Floating ground in AC power has a meaning ... in RF it mostly doesn’t
and that’s where the confusion starts
People use the word ground to mean at least four different things:
- Protective earth (PE / safety ground): the green-yellow wire whose job is human safety (fault-current path + fast breaker/RCD operation).
- Neutral: a current-carrying conductor in many systems that may be bonded to earth at one point (service entrance).
- Chassis ground: metalwork bonding for safety, shielding, EMC.
- Signal/RF “ground”: the reference node / return path of a circuit at a given frequency.
Most ham “ground debates” are really arguments caused by swapping those meanings mid-sentence.
What “floating to ground” means in AC electrical work
A circuit is floating (with respect to earth) when it has no intentional galvanic connection to earth/PE (and often no intentional connection to any other grounded system either).
Transformer-isolated secondary
A mains transformer’s secondary is galvanically isolated from the primary. If you don’t bond either secondary conductor to earth, that secondary system is “floating”.
- Secondary terminals are “live” relative to each other.
- But neither terminal is defined as “0 V to earth”.
Battery-powered equipment
A battery-powered circuit is typically floating relative to earth unless you intentionally bond one node to chassis/PE.
IT (ungrounded) power systems
Some installations deliberately run an ungrounded AC distribution where neither conductor is bonded to earth, with insulation monitoring to detect faults.
Why floating can be helpful (at 50/60 Hz)
It can reduce shock risk in single-touch scenarios
If a supply is floating and you touch one conductor while standing on earth, you often don’t get a dangerous current because there’s no low-impedance return path ... in the idealized model.
(Reality is messier: stray capacitance and EMI components can still create measurable leakage.)
It can break ground loops (hum/noise)
In audio and mixed-signal systems, floating (or using differential inputs) can help avoid ground loops where tiny voltage differences drive unwanted 50/60 Hz currents in shields and references.
It can make measurements more flexible
Floating sources let you reference a measurement where you want (with the right instruments), which is why many lab setups carefully manage what is earth-referenced and what is floating.
Why floating can be harmful in AC systems
“Floating” isn’t truly floating: stray capacitance creates a path
Even without a wire to earth, there is always capacitance from the circuit to nearby metalwork, building wiring, you, etc. At 50/60 Hz it’s usually high impedance, but not infinite.
A floating system can drift to surprising potentials
Without a defined reference, a circuit can “float up” due to leakage or capacitive coupling. You may measure something like “half mains” to earth on a high-impedance meter ... and it can confuse (or scare) people.
The first-fault / second-fault reality
In intentionally ungrounded distribution, the first fault may not trip protection (continuity benefit), but it shifts safety into monitoring, and the second fault can become severe.
“Floating neutral” is a different (and dangerous) problem
In many systems, an open neutral (“neutral floats”) can drive loads into wildly incorrect voltages. That’s a fire/shock/equipment hazard.
Never “float” a mains-referenced instrument by lifting PE
Using cheater plugs to lift protective earth can put the instrument chassis at live potential and create lethal hazards.
- Use an isolation transformer for the DUT (not for the scope),
- or use differential probes / battery-powered instruments,
- and follow proper measurement practice.
Now the RF part: why “floating ground” mostly stops being a meaningful idea
At RF, the core reality is simple: current always flows in a loop. If you don’t design a return path, RF will invent one through parasitics (coax shield, chassis, mic cable, house wiring, your body ...).
Capacitive reactance: XC = 1 / (2π f C)
Take a very ordinary stray capacitance: 50 pF.
At 50 Hz it looks “open”. At HF/VHF it often looks like a usable return path.
| Frequency | Example | XC for 50 pF | What it “feels like” |
|---|---|---|---|
| 50 Hz | Mains | ≈ 63 MΩ | Basically open ... “floating” feels real |
| 7 MHz | 40 m band | ≈ 455 Ω | Not open at all ... RF can flow |
| 144 MHz | 2 m band | ≈ 22 Ω | Practically tied to surroundings at VHF |
So in RF terms, the thing you thought was “floating” is often coupled to everything else through capacitance that becomes low-ish impedance as frequency rises.
“Ground” in RF is not a magic sink ... it’s a reference and return system
It’s true that RF does not need “ground” in the safety sense to exist or radiate.
