Back to Contents

The Site Survey

What is a Site Survey?

A site survey is a critical element in the process of designing and implementing a wireless LAN. The purpose of the site survey is to determine the number of access points needed to support the users of the wireless LAN, and the best placement of the access points in the designated location.

The site survey process consists of setting up an access point at a particular location, then using a computer equipped with a wireless LAN device and site survey software to measure the strength of the signal signal from the access point. By moving the access point around and repeating the measurements, you can determine the optimum number and best locations for the access points in the area you wish to cover. You can also identify dead zones and areas where building materials or other environmental factors affect the performance of the network.

Network requirements will vary for 802.11a and 802.11b networks. A location implementing both 802.11a and 802.11b networks in the same coverage area will also have its own special requirements and guidelines.

If you have a very large or unusually complex site, it might be advisable to contract with a professional service to perform the site survey.

When to Perform a Site Survey

Intel recommends that you perform a site survey prior to installing a wireless LAN. Site surveys are especially important at these three times:

  • Before installing a new site - Evaluate the placement of the access points and antennas throughout the proposed site.
  • Before changing an existing site - When modifying or extending an existing network structure, re-evaluate the placement of the access points and antennas. If you need a different level of coverage in some areas, you may need to move, replace, or supplement access points and antennas.
  • After physical changes to the site - Remodeling may introduce new sources of interference, such as motors and metal structures within the coverage area of the access point, even if it does not directly effect the sites where the access points are located.

Elements of an Effective Site Survey

An effective site survey requires four elements. Failure to commit the appropriate time, money, and energy to accomplish a proper site survey in advance may result in greater expenditures of money and time later, when problems arise that require repeated adjustments to the wireless configuration. The four elements of an effective site survey are:

  1. Examine the network usage problems the wireless LAN solves.
    How many clients need a wireless LAN connection? What areas of the site require wireless LAN connectivity? How many hours each day is wireless LAN connectivity required? Which locations are likely to generate the largest amount of data traffic? Where is future network expansion most likely?
  2. Study blueprints of the proposed wireless LAN site.
    A site blueprint provides a map of the site as well as the location of objects such as walls, partitions, and anything else that could effect the performance of a wireless LAN. Examining the site blueprint prior to conducting the physical walk-through helps you identify areas where wireless equipment is likely to perform well and areas where it is not. Many obstructions are not readily visible and, in some cases, a room originally built for a specific purpose, such as a radiology lab, might have been converted into something completely different, such as a conference room. The blueprint may also show areas proposed for future building expansion.
    To prepare for the next step of the site survey, mark possible wireless device locations on the blueprint and refer to the marked blueprint during the physical walk-through and inventory.
  3. Conduct a physical walk-through and inventory.
    The primary purpose of the physical walk-through is to document any items or materials near a proposed device location that may interfere with reception or transmission and effect network performance. Document stock and inventory levels, current environmental conditions and any materials that may interfere with wireless LAN transmissions.
  4. Measure the radio frequency (RF) transmissions.
    Intel(R) PRO/Wireless LAN Site Survey provides the means to measure the signal strength, the transmission rate and the number of successfully transmitted packets. This step in the site survey involves mounting access points and antennas in the proposed locations and measuring the RF signal strength, transmission rate and packet throughput.

Network Usage Requirements

When designing a site, consider all of the possible uses for the wireless network. Important factors include mobility, performance, security, redundancy, and system interface requirements.

  • Mobility Requirements — Determine whether the customers are in motion continually, such as in a warehouse or hospital, or whether the customers work from different fixed locations throughout the site.
  • Performance Requirements — Consider whether the people using wireless devices need higher data rates for time-bound transmissions, such as audio and video, or whether the system most often handles burst or typical knowledge worker data traffic. If computers require higher data rates, put access points closer together, because the higher transmission rates can only be obtained near the access points.
  • The number of computers per access point varies with the customer applications and other system factors. Typically, an access point readily accommodates 10 to 30 computers. Areas with a high density of computers, such as conference rooms, may require additional access points.
  • Security Requirements — Because wireless LANs use unsecured radio waves to transmit data, security is an important consideration for network administrators designing wireless networks. Network administrators should consider whether encryption such as Wired Equivalent Privacy (WEP) is necessary to protect sensitive data transmitted over wireless networks or if the customer can afford a lower degree of security.
  • Redundancy Requirements — It is recommended that you design your wireless system so that each access point has redundancy from another access point. This provides not only for redundancy, but improved overall system performance.
  • System Interface Requirements — Roaming from one access point located on one IP subnet to an access point located on another IP subnet causes loss of IP connectivity, unless special arrangements are made.

