Mandatory Coverage Analysis

After conducting a spectrum analysis site survey, the next step is the all-important determination of proper 802.11 RF coverage inside your facility. During the site survey interview, capacity and coverage requirements are discussed and determined before the actual site survey is performed.

In certain areas of your facility, smaller cells or co-location may be required due to a high density of users or heavy application bandwidth requirements.

Once all of the capacity and coverage needs have been determined, RF measurements must be taken to guarantee that these needs are met and to determine the proper placement and configuration of the access points and antennas.

Proper coverage analysis must be performed using some type of received signal strength measurement tool. These tools could be something as simple as the received signal strength meter in your wireless card’s client utility, or they could be a more expensive and complex site survey software package.

So how do you conduct proper coverage analysis? That question is often debated by industry professionals. Many site survey professionals have their own techniques; however, we will try to describe a basic procedure for coverage analysis.

The first mistake that many people make during the site survey is leaving the access point radio at the default full power setting. A 2.4 GHz 802.11b/g radio transmitting at 100 mW will often cause interference with other access point coverage cells simply because it is generating too much power.

Also, many legacy client cards have a maximum transmit power of 30 mW. The RF signal of a 30 mW client might not be heard at the outer edge of an access point’s 100 mW coverage cell.

A good starting point for a 2.4 GHz access point is 30 mW transmit power. After the site survey is performed, the power can be increased if needed to meet unexpected coverage needs, or it can be decreased to meet capacity needs.

The hardest part of a coverage analysis site survey is often finding where to place the first access point and determining the boundaries of the first RF cell. The procedure outlined here is generally how this is achieved and is further illustrated in Figure below:

  1. Place an access point in the corner of the building with a power setting of 30 mW.
  2. Walk diagonally away from the access point toward the center of the building until the received signal drops to -65 dBm. This is the location where you place your first access point.
  3. Temporarily mount the access point in the first location and begin walking throughout the facility to find the –65 dBm end points, also known as cell boundaries or cell edges.
  4. Depending upon the shape and size of the first coverage cell, you may want to change the power settings and/or move the initial access point. A good portion of a proper coverage analysis involves starting over and trying again.

Once the first coverage cell and boundaries have been determined, the next question is where to place the next access point. The placement of the next access point is performed using a technique that’s similar to the one you used to place the first access point:

  1. Think of the cell boundary of the first access point, where the signal is –65 dBm, as the initial starting point, similar to the way you used the corner of the building as your initial starting point. From the first access point, walk parallel to the edge of the building, and place an access point at the location where the received signal is –65 dBm, as pictured in Figure 16.4.

  1. Now walk away from this access point, parallel to the edge of the building, until the received signal drops to –65 dBm.
  1. This is the farthest point to place the access point if you do not want cell overlap.
  1. Using the distance from the previous access point and this location, the placement of this next access point should be about 15 to 20 percent (depending upon cell overlap requirements) closer to the previous access.
  1. Move to that location and temporarily mount the access point and begin walking throughout the facility to find the –65 dBm end points, or cell boundaries.
  1. Again, depending upon the shape and size of the first coverage cell, you may want to change the power settings and/or move this access point.

It is important to avoid excessive overlap because it can cause frequent roaming and performance degradation. The shape and size of the building and the attenuation caused by the various materials of walls and obstacles will require you to change the distances between access points to ensure proper cell overlap.

After finding the proper placement of the second access point and all of its cell boundaries, repeat the procedure all over again. The rest of the site survey is basically repeating this procedure over and over again, effectively daisy-chaining throughout the building until all coverage needs are determined.

The following cell edge measurements are taken during the site survey:

  • Received signal strength (dBm), also known as received signal level (RSL)
  • Noise level (dBm)
  • Signal-to-noise ratio, or SNR (dB)
  • Data rates

The received signal strength measurements that are recorded during a site survey typically depend upon the intended use of the WLAN. If the intent of the WLAN is solely coverage and not capacity, a lower received signal of –85 dBm might be used as the boundary for your overlapping cells.

If throughput and capacity are issues, using a stronger received signal of –65 dBm is recommended. The SNR is an important value because, if the background noise is too close to the received signal, data can get corrupted and retransmissions will increase.

The SNR is simply the difference in decibels between the received signal and the background noise, as pictured in Figure below.

Many vendors recommended a minimum SNR of 18 dB for data networks and a minimum of 25 dB for voice networks. Some site survey professionals prefer to use the data rate measurements as opposed to the received signal strength measurements when determining their cell boundaries.

