Why do Wi-Fi calls drop when users walk between access points?

Almost always roam latency or a weak cell edge. Without fast transition the client repeats full authentication on each roam, and if the cell edge is weak or there is no viable second AP, the handset roams late. Both produce interruptions long enough to break audio.

What coverage does Wi-Fi voice need?

Voice designs are usually specified with a stronger cell-edge signal than data, healthy SNR at that edge, genuine secondary coverage from a second AP above the same threshold, controlled roaming, and end-to-end QoS. The exact figures should be set per environment and handset, then validated by measurement.

Wireless problems we solve

Voice and Real-Time Application Roaming Issues

Voice, video and real-time control tolerate far less loss, jitter and roam latency than data, so they fail on designs that look healthy.

Engineer’s questionWhy do Wi-Fi voice calls drop or degrade when users walk between access points?

1 — Direct answer

The short answer

Voice over Wi-Fi and other real-time traffic exposes problems that bulk data hides. A roam that costs 400 ms is invisible to a file transfer and fatal to a call. A loss rate of one percent is unnoticeable in email and audible in voice.

The three recurring causes are roam latency (usually full re-authentication instead of fast transition, or late roaming decisions at an oversized cell edge), insufficient secondary coverage so there is no good AP to roam to, and QoS that is not applied consistently from the air through to the wired network.

Design for voice with tighter targets: stronger cell-edge RSSI, healthy SNR at the edge, genuine secondary coverage overlap, 802.11r/k/v where clients support it, and end-to-end DSCP/WMM mapping that survives the wired path.

Likely causes, roughly in order of frequency

Roam latency from full re-authentication: Without fast transition or key caching, each roam repeats the EAP exchange — commonly hundreds of milliseconds.
Insufficient secondary coverage: Voice designs require a viable second AP at the cell edge; coverage-only designs often provide only one.
Late roaming decisions: Oversized cells and high transmit power let handsets hold a degrading AP past the point where audio breaks up.
QoS misconfiguration: WMM markings dropped or re-written at the AP, switch or WAN edge means voice competes as best-effort.
Airtime contention: High channel utilization introduces jitter that voice codecs cannot conceal.
Handset and softphone behaviour: Power-save, scan behaviour and roaming thresholds differ substantially between handsets, softphones and headsets.

2 — Symptoms

Symptoms and what they usually mean

Symptom	What it usually indicates
Audio breaks exactly when walking between areas	Roam latency or cell-edge design, not general coverage.
Calls degrade only at busy times	Airtime contention and jitter rather than roaming.
One-way audio after a roam	Often session or QoS state rather than RF.
Softphones fine, handsets poor, in the same area	Client radio and roaming-threshold differences.
Voice poor while data tests pass	Expected — data tests do not measure jitter, loss or roam duration.

3 — Do this first

Safe checks you can run yourself

These are non-disruptive checks a competent network engineer can complete with the controller or cloud dashboard already in place. Do them before changing configuration — they frequently identify the cause, and they always make an external investigation faster and cheaper.

1
Measure the actual roam interruption
A continuous ping or an RTP-aware test walked along the route quantifies the gap that users experience.
2
Verify fast transition end to end
Confirm 802.11r is enabled, advertised, and actually used by the handset model deployed.
3
Check QoS marking at each hop
AP, switch access port, uplink and WAN — markings are frequently reset somewhere in the path.
4
Review cell-edge RSSI and secondary coverage targets
Voice typically needs a stronger edge and a genuine second AP above the same threshold.
5
Check channel utilization on the affected radios
Jitter from contention is a distinct cause from roaming.
6
Test with the exact handset model in use
Behaviour is device-specific; generic client testing will mislead you.

4 — Watch out

Common mistakes and misleading indicators

“Data speed tests are fine, so Wi-Fi is fine”
Throughput tests do not measure jitter, loss or roam duration, which are what voice depends on.
“Coverage meets −67 dBm, so voice is supported”
A general coverage target is not a voice target; voice needs secondary coverage and a controlled roam as well.
“QoS is enabled on the WLAN”
Enabled at the SSID means nothing if markings are lost on the switch or WAN path.
“The handsets are faulty”
Sometimes true, but almost always worth measuring the roam and the RF edge first.

5 — Escalation point

When normal troubleshooting is no longer enough

Clinical, public-safety, retail or operational voice is affected and outages have business impact.
You need packet-level evidence of roam timing and QoS treatment.
The WLAN was designed for data coverage and now must support real-time traffic.
Voice problems persist after enabling fast transition and adjusting power.
Multiple sites need a repeatable voice-grade design standard.

6 — Professional investigation

What a professional wireless investigation should measure

Voice-grade survey: Measured primary and secondary coverage, cell-edge SNR and overlap against voice design targets along real walking routes.
Roam measurement: Roam duration, loss and jitter captured across boundaries with the actual handset model.
Packet capture: Over-the-air capture of the roam exchange, fast transition, and QoS markings in both directions.
End-to-end QoS review: WMM to DSCP mapping, trust boundaries, switch and WAN treatment of voice traffic.
Controller and RF profile review: Power ranges, data rates, 802.11k/v/r configuration, band steering and load balancing behaviour.
Capacity and airtime analysis: Ensuring contention-induced jitter is not mistaken for a roaming fault.

7 — Relevant service

Which NetRobin service applies

Validation Survey

Real-time applications need measured proof against explicit targets — coverage, overlap, roam duration, loss and jitter. A validation survey establishes whether the environment meets voice-grade criteria and documents exactly where it does not.

Explore Validation Survey

Often combined with

8 — Deliverables

What you should expect to receive

Measured coverage, overlap and cell-edge results against voice design targets
Roam duration, loss and jitter measurements along the tested routes
QoS findings across the wireless and wired path
Prioritised remediation and configuration recommendations
Documented pass/fail criteria for re-validation

Direct answers

Frequently asked questions

Why do Wi-Fi calls drop when users walk between access points?: Almost always roam latency or a weak cell edge. Without fast transition the client repeats full authentication on each roam, and if the cell edge is weak or there is no viable second AP, the handset roams late. Both produce interruptions long enough to break audio.
What coverage does Wi-Fi voice need?: Voice designs are usually specified with a stronger cell-edge signal than data, healthy SNR at that edge, genuine secondary coverage from a second AP above the same threshold, controlled roaming, and end-to-end QoS. The exact figures should be set per environment and handset, then validated by measurement.

10 — Next step

Talk to a Wireless Engineer

Tell us the handset models, the routes affected and your current roam configuration.

Talk to a Wireless Engineer All wireless problems