Wi-Fi and Network Troubleshooting
Wi-Fi complaints arrive as "the internet is slow", which is almost never what is actually wrong. The useful first move is to establish whether the problem is wireless, wired, or the internet connection — because those three have nothing in common and testing them takes about two minutes.
Narrow it down first
- Test on a wired connection. If wired is slow too, it is not Wi-Fi.
- Test to a local device, not the internet. Separates the LAN from the WAN link.
- Establish scope — one device, one area, or everyone.
- Establish timing — constant, or at particular times of day.
Start here
| Symptom | Most likely cause |
|---|---|
| Full signal but slow | Too many devices per access point, or interference |
| Slow only at certain times | Congestion at peak, or scheduled backups saturating the uplink |
| Drops while walking the site | Roaming gap, or inconsistent transmit power |
| Dead spots in aisles | Racking blocking signal — surveyed empty, deployed full |
| One device cannot connect | Band or security mismatch on that client |
| Everything slow, wired included | Uplink saturation or the internet connection |
| Access point offline | PoE budget exceeded, or the switch port |
| New access point underperforms | Insufficient PoE — radios disabled quietly |
| Guest users reach internal systems | VLAN or native VLAN misconfiguration |
| Intermittent on one port only | Cable, connector, or exceeded run length |
| Whole network drops briefly, repeatedly | Loop, spanning tree event, or a failing uplink |
| 2.4 GHz poor, 5 GHz fine | Normal — 2.4 GHz is congested nearly everywhere |
Full bars but slow
This is the most common and most misdiagnosed complaint. Signal strength shows the device can hear the access point. It says nothing about whether there is airtime available to use.
- Count clients on that access point Management software will show it. An access point carrying far more devices than its neighbours is the answer, and the fix is more access points rather than a faster one.
- Check which band the device is on A laptop that landed on 2.4 GHz in a congested area will be slow while 5 GHz sits idle. Band steering helps; some clients still choose badly.
- Look for interference and neighbouring networks Especially on 2.4 GHz. A spectrum view in the management software shows what else is using the channel.
- Check for slow legacy clients One old device connected at a very low rate consumes disproportionate airtime and slows everyone on that radio. Minimum data rate settings help.
- Check the uplink An access switch with a saturated gigabit uplink limits every device behind it regardless of Wi-Fi quality.
- Test to a local server, then to the internet Fast locally and slow externally means the problem is the WAN link, not the network you were about to redesign.
The instinct when Wi-Fi is poor is to increase transmit power. In a multi-access-point deployment this enlarges each cell so they overlap more, devices cling to distant access points instead of roaming, and channel reuse degrades. Dense networks generally run at lower power with more access points. If coverage is genuinely short, add an access point rather than shouting louder.
Devices dropping while moving
A scanner that disconnects mid-aisle or a call that breaks up while walking is a roaming problem, and roaming is decided mostly by the client.
- Check for a coverage gap along the actual route. Walk it with a survey tool rather than standing still.
- Check transmit power consistency. One access point running hotter than its neighbours creates a cell devices will not leave until the signal is unusable.
- Confirm the SSID and security settings match everywhere. Any difference forces a full reconnect instead of a roam.
- Check the client's roaming behaviour. Devices decide when to roam, and some hold on far too long. Scanner vendors often expose roaming aggressiveness settings — this is frequently the actual fix.
- Verify the VLAN is consistent across access points, or a roam means a new address and a broken session.
- Test with the oldest device in the fleet. It defines real-world roaming performance.
Loaded racking attenuates signal dramatically. Coverage that was fine during fit-out disappears once stock is in, and the symptom is scanners failing in specific aisles while the office is perfect. Re-survey loaded and expect to add access points along aisles — see deployment planning.
One device will not connect
- Check band support. A 2.4 GHz-only device cannot see a 5 GHz-only SSID. Common with older scanners and sensors.
- Check the security mode. Older devices may not support WPA3. If they force WPA2, give them their own SSID and VLAN rather than downgrading the main network.
- Check whether it is MAC filtered or on a blocklist, if you use those.
- Confirm DHCP has addresses free on that VLAN. An exhausted scope presents exactly as "cannot connect".
- Forget and rejoin the network on the client. Stale credentials are a genuine cause, not just a ritual.
- Check certificate validity for enterprise authentication. Expired certificates take out whole classes of device at once.
Wired faults
| Symptom | What to check |
|---|---|
| Intermittent on one port | Check the port's error counters. Rising CRC errors mean a cable, connector, or run length problem. Swap the patch lead first — it is free and it is usually that. |
| Port negotiated at low speed | Usually a cable fault or a damaged pair. A gigabit link falling back to 100 megabit is a classic bad-cable signature. |
| Access point or camera will not power on | PoE budget exceeded on the switch, or the port delivers an older PoE standard than the device requires. |
| Whole network drops briefly and repeatedly | Likely a loop causing spanning tree recalculation. Look for a cable patched back into the same switch, which happens more often than anyone admits. |
| Guest reaching internal systems | VLAN assignment or a native VLAN mismatch on a trunk. Check both ends of every trunk. |
| Everything slow at a particular time | Scheduled backups or updates saturating the uplink. Check the schedule before redesigning anything. |
Networks that worked yesterday and not today usually had something change — a firmware update, a new device, a contractor plugging something in, a certificate expiring, or a DHCP scope filling up. Ask what changed before starting a deep investigation. The answer is often immediate and saves hours.
Tools worth having
- Your management platform's client and RF views. Client counts, data rates, channel utilisation, and interference — the fastest route to a real diagnosis.
- A Wi-Fi survey app on a laptop or phone for walking a site and spotting coverage gaps.
- Switch port error counters. The single best indicator of a cable fault, and free.
- A cable tester, for anything you suspect is a physical layer problem.
- A known-good patch lead, which resolves a surprising share of "network faults".
- A spare access point, pre-configured. Swapping it in settles whether the unit or the environment is at fault.
A file copy to a local server tells you what the LAN is doing. A speed test tells you about the internet connection and everything between. When someone reports slowness, run the local test first — it immediately splits an internal problem from a WAN one and stops you redesigning a network that is working correctly.
Official manufacturer resources
Diagnostic tools, RF views, and error counter interpretation vary by platform and firmware. Confirm against documentation for your exact hardware.
Wi-Fi that is fine in the office and useless in the warehouse?
That pattern is almost always coverage designed before racking went in. Tell us the layout and what devices are failing, and we will tell you whether it needs more access points or a configuration change — and where.
Still stuck? Talk to someone who works on this hardware daily.
Tell us the model, what you have already tried, and what you are seeing. We will tell you whether it is a setting, a consumable, or a part — and we will say so if you do not need to buy anything.
