2.2 Reflection, Multipath, VSWR and Free-Space Path Loss

Why impairments show up as retries, not as a physics lecture

Harborline's steel pallet racks turn Amira's clean sine wave into a crowd of copies. Clients see retries, roaming flaps, and complaints that the AP is right there but the laptop is slow. CWNA-109 Domain 1.1 asks you to name reflection, refraction, diffraction, and scattering, then connect them to multipath, voltage standing wave ratio (VSWR), return loss, and free-space path loss (FSPL). Independent OpenExamPrep teaching uses Harborline's warehouse so those names become mechanisms you can design around, not a vocabulary list.

The next chapter covers antennas, EIRP, SNR, and RSSI. This section stays on how the wave is redirected, delayed, reflected back into a cable, or diluted by distance.

Four ways a wave changes direction

When RF leaves the AP, anything that is not empty air can redirect it.

Reflection

Reflection is a bounce from a surface that is smooth and large compared with the wavelength. Metal is the classic reflector: rack uprights, HVAC ducts, reinforced concrete with rebar, elevator shafts, and Harborline's insulated dock doors. Glass and still water also reflect. For a mirror-like (specular) bounce, the angle of incidence equals the angle of reflection.

A 12.5 cm 2.4 GHz wave sees a 2-meter steel panel as a huge mirror. The same panel is still huge at 6 cm (5 GHz). Reflection does not eat the energy the way absorption does; it sends it somewhere else — sometimes to a client, sometimes back into a null.

Refraction

Refraction is bending as the wave crosses a boundary between materials with different densities (different propagation speeds). Glass partitions, plastic skylights, and layered atmospheric ducts outdoors are typical stories. The wave enters the new material and changes direction; it does not merely bounce. Indoor WLAN refraction is real but usually milder than metal reflection. Coated or wet glass can also add loss on the way through, so refraction is not a free lunch.

Diffraction

Diffraction is bending around an edge or into a shadow. A knife-edge roof, a rack corner, or a concrete doorway lets some energy leak into a region that simple line-of-sight geometry said should be dark. Longer wavelengths diffract more usefully, which is one reason 2.4 GHz still talks to a scanner behind a forklift when 5 GHz does not. Diffraction is not a guarantee of a good link; the diffracted component is usually weaker than the line-of-sight ray.

Scattering

Scattering is redirection from rough or small objects — chain-link, textured masonry, foliage, hanging chains, or a stack of mixed cartons. Instead of one clean bounce, energy sprays into many weak copies. Scattering fills in some coverage and also feeds multipath.

Harborline aisle 7 is a field trip in all four: steel faces reflect, polycarbonate skylights refract, rack corners diffract, and mixed cartons scatter. Amira cannot delete physics; she can place APs so a client still has a usable first path.

Multipath and RF interference

Multipath means the receiver collects two or more copies of the same transmission that traveled different routes. Copies differ in amplitude, phase, and arrival time.

When copies are in phase, they constructively interfere and the amplitude rises. When they are about 180° out of phase, they destructively interfere and the amplitude drops — a null. Harborline pickers walking an aisle move through peaks and nulls; a laptop on a cart can fade without the AP changing power. At 5 GHz a half-wavelength extra path is only about 3 cm, so a small cart motion can swing the result.

RF interference in this objective is both:

• Self-interference from multipath (the WLAN talking to itself via extra paths)
• Other RF occupying the same frequencies (microwave ovens near 2.4 GHz, neighboring WLANs, analog cameras)

Domain 6 treats interference hunting in depth. Here you need the physical idea that extra energy on the channel, or extra delayed copies of your own energy, both degrade decoding.

Delay spread and intersymbol interference (ISI)

Delay spread is the time difference between the first significant copy and the last significant copy. In a small office it might be tens of nanoseconds. In Harborline's long metal aisle it can be larger because a bounce off a far dock door returns late.

Intersymbol interference (ISI) happens when delay spread is large compared with the symbol duration. Energy from symbol N is still arriving when symbol N+1 starts, so the receiver smears bits together. Older single-carrier PHYs suffered this badly. OFDM (Domain 2) chops data into slower symbols and uses a guard interval to absorb delay spread. You do not need 802.11ax numerology here; you need the cause: late copies overlapping the next symbol.

How MIMO can exploit multipath

Legacy thinking treated multipath as only an enemy. Multiple-input multiple-output (MIMO) radios with multiple radio chains and antennas can treat sufficiently independent fading paths as extra spatial channels. Spatial multiplexing sends different data on different paths; spatial diversity sends the same data so that a null on one antenna is less likely on another.

Harborline's metal racks, which create rich multipath, can therefore help a 2×2 or 4×4 client in the aisle even while they also create nulls for a cheap single-antenna scanner. Exam trap: multipath is always bad. The accurate statement is: uncontrolled delay spread causes ISI and nulls; MIMO can harvest independent paths when the radio is built for it. Antenna diversity and radio-chain hardware detail belong with antennas in chapter 3; this section only connects multipath to that later topic.

Voltage Standing Wave Ratio (VSWR)

Not every reflection happens in the room. Some happens on the cabling and antenna side of the radio.

WLAN radios, jumpers, and antennas are commonly 50-ohm systems. If the transmitter, cable, and antenna do not present the same impedance, part of the forward wave reflects back toward the radio. Forward and reverse waves on the same cable form a standing wave.

