At sites visited for a “Wi-Fi is slow” complaint, most of the problem turns out to be not in the hardware but in the RF profile: channels opened up to 80 MHz, radios set to maximum power, 2.4 GHz left on on every AP. These settings are chosen at installation as “faster, wider coverage” and do the exact opposite. Below we explain how to set channel width, the channel plan and transmit power on a running enterprise wireless network, where automatic RF management falls short, and how to verify the change by measurement.
Co-channel interference and adjacent channel interference
Wi-Fi talks by listening to the medium (CSMA/CA). If two APs on the same channel hear each other above a certain level, they have to share the air; while one transmits, the other waits. This is called co-channel interference (CCI), and it directly divides capacity. The annoying part is this: the level needed for two APs to “hear” each other is much lower than the level needed for data transmission. Even a neighboring AP too weak to connect to in your own cell eats your channel time. Weak signal, big impact.
Adjacent channel interference (ACI) comes from channels with partially overlapping frequencies and shows up as noise that corrupts transmission. Using channels 1 and 3 together at 2.4 GHz is the typical example. The distinction to keep in mind: CCI divides capacity, ACI increases error rates and retransmissions. Almost all of the settings below are about reducing these two.
The 2.4 GHz channel plan and the decision to turn off the band
At 2.4 GHz there are three non-overlapping 20 MHz channels: 1, 6 and 11. Because channels 1–13 can be used in Europe and Turkey, a 1/5/9/13 plan is sometimes recommended too; however, the spacing between channels is narrower in this plan, and partial overlap occurs, especially with older 802.11b transmissions. In enterprise network design, the safe choice is 1/6/11. A 40 MHz channel at 2.4 GHz reduces the three-channel space to two; don’t use it in a multi-AP environment.
Three channels can’t carry a dense AP layout. If 2.4 GHz stays on in all APs placed 10–15 meters apart for 5 GHz coverage, dozens of radios on the same channel hearing each other result. This is a picture you may encounter. The solution is to keep the band on in the minimum number of radios needed for coverage:
- Turn off the 2.4 GHz radio on one in every two or three APs, or (if supported) put it into monitor/WIPS mode.
- On the 2.4 GHz radios that stay on, keep the power a few dB lower than 5 GHz; 2.4 GHz propagates better.
- Mark on the map the areas with devices that support only 2.4 GHz (label printers, some IoT and handheld terminals); make the turn-off decision outside these areas.
- Moving the corporate SSID to 5/6 GHz only and leaving 2.4 GHz to the IoT SSID is also a valid option.
Choosing channel width and DFS at 5 GHz
There are many channels at 5 GHz, but not unlimited. In European countries where ETSI rules apply (verify the country code and channel list for Turkey with the AP vendor too), the commonly used ranges are: 36–48 (no DFS), 52–64 (DFS), 100–140 (DFS). If you don’t use DFS channels, you’re left with four 20 MHz channels. That’s it.
| Channel width | Without DFS | With DFS (36–64, 100–140) |
|---|---|---|
| 20 MHz | 4 channels | 19 channels |
| 40 MHz | 2 channels | 9 channels |
| 80 MHz | 1 channel | 4 channels |
| 160 MHz | None | 2 channels |
What the table says is clear: if you choose 80 MHz in an office without DFS, all APs share the same channel. A wide channel increases a single client’s peak speed; but if the number of APs that hear each other exceeds the number of usable channels, total capacity drops. The order of decisions:
- From the location of an AP, measure the number of neighboring APs heard above a certain threshold (e.g. -85 dBm).
- Compare this number with the number of channels at the width to be chosen; if the neighbor count exceeds the channel count, narrow the width.
- Consider 20 MHz in dense offices and meeting areas, 40 MHz in standard offices, and 80 MHz in structures with few APs and many walls.
Risks of DFS channels
On a DFS channel, an AP must leave the channel when it detects radar. According to the EN 301 893 standard published by ETSI, a 60-second channel availability check (CAC) is done before a channel is used; in the 5600–5650 MHz range where weather radars are found (channels 120–128), this period is 10 minutes. A channel where radar was detected can’t be used for 30 minutes. At facilities near airport, port or weather radars, frequent channel changes and client disconnections are seen. Some clients don’t support DFS channels at all, or find the AP late because they only scan passively on these channels.
Don’t turn off DFS blindly. Watch radar detection records in the event logs for a few weeks; keep channels with no radar events in the plan, and leave out the problematic sub-bands (mostly 120–128). To check older clients’ DFS support, you can use the inventory steps in the Wi-Fi generations and the client ecosystem article.
Transmit power and client power imbalance
Wi-Fi is a two-way link. Increasing AP power only strengthens the direction from the AP to the client. Phones, tablets and handheld terminals usually transmit below the upper power levels an AP can reach. Worked example: if the AP transmits at 20 dBm and the phone at 14 dBm, there’s a 6 dB difference between the two directions. If the client sees the AP at -65 dBm, the AP hears the client at roughly -71 dBm (assuming equal antenna gains and receiver sensitivity). The user sees full bars on the screen; in the upstream direction, they experience low data rates and retransmissions. Full bars, half a connection.
