RF Exposure in an Attic Install: Running the Evaluation You Are Required to Do
The FCC removed the amateur power exemption in 2019. Here is the MPE math for an antenna above a bedroom, worked band by band, and where a 100 watt attic wire actually fails.
An attic dipole at 100 watts running FT8 on 10 meters, eight feet above a bedroom, exceeds the FCC general population exposure limit by roughly 60 percent. The same antenna, same power, same room, running SSB on 40 meters sits at 2 percent of the limit. Nothing changed but the band and the mode.
This article contains affiliate links. I may earn a small commission at no extra cost to you.
That spread is the whole reason this evaluation exists, and it is why the old approach of checking your power against a table does not survive contact with an indoor antenna. The attic is the strongest hand a restricted-space operator can play, as the attic antenna article argues. It is also the only install where the antenna and the people share a building.
What changed in 2019
Amateur radio used to have a categorical exemption from routine RF exposure evaluation based on power alone. Below a per-band threshold, you did nothing. Above it, you evaluated.
The FCC eliminated that in its 2019 Report and Order (ET Docket 19-226) and folded amateurs into the same general exemption criteria every other service uses. The compliance date for existing stations was May 3, 2023. Every amateur station now needs an exposure evaluation on file, at every power level, including QRP.
Two things about that are worth stating plainly because they get garbled constantly.
You do not file anything. There is no form, no fee, no submission. The requirement is that you perform the evaluation and keep the record at the station. A page of arithmetic in a binder satisfies it.
And the exemption did not disappear, it stopped being a simple power table. The surviving criteria turn on effective radiated power and separation distance rather than raw transmitter output, and they assume separation an attic install does not have. You are going to end up running the numbers regardless, so run them.
Two limits and two clocks
The rules define two exposure environments, and the difference between them is roughly a factor of five.
Controlled (occupational) applies to people who know the exposure exists and can act on that knowledge. You qualify. So, per FCC guidance, do members of your own household, provided they have actually been told. Averaging period is 6 minutes.
Uncontrolled (general population) applies to everyone else: neighbors, guests, anyone on the other side of a shared wall. Averaging period is 30 minutes.
For an HF antenna, the maximum permissible exposure in power density is:
| Environment | 1.34 to 30 MHz | 30 to 300 MHz | Averaging |
|---|---|---|---|
| Controlled | 900 / f² mW/cm² | 1.0 mW/cm² | 6 minutes |
| Uncontrolled | 180 / f² mW/cm² | 0.2 mW/cm² | 30 minutes |
f is in MHz. The inverse-square term is why the low bands are nearly free and 10 meters is tight: the limit at 3.5 MHz is 14.7 mW/cm², and at 28 MHz it is 0.23 mW/cm². Sixty-four times stricter across the HF spectrum.
For an attic install, the honest question is which environment applies to the room below. If the antenna is over your own shack, controlled. If it is over a child's bedroom, a guest room, or a shared wall with the neighboring condo, use the uncontrolled limit and stop arguing with yourself about it. The margin you gain by claiming the controlled limit is exactly the margin you would want if you were wrong.
The equation
The FCC expects a far-field estimate, from OET Bulletin 65 and its Supplement B:
S = 0.00796 × P × G × F / R²
S is power density in mW/cm², P is transmitter output in watts, G is antenna gain as a numeric ratio (not dB), R is distance in meters, and F is the ground reflection factor.
Two notes on the terms that people get wrong.
G is a ratio. A dipole's 2.15 dBi is 1.64, not 2.15. Using the dB number directly is the single most common arithmetic error in these calculations and it understates your exposure.
F is 4. Supplement B applies a worst-case reflection assumption that doubles field strength, which quadruples power density. In an attic you have a roof deck above and a floor below, both of them reflecting, so this is not a conservative padding factor you get to argue away. It is closer to a description of the geometry.
The equation is a far-field formula being applied at near-field distances, which is technically the wrong tool. In most attic geometries it overestimates, which is the direction you want an error to run, and it is what the Commission's own guidance directs you to use. Document the conservative number.
Duty cycle is where the real numbers live
Peak power density almost never matters, because both limits are time-averaged. What matters is the fraction of the averaging window you spend with a carrier up, and that is two multiplied factors: how much of a transmission is key-down, and how much of the window is transmission.
| Mode | Key-down within a transmission | Share of window transmitting | Combined factor |
|---|---|---|---|
| SSB, conversational | ~20% | ~50% | 0.10 |
| SSB, heavy processing | ~40% | ~50% | 0.20 |
| CW | ~40% | ~50% | 0.20 |
| FM / RTTY | 100% | ~50% | 0.50 |
| FT8, alternating slots | 100% | 12.64s of every 30s | 0.42 |
FT8 is the entry that surprises people. It is a constant-envelope mode with no gaps, transmitting 12.64 seconds out of every 15-second slot, and in a normal exchange you hold every other slot. The digital modes article makes the case that FT8 buys back the dB a compromise antenna gives away, and that case still holds. But it does so at four times the average exposure of SSB at identical power. If you moved indoors and moved to FT8 at the same time, which describes a lot of operators on this site, your exposure went up by a factor of four while your S-meter reading stayed the same.