What About Range?

Range is one of the most complicated of microwave subjects. Learned professors have written long technical articles with complex equations attempting (with only marginal success) to determine the range of a microwave signal in a given environment. A few facts are clear:

  • Range is inversely proportional to data rate. The faster the data, the shorter the range. This has to do with the modulation technology used. Very fast data rates require extremely complex signal waveforms, where even minor distortions can result in data errors. Slower data rates are much more tolerant, and consequently will get through even in the presence of some amount of noise, interference, distortion and echo. This is a normal condition.
  • Range is highly dependent on the physical environment. In a line-of-sight location, with elevated and calibrated antennas, range predictions are quite accurate. This is not true in a typical office building, where the walls may be simple drywall (which is almost transparent to microwaves), or could be plaster with metal lath (which is mostly microwave opaque.) Most sites are somewhere between these two extremes, and consist of a mixture of surfaces. You can’t tell what is inside a wall by just looking at it, and we can’t tell you exactly what distance you will achieve. Consider published range information to be typical, average, common or usual. Do not expect it to be exact!
  • Range also depends on the electronic environment. If other equipment that could cause interference is nearby, the range of your transceiver could vary widely, and could change suddenly when the other equipment activates. This is particularly true for 802.11b installations, which share their frequencies with microwave ovens, cordless phones, wireless hi-fi speakers, electronics toys and similar devices.
  • If your range seems unexpectedly short, repeat some tests late in the evening, or on a weekend, when there may be less interference. However, some users tend to leave their nets turned on 24/7 so this test is not foolproof. By all means, try more than one channel. Your range problem may just be a nearby user whose system uses your present test channel. 
  • Try to keep your system away from other transmitters, and from other sources of electrical noise, such as large motors, spot welders, and similar electronically noisy devices.
  • Do not mount access points closes to fluorescent light fixtures! The lamp glow appears constant, but inside the lamp tube, ionization appears and disappears 120 times a second. This can modulate or chop and incoming signal, and interfere with reception.
  • Too much range can be a very bad feature. At first it would appear that you would want as much range as possible, but with the increase in range comes the increase in interference potential, as your unit hears not only your other units but also manages to hear the systems of other companies up and down the street. If you have a large installation, you will also wind up with more than one access point using the same channel. Now a remote unit will hear two or more access points, and you will wind up with a slow, clogged up, network.
  • Excess transmit range presents a special reverse problem. For example, putting an access point adjacent to a second floor bay window invites anyone parked on the street nearby (with special software) to pick up and enjoy all of your transmissions! We discuss some possible solutions to this problem further on.

Wireless LAN radio devices are extremely sensitive. They are designed to work reliably in a typical office or warehouse environment, where the receiver intercepts only a tiny portion of the transmitted signal. Nevertheless it is instructive to list some typical losses. The table below is based on 802.11b frequencies, and should be considered a rough guide only:

  • Loss through a typical office window: 38%
  • Loss through a typical brick wall: 60%&
  • Loss through a typical office wall: 75%
  • Loss through a typical metal door in an office wall: 93%

The bottom line is that the exact range of a given microwave product is almost impossible to calculate, due to the variables involved. If you own a cell phone, which uses frequencies close to microwaves, you have undoubtedly experienced the common problems due to range limitations and dead spots. Data transceivers have the same problems as cell phones, and the problems will show up at much shorter distances.

We can give you hints and suggestions, but the only way to know your range accurately is to survey your site, and measure and map the signal strengths in the various areas you want to cover.