Rules we apply when setting power:
- Keep AP power close to the power of the weakest of the target clients; in dense offices, medium power levels for 5 GHz (e.g. between 11 and 17 dBm) are often enough.
- Don’t keep large power differences between adjacent APs; a large difference causes clients to stick to the wrong AP (a client staying connected to a distant access point, a sticky client).
- Steer clients to 5 GHz by keeping 2.4 GHz power lower than 5 GHz.
- Think of power values in terms of EIRP: when external antenna gain is added, ETSI limits can be exceeded; an AP with the correct country code limits this, but care is needed with manual settings.
When is automatic RF management (RRM) not enough?
The automatic RF management of enterprise platforms (Cisco RRM, HPE Aruba ARM/AirMatch and similar) assigns channels and power based on neighbor AP measurements. In most offices it gives a good start. The problem is that the “automatic” label comes to mean nobody looks at the settings. In these cases manual intervention or defined limits are needed:
- Floor and building transitions: the algorithm measures leakage between floors, but unexpected channel neighbor relationships can form in areas such as atriums and stairwells.
- Frequent channel changes during the day: every channel change forces clients to reconnect. Use settings that move changes outside business hours or reduce sensitivity.
- Lower and upper power limits: define minimum and maximum power in the RF profile to prevent the algorithm from raising power to the top level.
- Channel width: dynamic width selection can choose wider channels than needed in dense areas; set a fixed 20 MHz in these areas.
- Foreign networks: neighboring tenants’ APs are included in the measurements, but you can’t manage them. In this case, preserving the number of channels in your own plan becomes even more important.
Table of recommended settings by environment
The values below are a starting point. They don’t count as permanent settings until verified by measurement.
| Medium | 5 GHz width | 5 GHz power | 2,4 GHz |
|---|---|---|---|
| Dense open office | 20 MHz | Low to medium | On for selected APs |
| Standard office, with rooms | 40 MHz | Intermediate | One in every 2–3 APs |
| Meeting / training room | 20 MHz | Low | Off or separate SSID |
| Warehouse, high ceiling | 20–40 MHz | Medium–high | On, depending on clients |
| Small branch with few APs | 40–80 MHz | Intermediate | On, low power |
Disabling the lowest data rates (1, 2, 5.5 and 11 Mbit/s at 2.4 GHz) lets beacons and management frames be sent in less time and narrows the cell boundary. This setting leaves out 802.11b clients; verify in advance that there are no such devices in the inventory. More than a few settings have been rolled back because of an old handheld terminal in the warehouse.
Measurement and verification after changes
To be able to compare the change, take measurements before the change as well. Without a before, the after doesn’t prove anything. Verification looks at three sources: AP/controller statistics, client-side measurement and, if needed, a walk-through verification measurement.
To see a summary of channels, power and channel utilization on a Cisco Catalyst 9800 controller:
show ap dot11 5ghz summary
show ap dot11 24ghz summary
show ap auto-rf dot11 5ghz
On a Windows client, the connected channel, signal and receive/transmit rate:
netsh wlan show interfaces
On a Linux client, the signal level and data rates of the connected AP:
iw dev wlan0 link
Checklist:
- Has channel utilization dropped during busy hours compared with before?
- In every area, is at least one AP heard at the target signal level (e.g. -67 dBm for voice), and is no second strong AP heard on the same channel?
- On the client side, is the transmit rate close to the receive rate? A large difference is a sign of power imbalance.
- Have retry rates and client disconnect events decreased?
- Is the number of RRM channel changes during the day at a reasonable level?
As cells get smaller, clients move between APs more often; for this transition to be seamless, address roaming with 802.11k, 802.11v and 802.11r settings together. The measurements to be done before design are the subject of the site survey article. For networks where RF settings need to be redone from scratch based on measurements, work can be done under the wireless network installation and optimization service.
Frequently asked questions
Which channel width should be used for 5 GHz on an enterprise network?
In multi-AP offices, the starting point is 20 MHz or 40 MHz. 80 MHz should be considered only if DFS channels can be used, the number of APs is small and wall attenuation is high. The decision is made by comparing the number of usable channels with the number of APs that hear each other.
Is it right to turn off the 2.4 GHz band completely?
Not on all APs. More 2.4 GHz radios than needed for coverage create co-channel interference; turning some of them off or putting them into monitor mode is common practice. If there are devices that support only 2.4 GHz, the band should stay on in the areas where those devices are.
What happens when radar is detected on DFS channels?
The AP leaves that channel, and the channel can’t be used for a set period; under ETSI rules this period is 30 minutes. Before moving to a new DFS channel, a channel availability check (CAC) is done. Meanwhile, clients have to move to another AP or channel, and short interruptions can occur.
Doesn’t setting transmit power to the highest level increase coverage?
The AP’s signal is heard from farther away, but the client’s signal back to the AP doesn’t increase by the same amount. The result is large cells where the client sees a good signal but the connection stays weak in one direction, more co-channel interference and poor roaming decisions.