A worked attic install
The setup: an end-fed half-wave along the rafters, 100 watts, gain taken as 2.15 dBi, reflection factor 4, and a person eight feet (2.44 m) directly below in the room underneath. Peak power density comes out at 0.88 mW/cm². Against the uncontrolled limit:
| Band | Uncontrolled limit | SSB (0.10) | CW (0.20) | FT8 (0.42) |
|---|---|---|---|---|
| 40m (7.1 MHz) | 3.57 mW/cm² | 2% | 5% | 10% |
| 20m (14.2 MHz) | 0.893 mW/cm² | 10% | 20% | 41% |
| 15m (21.1 MHz) | 0.404 mW/cm² | 22% | 43% | 91% |
| 10m (28.1 MHz) | 0.228 mW/cm² | 38% | 77% | 162% |
Percentages are of the limit, so anything at or above 100% is non-compliant.
The shape of that table is the finding. Every SSB and CW entry passes. The low bands pass by so much that the calculation is a formality. And then 10 meter FT8 fails outright, with 15 meter FT8 sitting at 91 percent, which is not a pass so much as a rounding error away from failure. Both of those are exactly the operating that a small-space station gravitates toward, because 10 and 15 meters are the bands where a short attic wire is closest to full size and where low height costs least.
Three ways out, and only the first two are worth much.
Cut power. Exposure scales linearly with it. That 162 percent becomes 81 percent at 50 watts and 41 percent at 25 watts. This is the same conclusion the indoor QRP article reaches from the RFI and efficiency direction, arriving from a third direction: low power solves the exposure math, the neighbor-interference problem, and the breaker problem at once.
Add distance. The term is squared, so it moves fast. The same 10 meter FT8 case complies at 10.2 feet instead of 8. Moving the wire to the far end of the attic, over a hallway or a garage instead of over a bed, is often free.
Reclassify the space. If the room below is genuinely yours and only yours, the controlled limit is five times higher and everything in that table passes. Be careful here. The 6-minute averaging window is less forgiving of a sustained FT8 run than the 30-minute one, and "nobody is ever in that room" is a claim that ages badly.
The feedline nobody calculates
The evaluation above assumes the antenna radiates and the coax does not. Common-mode current on the shield breaks that assumption, and it breaks it in the worst possible geometry: the feedline runs down through occupied rooms, at arm's length, unlike the antenna.
A radiating feedline is not in your calculation at all. It is a second antenna, closer to people, that you did not evaluate.
The fix is the same 1:1 current choke that solves hot microphones and USB dropouts, and the reasoning is laid out fully in the RF grounding and counterpoise article. Choke at the feed point, choke again where the coax enters the shack. Five to seven turns through an FT240-31 core at each end covers HF, and a clamp-on mix-31 bead is the fast version for testing whether it helps before you commit. Purpose-built current chokes are stocked by DX Engineering if you would rather buy than wind.
Your calculation is only valid if the coax is quiet. Choking it is not an exposure accessory, it is what makes the arithmetic describe your station.
What a meter will and will not tell you
The equation needs your actual PEP output, and the number on the front panel is a setting, not a measurement. An in-line directional SWR and power meter reading peak power is the one instrument that genuinely improves the input to this calculation.
What will not help is a consumer EMF meter. The inexpensive units sold for household field surveys are not calibrated for HF and produce a reading with no traceable relationship to the MPE limits. Pointing one at the ceiling is not an evaluation. The FCC asks for a calculation. Give it a calculation.
What I would document
One page, kept with the license. Antenna type and gain assumption, height above the occupied space below, transmitter output measured rather than assumed, the reflection factor, the duty cycle factor for each mode you actually run, and the resulting percentage of limit per band. Note which rooms you treated as controlled and which as uncontrolled, and why.
Then act on it. In the worked case above, that means the wire moves over the hallway rather than the bedroom, and 10 and 15 meter digital work runs at 25 watts. Neither of those costs anything worth mentioning. A compromise antenna in an attic is already trading efficiency for a place to exist, as the decision tree lays out, and dropping to 25 watts on the two bands where a short wire performs best is a smaller sacrifice than any of the trades that got you into the attic in the first place.
Redo the page when you change antennas, add an amplifier, or start running a mode you were not running before. That last one is the trigger people miss, and it is the one that moved a compliant station to 162 percent of the limit in the table above without a single watt changing hands.