OT Wireless
Search/Menu

Search the field guide, causes, requirements, and glossary

Multipath in reflective spaces

Metal racking, tanks, enclosures, and vehicles bounce the signal back on itself, so a client can show strong signal and still lose frames, because the copies cancel each other at the spot where its antenna sits.

Ben RutterIndustrial network engineer
  • Updated
  • Reviewed
  • 3 min read
  • 7 sources
On this page5 sections
  1. How it presents
  2. Why it happens
  3. How to confirm it
  4. The fix
  5. Prevent it at design time

Is this your problem? It presents like this

  • Signal strength looks good, yet the data rate is low and retries are high.
  • Moving the client or its antenna a short distance changes link quality sharply.
  • Problems cluster near large flat metal, such as rack faces, tanks, machine enclosures, and parked vehicles.
  • A single-antenna client struggles at a spot where a multi-antenna client in the same position works.
  • At a fixed spot, rate and retries swing over time as people and vehicles move nearby.
Not sure?
Rank the likely causes from your symptoms

Why it happens

In a metal building, a strong signal is not proof of a good link. Conductive surfaces reflect most of the energy back, and NIST warns that the resulting multipath can cause significant interference and information loss even if signal power is high. Each reflection arrives with its own delay and phase. Where copies arrive out of phase they cancel, which one vendor’s technical note calls signal nulling, naming racks, shelving, and other metal among the usual causes.

Plants measure as highly reflective. NIST found mean delay spreads of 115 to 162 ns in an automotive factory, against about 6 ns at an open test site1, and average K-factors of -5 to 6 dB, which it reads as a strong multipath channel. The vendor note puts manufacturing floors higher, at 200 to 300 ns2.

The strongest objection is that OFDM was built for this. For echoes, it largely was. Wi-Fi’s legacy cyclic prefix is 0.8 µs, and 802.11ax adds 1.6 and 3.2 µs3, longer than either range above. What OFDM does not remove is fading. Channel gain still varies across the subcarriers, and where the copies cancel, rate adaptation steps the link down to a lower modulation and coding scheme, which is more robust but slower, so each frame holds the air longer.

MIMO buys margin, not immunity. Antennas spaced at least half a wavelength apart rarely fade at the same moment, and one vendor’s note says a second antenna typically sits outside the null the first one is in. A receiver with one antenna has no second copy to fall back on. Beamforming steers from channel measurements refreshed at a suggested 10 ms interval, and the reflections keep moving: a field study traced industrial fading to metal surfaces and, more significantly, to moving operators and vehicles. A vendor design guide even disables 802.11n rates for real-time EtherNet/IP, because they may decrease reliability. NIST’s remedy is geometry: a line-of-sight link between each transmitter and receiver. Signal strength tells you the energy arrived, not that it arrived in one piece.

How to confirm it

  1. At the problem spot, compare signal strength with SNR, data rate, and retries. Strong signal with a low rate and high retries points to multipath, not weak coverage.
  2. Survey the spot with the client's own antenna at its real height and mounting, moving it in small steps, and log how much the link swings.
  3. Check how many antennas the client receives on, and whether antenna diversity is enabled.
  4. Log the rate the client and access point actually use over time at a fixed spot. Frequent rate changes with nothing else changing point to fading.
  5. Take a spectrum capture to rule out interference. Multipath leaves the noise floor clean; interference raises it.
  6. Repeat the measurements with production running, including people and vehicles moving through the area.

The fix

  • Give each control client a clear line of sight to an access point by raising or moving access point or client antennas.
  • Move client antennas off flat metal and out of metal corners, or mount them where the reflections work in their favor.
  • Use clients and access points that receive on more than one antenna, spaced as the maker specifies, with diversity enabled.
  • Add a second access point so each client has a second, independent path.
  • For real-time control traffic, limit the access point to robust data rates instead of letting the link chase the highest rate.

Prevent it at design time

  • Survey with the real client antenna at its real height and mounting, under production conditions.
  • Design so every control client has line of sight to at least one access point.
  • Design to SNR and data rate targets, not signal strength alone.
  • Specify multi-antenna clients for control traffic in metal-dense areas.

Related causes

Terms on this page

About this page

Built from 7 sources: 2 standards bodies and labs, 3 research papers and theses and 2 vendor documents. Researched and drafted with AI assistance, then reviewed and approved by Ben Rutter on . How pages are made

First published
Last updated
Change history (1)
  • First published
Cite this page

Plain

Ben Rutter. "Multipath in reflective spaces." OT Wireless, published October 5, 2026. https://otwireless.com/causes/multipath-in-reflective-spaces/

APA 7

Rutter, B. (2026, October 5). Multipath in reflective spaces. OT Wireless. https://otwireless.com/causes/multipath-in-reflective-spaces/

BibTeX

@misc{rutter2026multipathinreflectivespaces,
  author = {Rutter, Ben},
  title = {{Multipath in reflective spaces}},
  year = {2026},
  howpublished = {\url{https://otwireless.com/causes/multipath-in-reflective-spaces/}},
  organization = {OT Wireless},
}