thing in receiver design. The catch is that on HF your antenna is almost never looking at a quiet 290 K source. It’s looking at a sky + lightning + man-made noise source with an equivalent noise temperature that can be millions to billions of kelvin. In that regime, shaving 1–5 dB off an LNA’s NF makes essentially no difference in received SNR.
The building blocks: Johnson noise and ITU external noise
Johnson / thermal reference (kTB)
Thermal noise available from a matched source is:
N = kTB
At 290 K, the noise power spectral density is about −174 dBm/Hz. Thermal noise power in bandwidth B is:
Nth(dBm) = −174 + 10·log10(BHz)
Thermal noise power at 290 K (reference only)
| Bandwidth | Thermal noise power (dBm) |
|---|---|
| 1 Hz (per-Hz reference) | −174.0 |
| 500 Hz (typical CW filter) | −147.0 |
| 2.4 kHz (typical SSB) | −140.2 |
| 10 kHz (wider digital) | −134.0 |
External noise on HF is not 290 K
ITU-R P.372 models median external noise as an external noise figure Fa (in dB above kT0B):
Fa = c − d·log10(f) (with f in MHz and c,d depending on environment)
Note: the “galactic” model is commonly stated as neglecting ionospheric shielding. For 10 m and 20 m, the ionosphere is often transparent enough that galactic noise can be a fair reference... but in real neighborhoods, local man-made noise usually dominates anyway.
External noise by band: low bands plus 20 m and 10 m
What’s being computed here:
- Thermal noise in the bandwidth (
Nth) - External noise figure (
Fa) using the ITU median model - External noise power at the antenna terminals:
Next(dBm) = Nth(dBm) + Fa(dB)
Assumptions used for the table: Quiet site = max(Quiet-rural man-made, Galactic). Residential site = Residential man-made (dominates galactic across these HF bands in typical QTH reality).
External noise levels (median model) at the antenna terminals
| Band | f (MHz) | Quiet external Fa (dB) | Quiet noise @ 500 Hz (dBm) | Quiet noise @ 2.4 kHz (dBm) | Residential Fa (dB) | Residential noise @ 500 Hz (dBm) | Residential noise @ 2.4 kHz (dBm) |
|---|---|---|---|---|---|---|---|
| 600 m | 0.472 | 62.9 | −84.1 | −77.3 | 81.5 | −65.5 | −58.7 |
| 160 m | 1.8 | 46.3 | −100.7 | −93.9 | 65.4 | −81.6 | −74.8 |
| 80 m | 3.5 | 39.5 | −107.5 | −100.7 | 57.4 | −89.6 | −82.8 |
| 40 m | 7.0 | 32.6 | −114.4 | −107.6 | 49.1 | −97.9 | −91.1 |
| 20 m | 14.0 | 25.6 | −121.4 | −114.6 | 40.8 | −106.3 | −99.4 |
| 10 m | 28.0 | 18.7 | −128.3 | −121.5 | 32.4 | −114.6 | −107.8 |
How to read that in plain ham terms:
- On 80 m, a quiet rural site still sits around ~40 dB above kT.
- On 20 m, it’s still ~26 dB above kT.
- Even on 10 m (quiet rural), it’s still ~19 dB above kT.
So the antenna is still “feeding” the receiver a noise source far hotter than 290 K... even up at 20 m and 10 m.
How much does NF really change the noise floor on 20 m and 10 m?
This is the question NF marketing implies:
“If I improve my active antenna amp from a bad NF to a great NF, how much lower will my received noise floor get?”