The problem with using the data rate is that vendors have different receive signal strength indicator (RSSI) thresholds and different vendor cards will shift between data rates at different dBm levels.

Cell design can be performed using one vendor’s RSSI threshold values if the company deploying the WLAN intends to use just that one vendor’s radios.

If measurements are based on received signal levels (RSLs), then the WLAN surveyor can always go back and map different client cards and data rates without having to resurvey.

A site survey using just data rates or a proprietary signal strength measurement threshold does not allow for any flexibility between vendors.

Table below depicts the recommended minimum received signal and minimum SNR for a WLAN data network using one vendor’s highly sensitive radio card.

Data Rate Minimum Received Signal Minimum Signal to Noise Ratio
54 – 71 dBm 25 dB
36 – 73 dBm 18 dB
24 – 77 dBm 12 dB
12/11 – 82 dBm 10 dB
6/5.5 – 89 dBm 8 dB
2 – 91 dBm 6 dB
1 – 94 dBm 4 dB

Most VoWiFi manufacturers require a minimum received signal of –70 dBm, therefore overlapping cells of –65 dBm is a good idea for VoWiFi wireless networks in order to provide a buffer.

The recommended SNR ratio for a VoWiFi network is 25 dB or higher. Cell overlap of 15 to 20 percent will be needed and the separation of same channel cells should be 20 dB or greater.

Figure below depicts the recommended coverage for a VoWiFi network.

AP Placement and Configuration

As you have just read, coverage analysis also determines the proper placement of access points and power settings. When the site survey is conducted, all the cell edge measurements will be recorded and written on a copy of the floor plan of the building.

An entry with the exact location of each access point must also be recorded. Next to the entry of each access point should be the transmission power level of the AP’s radio card when the survey was conducted.

The location of all the wiring closets will also be noted on the floor plan, and care should be taken to ensure that the placement of any access point is within a 100 meter (328 feet) cable run back to the wiring closet due to CAT5 cabling distance limitations.

Another very often overlooked component in WLAN design during coverage analysis is the use of semi-directional antennas. Many deployments of WLANs only use the manufacturer’s default low gain omni-directional antenna, which typically has about 2.14 dBi of gain.

Buildings come in many shapes and sizes and often have long corridors or hallways where using an indoor semi-directional antenna would be much more advantageous. Using a uni-directional antenna in areas where there are metal racks, file cabinets, and metal lockers can be advantageous because you can cut down on reflections.

Using indoor semi-directional antennas to reduce reflections will cut down on the negative effects of multipath, namely the data corruption caused by the delay spread and inter-symbol interference (ISI).

If data corruption is reduced, so is the need for retransmissions, thus the performance of the WLAN is enhanced by the use of semi-directional antennas in the correct situations.

Figure below depicts the use of semi-directional antennas in a warehouse with long corridors and metal racks that line the corridors.

A good site survey kit should have a variety of antennas, both omni-directional and semidirectional. The best way to provide proper coverage in most buildings is to use a combination of both low gain omni-directional antennas and indoor semi-directional antennas together, as pictured in Figure below.

When a semi-directional antenna is used, recording the received signal strength, SNR, and noise level measurements is still necessary to find the coverage edges.

The coverage area should closely resemble the radiated pattern of the semi-directional antenna. Simply record the signal measurements along the directional path and the edges of the directional path where the antenna is providing coverage.

Optional Application Analysis

While spectrum analysis and coverage analysis are considered mandatory during 802.11 wireless site surveys, application analysis is considered optional. Capacity planning is an important part of the site survey interview.

Cell sizing or co-location can be planned and surveyed during the coverage analysis portion of the survey. Capacity testing using application analysis and throughput verification is not normally part of a standard site survey.

However, tools do exist that can perform application stress testing of a WLAN. These tools may be used at the tail end of a site survey, but they are more often used during the deployment stage of the WLAN network.

One company, IXIA, makes an 802.11a/b/g multistation emulation module and hardware device that can simulate multiple concurrent virtual wireless client stations. The virtual client stations can have individual security settings. Roaming performance can also be tested.

The 802.11a/b/g multistation emulator works in conjunction with another component that can emulate hundreds of protocols and generate traffic bidirectionally through the virtual client stations.

A great use of such a device could be to test the performance of a simulated wireless data network along with simulated wireless VoIP traffic.