It is only fair to note in passing that the above and following comments are not unique to Intel products. The nature of microwave signals has been studied for more than 50 years, and most of the quirks and oddities have been noted and documented long ago. Further, our products, and in fairness those of our competitors, are designed and manufactured in compliance with industry specifications that define such parameters as output power, receiver sensitivity and data rates. While we obviously concentrate on our own products, and just happen to think that they are better made and more carefully tested, you might pick up some good information here regarding competitive products, your cell phone, baby monitor, wireless speaker system or other microwave devices. Think of us when you remember to move to the other side of that all-metal truck before trying to use your cell phone!

Channel Selection Considerations

The exact number of available channels is dependent on local regulation. To avoid inter-channel interference between access points, configure your access points to local clear channels or their equivalents in other countries. Arrange your access points to avoid having same channel units adjacent to one another.

You cannot adhere to this recommendation in some situations, such as in a building with several access points on multiple, adjacent floors, where the access points transmit through many floor structures. Adjacent access points operate properly when they use the same channels. Each access point "hears" the other, and takes its turn transmitting. This situation may cause a reduction in data throughput if it occurs in a very active area, so design your channel layout accordingly.

Intel(R) PRO/Wireless 5000 LAN Adapters operating in ad hoc mode must operate on a clear channel. Ad hoc clients should NOT be configured to a channel used by an Access Point servicing the same area. 

Preparing for A Site Survey

Installing an Intel(R) PRO/Wireless LAN begins with conducting a site survey. A site survey involves using the Site Survey Utility to determine the physical requirements for a site-specific Intel(R) PRO/Wireless LAN. A site survey analyzes the installation environment and provides recommendations for equipment and placement.

Use the Site Survey Utility to determine the placement of access points and antennas, as well as the number of devices necessary to provide optimal service. The facility can be a warehouse, manufacturing plant, office building or retail store.

In the installation of an Intel(R) PRO/Wireless LAN, complete radio coverage could require optional antennas.

Inspecting the Survey Area

During the planning stages of the site survey, a network administrator or equivalent visits the proposed radio coverage site. As a standard practice in the site survey consultation, the network administrator gathers facility drawings and completes a Site Survey Requirements document and a site survey questionnaire. The network administrator documents the wiring used within the facility (Category 5 for 10 Mbps Ethernet or 10/100 switched Ethernet, fiber optic cabling, etc.) and assesses its applicability to Intel(R) PRO/Wireless LAN components.

Select several trial installation areas. The site survey team analyzes each proposed installation area to document radio transmission constraints and to develop preliminary access point placement alternatives to be tested during the actual site survey. Document the findings from the initial site inspection in a Site Survey Request Form and use that form as the outline of the site survey.

Consider the following variables in the site survey requirements definition:

  • RF systems already in use.
  • Availability of lifts for mounting access points.
  • Location of host system or systems.
  • Available AC power.
  • Interfering metal fire breaks and wall structures.
  • Availability of customer technical personnel to answer questions during the survey.
  • Doorways and passages effecting RF propagation.

The completion of the RF Site Survey Requirements document is a coordinated effort between the network administrator and the customer management team.

This document does not identify potential installation constraints within the customer site, nor does it recommend access point and antenna placement location. The document represents a preliminary overview of the customer site used as a baseline for refining site survey requirements.

Assessing Environmental Radio Coverage

The network administrator selects trial component installation areas away from transformers, heavy-duty motors, fluorescent lights, microwave ovens, refrigerators and other industrial equipment. Areas with excessive moisture, heat and dust are inappropriate for staging a wireless network.

Signal loss can occur when metal, concrete, walls or floors block access point transmission areas. Access point antennas are trial-mounted in open areas or added to an existing access point to boost the Intel(R) PRO/Wireless LAN coverage area.

The positioning of an access point depends on the floor plan of the site. The network administrator makes access point and antenna placement recommendations based on the following installation site variables:

  • Outdoor or indoor installation site.
  • Large or small proposed radio coverage area.
  • Wide or narrow proposed coverage area.
  • Number of additional devices operating in the proposed coverage area.
  • Open coverage area or area with documented obstructions.