Using noise-temperature form (same physics as Friis, just easier to see), receiver equivalent noise temperature is:
Te = (F − 1)·T0
Now compare an intentionally extreme change:
- NF = 10 dB (awful for an LNA)
- NF = 1 dB (excellent)
Noise floor change when NF improves from 10 dB to 1 dB
Shown as the reduction in total system noise power (dB). If this number is tiny, you won’t “hear” it... and it won’t change decode thresholds in any meaningful way.
| Band | Quiet Fa (dB) | Noise floor change (quiet site) | Residential Fa (dB) | Noise floor change (residential site) |
|---|---|---|---|---|
| 600 m | 62.9 | 0.000019 dB | 81.5 | ~0.000000 dB |
| 160 m | 46.3 | 0.000890 dB | 65.4 | 0.000011 dB |
| 80 m | 39.5 | 0.0043 dB | 57.4 | 0.000069 dB |
| 40 m | 32.6 | 0.021 dB | 49.1 | 0.000468 dB |
| 20 m | 25.6 | 0.102 dB | 40.8 | 0.0032 dB |
| 10 m | 18.7 | 0.481 dB | 32.4 | 0.022 dB |
The takeaway is blunt:
- 20 m (quiet rural): a massive “10 dB NF → 1 dB NF” improvement buys you about 0.10 dB.
- 10 m (quiet rural): the same extreme improvement buys you about 0.48 dB.
- 10 m (residential): it’s about 0.02 dB... basically undetectable.
That’s why “low NF” is still mostly nonsense as a selling point on HF (including 20 m and 10 m). The improvement is tiny compared to real HF variability... and in typical man-made noise it’s essentially zero.
Reality check: This is before you even touch the real reasons active antennas disappoint on HF: overload, IMD/cross-mod, common-mode pickup (coax becomes the antenna), and mismatch effects. None of those show up in an NF headline.
A practical “max NF” target for HF
- The quietest part of HF (lowest external noise) is usually the highest-frequency end, so 10 m is the hardest case if your goal is to remain external-noise-limited across HF.
- Using the median model, quiet-site external noise on 10 m is about Fa ≈ 18.7 dB above kT0.
Single-number rule of thumb
Keep total HF receive system NF (referenced to the antenna terminals) at or below ~10 dB.
That’s “good enough” to stay external-noise-limited even on 10 m in a very quiet place, with comfortable margin. Below that, NF is simply not what’s limiting you... your wins come from the rest of the system.
Max NF vs how external-noise-limited you want to be (worst case: quiet 10 m)
This table sets receiver-added noise to be some margin below the quiet external noise at 10 m. It uses the standard NF ↔ noise temperature relationship Te = (F−1)·T0.
| Goal on quiet 10 m | Receiver noise relative to external noise | Max system NF (dB) |
|---|---|---|
| Comfortably external-noise-limited | 10 dB below | ≈ 9.3 dB |
| Still clearly external-noise-limited | 6 dB below | ≈ 12.9 dB |
| Barely matters | 3 dB below | ≈ 15.8 dB |
| Equal contribution | 0 dB (equal) | ≈ 18.8 dB |
What actually matters for active HF receive antennas
If your active antenna / LNA system NF is ≤ ~10 dB, NF is not what’s limiting you anywhere on HF. Spend your effort where it changes what you can hear:
- Linearity / overload handling: IIP3, P1dB, cross-mod behavior with real-world RF levels
- Front-end filtering: MW/BC management when needed (notch/high-pass/band-pass based on your location)
- Common-mode control: choking, grounding/layout, and not letting the coax shield become the antenna
- Antenna type + placement: loop/balanced/array choices that reduce local E-field noise pickup
Mini-FAQ
- Does NF ever matter on HF? It can, but mostly in edge cases: extremely quiet site, higher HF, low-loss antenna system, and careful overload management. On 600/160/80/40 at most real QTHs, external noise dominates.
- Is “0.9 dB NF” a lie? Not necessarily... it can be a real lab measurement into a 50 Ω source. The problem is that it usually doesn’t predict on-air SNR on HF, and it ignores overload and common-mode failure modes.
- What single spec should I care about instead? Linearity (IIP3/P1dB) and overload behavior. IMD can raise your apparent noise floor far more than any NF difference.
- What’s a practical NF target for HF systems? Around 10 dB or better referenced to the antenna terminals is generally enough to remain external-noise-limited across HF, even in quiet conditions.
Questions or experiences to share? Feel free to contact RF.Guru.