It’s also true that saying “ground has absolutely nothing to do with RF” is a trap, because RF still needs a return path and a reference ... which may or may not involve earth.
- Protective earth (PE) is a safety function.
- RF return / counterpoise / reference plane is an RF circuit function.
- Sometimes they are bonded together (often for safety and lightning reasons), but they are not the same concept.
Examples that make this obvious
A dipole doesn’t need earth
A half-wave dipole is a fairly closed RF system: currents flow out one leg and return via the other. Earth can be nearby, but it’s not “the other half” in the way people imagine.
A quarter-wave vertical needs a return path
A 1/4-wave vertical needs something that plays the role of “the other quarter-wave”. In practice that’s radials/ground plane, a counterpoise, conductive structures, and sometimes lossy earth.
(A single ground rod at HF is usually a poor substitute for a real radial system: too little surface area, too high impedance, and often too lossy.)
An end-fed “works” because the return path becomes your station
End-feds often rely on common-mode current on the coax shield and on everything connected to the rig unless you provide a deliberate counterpoise and proper choking.
That’s the practical face of the RF rule: if you don’t provide the return where you want it, RF will return where it can.
Why hams get stuck: mixing DC logic with RF impedance
At DC/low frequency it’s intuitive to say “this node is floating”, “ground is 0 V”, “current goes into ground”.
At RF:
- “0 V” is local and frequency-dependent,
- “ground wires” behave like inductors,
- return currents follow the path of lowest impedance, not lowest resistance.
That’s why long “RF ground wires” in the shack often make things worse: a long wire is a decent inductor, and at HF its reactance can be tens or hundreds of ohms.
A technically bulletproof way to say it
- Protective earth is for safety. RF needs a return path/counterpoise, which may or may not be earth.
- At RF nothing is truly floating. Parasitic capacitance and unintended conductors create a reference and return whether you planned it or not.
- “RF ground” usually means a low-impedance RF return/reference plane ... not a literal earth connection.
Practical takeaways that stop pain fast
Separate safety grounding from RF current control
- Safety: bond equipment per electrical code, use proper PE, bond metalwork, and use proper surge/lightning practices.
- RF performance: provide a deliberate counterpoise/balanced feed/choking so RF currents stay where you intend.
If you have RF in the shack, assume you have an unintended return path
Symptoms include hot mic/key, RF in audio, USB glitches, and SWR changes when you touch things.
This is rarely because “you lack a ground rod”. It’s because RF return current is flowing on feedline shields, station wiring, and chassis.
The cure is usually current control
- Use a proper current balun at balanced antennas.
- Use common-mode chokes on coax where needed (often at the feedpoint first).
- Provide a real counterpoise/radials where the antenna type needs it.
“RF ground straps” only help when they’re short and wide
A short, wide strap can be lower impedance than a long round wire at RF. But it’s still not magic ... it’s just controlling impedance in the return path.
Closing thought
In AC power systems, “floating” is a real galvanic isolation concept ... useful, but with specific hazards.
In RF, “floating” is mostly an illusion because capacitance to everything becomes a return path at high frequency, and every RF system still needs a loop.
Earth ground isn’t what makes RF work. But a return path always exists ... and if you don’t design it, RF will choose it for you.
Mini-FAQ
- Does RF need an earth ground to work? No. RF needs a loop (a return path). Sometimes that return is a radial field/counterpoise, sometimes it’s the other half of a balanced antenna, and sometimes it becomes your station wiring if you don’t control it.
- Why do I measure “half mains” on a floating device? Stray capacitance (and sometimes intentional EMI capacitors) couples AC into the floating side. A high-impedance meter reads voltage easily, even when the source collapses under load.
- Why can a long “shack ground wire” make RFI worse? At HF a long wire is inductive. It raises impedance and can encourage RF current to flow on coax shields and station wiring instead of where you intended.
- What’s the first thing to try when I get hot mic/RF in audio? Common-mode current control: add/verify a feedpoint choke, fix balance at the antenna, and provide a deliberate counterpoise (or proper radials) where the antenna type needs it.
- Are radials the same as a ground rod? No. Radials are RF conductors that form an RF return/reference plane for a vertical. A single rod is usually too small and too lossy at HF to replace a radial system.
Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.
Questions or experiences to share? Feel free to contact RF.Guru via our RF.Guru contact page.