General Guidelines

Site surveys for wireless devices require experimentation with different antenna coverages at various angles. Placing a directional antenna in a vertical position can often minimize multi-path problems.
NOTE: Do not locate access points near corners, against walls, against metal walls or inside metal conduits.

Preventing Channel Interference in a 2.4 GHz Environment

Intel(R) PRO/Wireless LAN access points require careful survey area testing to ensure radio transmissions do not overlap.

Intel(R) PRO/Wireless LAN access points supporting different coverage areas in close proximity to one another should be configured using direct sequence channels as far apart as possible, such as channels 1, 6, and 11.

Supplying Power from Site Electric Circuits

Intel(R) PRO/Wireless LAN access points draw power from wall outlets. Access point performance is subject to degradation due to inherent or random electrical problems or site-specific disturbances.

Access point electrical installation alternatives are listed below in order from most to least preferred for an installation area.

  1. Isolated ground circuit with an online, uninterruptible power supply (UPS) that also acts as a filter and surge suppressor.
  2. Isolated ground circuit with a surge suppressor.
  3. Dedicated circuit with a UPS.
  4. Dedicated circuit with a surge suppressor.
  5. Nondedicated circuit with a UPS.
  6. Nondedicated circuit with a surge suppressor.
CAUTION: When using a Network Controller Unit, items one through four are recommended. The use of nondedicated circuits for configurations five and six could cause data loss and serious transmission problems. By its nature, a nondedicated circuit contains open receptacles. The load and type of use of open receptacles in a nondedicated circuit cannot be predicted after installation.

If you must use a non-dedicated circuit, Intel recommends that the circuit not support any of the following:

  • Hard-wired devices.
  • Devices with components intended or known to produce heat such as space heaters, laser printers, heat guns and soldering irons.
  • A single device drawing more than 20% of the rated value of the circuit.
  • Devices drawing more than 60% of the rated circuit value.
NOTE: Ensure that each access point has power available 24 hours per day. It is recommended access point power never be provided from an Energy Management System (EMS).

Recommended Procedure for Performing a Site Survey

Using the Site Survey Utility to perform a site survey involves selecting possible locations for the access points from a review of the floor plan layout, based on estimated of the range of an access point within the site. Range estimates are based on previous experience, physical test and measurement or calculation from the rated hardware specifications.

Proposed locations are validated by temporarily installing one access point and using the Site Survey Utility to test the actual range and coverage area of that access point in that location. Adjust the position of the access point as required until you validate the coverage area.

After you validate the first proposed access point position, annotate it on the floor plan layout, move the access point to the next adjacent location, and test the access point coverage from that location. Repeat this procedure of testing, adjusting positions, and relocating the access point until the entire proposed coverage area and all of the proposed locations have been validated.

It is recommended that you accomplish the site survey using the continuous Internet Control Message Protocol (ICMP) "ping" test with a block size of 1024 bytes.

Helpful Hints and Tips

  • Floor to floor interaction between access points is usually not a problem, unless the floor is wood. If interaction can occur, ensure that the channel selections are appropriate for vertically adjacent access points. Floor to floor interaction can be a good thing, as it provides for redundancy.
  • Close all office and room doors before beginning the survey. This puts more material between the transmitter and receiver, allowing you to assess reception at its lowest everyday level.
  • Conference rooms may have a high density of computers, so ensure adequate access points for satisfactory throughput. Consider redundancy for these areas.
  • In high security areas where access may be limited, such as boardrooms and labs, consider placing access points adjacent or straddling the area, rather than directly in the area. When maintenance is required on the access points, access to the restricted area is not required.
  • Cafeteria areas can be a high usage and demand area, as informal employee meetings are often conducted there.
  • Computer to access point ratios are typically maintained at a maximum of approximately 20-to-1, depending on the type of applications and data.
  • Areas with multiple conference rooms in close proximity are often given additional access points to provide redundancy and ensure adequate throughput.

