RF noise is the quietest way for a wireless system to fail. There's no dramatic interfering signal to point at — just a steadily rising noise floor that means your receivers can no longer hear the signals they need. Range shrinks, links get flaky, and nothing obvious is "broken." Here's where that noise comes from, and how it's traced back to source.
What a rising noise floor does to you
Every receiver can only hear a signal that stands above the background noise. Lift that background — the noise floor — and you shrink the margin every link depends on. The effects creep in: a radio that used to reach across the site now drops at the far end, a sensor network gets patchy, a once-solid link starts retrying. Because there's no single culprit signal, it's easily mistaken for failing hardware when the real problem is the environment getting noisier.
Where the noise comes from
Most RF noise is incidental — a by-product of ordinary electrical equipment, not anything trying to transmit. The classic sources are anywhere power is switched hard and fast:
- Switch-mode power supplies and chargers — cheap, failing or unfiltered ones are a leading cause.
- Variable-frequency drives and motors — strong and often tied to machinery cycles.
- LED and lighting drivers — surprisingly noisy, and there are usually a lot of them.
- Welders, plasma cutters and arcing equipment — intense, broadband, sometimes intermittent.
- Power-line networking and faulty wiring — injecting noise straight onto the mains.
This kind of noise is broadband and usually strongest at lower frequencies, tailing off higher up — a signature an engineer recognises on sight.
Two paths: conducted and radiated
Noise reaches your equipment two ways, and the fix depends on which. Conducted noise travels along cables and the mains supply, hitching a ride into sensitive kit. Radiated noise leaves the offending equipment through the air and is picked up directly by antennas or cabling acting as one. Many real problems are a mix of both, which is why guessing at a cure rarely works first time.
How an RF engineer traces it
The starting point is a spectrum survey to see the noise floor itself — its level, its shape across frequency, and how it changes as equipment switches on and off. Near-field probing then narrows it to the actual leaking device, and direction finding handles anything radiating from further away. Establishing whether it's arriving conducted or radiated decides the remedy. Only then does the fix get chosen, because filtering a radiated problem (or shielding a conducted one) wastes time and money.
The fastest way to narrow it
If you safely can, switch suspect equipment off one item at a time and watch whether the problem eases. A noise source that vanishes when a specific drive, charger or lighting circuit goes off has just identified itself. Also worth a look: earthing and bonding, since poor grounding turns many minor noise sources into major ones.
See RF interference investigation →Fixing it
Once the source and path are known, the cure is usually containment or replacement: filtering on the mains or signal lines, ferrites and proper bonding, shielding, or simply swapping out an offending supply for a clean one. These are the kinds of remedies that can be designed and fabricated on the bench and fitted on site — diagnosis and fix under one roof, rather than a report that leaves you to find someone else to act on it.
When to get someone in
If range or reliability has quietly degraded with no obvious cause, if a new installation has made existing systems flaky, or if you suspect noise but can't see it, that's the point for a spectrum survey. Noise you can measure is noise you can fix.