Voltage Standing Wave Ratio (VSWR) is the ratio of maximum to minimum voltage on that standing wave.

• VSWR 1:1 is the ideal match: no reflected power.
• VSWR 1.5:1 is often acceptable for many WLAN antennas.
• VSWR 2:1 and higher means a serious mismatch: less power radiated, more power bouncing into the power amplifier, possible damage or shutdown on sensitive radios, and extra cable heating.

VSWR is dimensionless. It is not dB. Saying VSWR of 6 dB mixes it with return loss.

Harborline's loft AP had a crushed N-connector on a jumper. The analyzer showed a poor match. Replacing the jumper dropped VSWR toward 1:1 and the aisle coverage improved without raising transmit power, because more of the intended watts finally left the antenna instead of reflecting.

Return loss

Return loss (RL) expresses the same mismatch as a positive decibel figure: how many dB down the reflected power is relative to the incident power.

Return loss (dB) = −20 log₁₀ |Γ|, where Γ (gamma) is the reflection coefficient and |Γ| = (VSWR − 1) / (VSWR + 1).

• Perfect match: Γ = 0, VSWR = 1:1, return loss → infinity (no reflected power to measure).
• Total reflection: |Γ| = 1, VSWR → infinity, return loss = 0 dB.

Higher return loss is better (reflected power is farther below forward power). Lower VSWR is better (closer to 1:1). They move in opposite verbal directions, which is a favorite trap.

Worked pair:

• VSWR 2:1 → |Γ| = 1/3 ≈ 0.333 → RL = −20 log(0.333) ≈ 9.5 dB.
• VSWR 1.5:1 → |Γ| = 0.2 → RL = −20 log(0.2) ≈ 14 dB.

If a vendor quotes only one number, convert in your head so you can compare datasheets.

Free-space path loss (FSPL)

Free-space path loss is the reduction in power density caused by spreading over a sphere as distance increases. It assumes line of sight, no walls, no absorption, no extra reflections — an idealized path.

Exam trap: FSPL exists even with no walls. Harborline's empty center aisle still loses tens of decibels to spreading. There is nothing in the way does not mean zero loss.

A standard form, with distance in kilometers and frequency in megahertz, is:

FSPL(dB) = 20 log₁₀(d_km) + 20 log₁₀(f_MHz) + 32.44

An equivalent indoor-friendly form uses meters:

FSPL(dB) = 20 log₁₀(d_m) + 20 log₁₀(f_MHz) − 27.55

Both come from Friis' (4πd/λ)² identity. Use one consistently. Because of the 20 log(d) term, doubling distance adds about 6 dB of FSPL. Doubling frequency also adds about 6 dB.

Worked indoor-scale example: 50 meters, no walls

Amira stands 50 m from an AP in the clear aisle. That is d_km = 0.05, not 50. Use 2400 MHz as a 2.4 GHz stand-in.

1. 20 log₁₀(0.05) = 20 × (−1.3010) = −26.02
2. 20 log₁₀(2400) = 20 × 3.3802 = 67.60
3. Add 32.44
4. FSPL ≈ −26.02 + 67.60 + 32.44 = 74.0 dB

Same 50 m at 5000 MHz (5 GHz stand-in):

1. 20 log₁₀(0.05) is still −26.02
2. 20 log₁₀(5000) = 20 × 3.6990 = 73.98
3. • 32.44 → FSPL ≈ 80.4 dB

5 GHz is about 6.4 dB worse at the same empty-air distance (20 log₁₀(5000/2400) ≈ 6.4 dB). That is spreading only. Add a concrete wall and you stack absorption and reflection on top.

Quick 10 m check with the meter formula at 2400 MHz:

20 log₁₀(10) + 20 log₁₀(2400) − 27.55 = 20 + 67.60 − 27.55 ≈ 60 dB

Remember about 60 dB at 10 m / 2.4 GHz and about 74 dB at 50 m / 2.4 GHz as order-of-magnitude anchors. If someone types 50 into the kilometer slot, 20 log(50) is positive and the formula explodes into a nonsense outdoor-scale number.

FSPL is not the entire link budget. Cables, antennas, walls, and receiver thresholds complete the picture in later RF-measurement sections. FSPL is the loss that remains when you have already imagined the walls away.

Harborline scenario: crushed connector plus empty-aisle math

Amira's ticket says dead zone in aisle 14. The aisle is open steel, not drywall. She measures on the order of 74 dB of spreading at 50 m on 2.4 GHz and knows a client still might work if the AP radiates well. The spectrum looks clean. VSWR on that AP's chain is ugly. Replacing the jumper fixes radiation. A second ticket at 5 GHz in the same aisle is partly FSPL: she needs another AP, not another amplifier, because about 80 dB of empty-air loss plus rack scattering is a density problem.

Impairment traps

1. Claiming FSPL is zero without walls.
2. Mixing VSWR (ratio) with return loss (dB) and forgetting that high return loss is good while high VSWR is bad.
3. Stating multipath is always harmful, ignoring MIMO.
4. Confusing absorption (heat) with reflection (bounce).
5. Using kilometers in the 32.44 formula but typing 50 instead of 0.05 for a 50-meter aisle.