Creating a Trial Wireless LAN

Before using the Site Survey Utility to test component radio signal strength, you must create a wireless LAN infrastructure. To create a wireless LAN and perform a site survey, you need the following equipment in addition to the Site Survey Utility:

  • Access points
  • Digital camera
  • Distance measurement wheel
  • Computers with extra batteries
  • AC power extension cords

Installing a Trial Wireless LAN

  1. Mount the access points.
    Mount access points at the locations recommended on the floor plan drawing. Start with the most difficult coverage area first. Position the access points so the antennas are not obstructed.
  2. Power on the access points. 
  3. Observe access point LED behavior.
  4. Observe the LEDs to verify normal boot operation.
NOTE: Component installations differ depending on the device installed. Refer to the documentation shipped with each component to ensure proper installation and configuration.
  1. Power on the computers used for the site survey. When powered on, the device displays a boot server connection failure message. This is normal, because the access point is not yet connected to an Ethernet network.
NOTE: The Site Survey Utility should already be loaded on the mobile devices and desktop computer used to perform the survey.
  1. Set the access point and computer Network Name (SSID).
  2. Set the computer and access point to the same Network Name. In a 2.4 GHz environment the access point default value 101.

After you set the Network Names, you can begin testing the Intel PRO/Wireless LAN for radio transmission effectiveness with the Site Survey Utility.

Conducting a Site Survey

Site surveys differ depending on the size of the survey site.

  • Smaller sites are surveyed for one or two access points to provide coverage over a space no larger than a single room.
  • Medium sites could require between 10 and 20 access points for the size of a warehouse or several rooms.
  • Large sites could require between 20 and 100 access points for a large building with different radio coverage service areas in different parts of the building.

Discuss any special installation requirements. Determine the types of cables to be connected to the access points (Category 5 for 10 Base T 10 Mbps Ethernet or 10/100 switched Ethernet, fiber optic cabling, etc.)

Survey for a Small Area

To conduct a survey in a site where one or two access points are anticipated:

  1. Document the size and layout of coverage area. Also, document RF systems already in use, location of host system, and available AC power.
  2. Set up an access point in the middle of the room or where it is estimated that the coverage cell is the largest.
  3. Walk the perimeter of the coverage area and measure radio coverage.
    • For a wireless network, study round-trip ping times and data rates. The individual round-trip ping time is usually about 7 ms before the rate decreases when using a packet size of 1024 bytes.
    • To display the number of packets transmitted and received and the rates for each, click Test Totals on the Site Survey Tests dialog box.
    • Move the mobile computer in different directions. Position the computer between the surveyor and the access point. Do not walk fast. If you walk too fast, you may miss coverage area dead spots.
    • Depending on the coverage requirements, you may have to determine the perimeter for each of the data rates separately. For example, in a 2.4 GHz environment, total test time should be less than 13 seconds for 1 Mbit or 12 seconds for 2 Mbit.
  4. Mark the location of access points on the blueprint, move the access point to the second survey location and repeat the procedure.
    Infrastructure backbone, hubs and patch panels should all be documented. Document the conditions existing in the final installation.
  5. Complete the Site Survey Report. The report indicates the number and location of the access points. Assign the access points the same channel for each coverage area.

Survey for a Medium or Large Area

To conduct a survey in a site where 10 or more access points are anticipated:

  1. Document the size and layout of the coverage area. Also document RF systems already in use, location of host systems, available AC power, possible antenna locations, interfering metal fire breaks and wall structures, doorways and passages that could help RF propagation, and amount of stock in coverage areas. For a warehouse, document how high the stock is kept and how high the lifts go so that the lifts do not damage the access points.
  2. Set up an access point at one side of the proposed coverage area.
  3. Walk the perimeter and measure radio coverage. If there are multiple floors, measure each floor separately. For a wireless network, document round-trip times and data rates. The individual round-trip ping time is usually about 7 ms before the rate decreases when using a packet size of 1024 bytes. Move the mobile computer in different directions. Position the computer between the surveyor and the access point. Do not walk fast. If you walk too fast, you may miss radio coverage dead spots.
  4. Document the boundary of each coverage area. The data rates could be different in each area. For results in better coverage, you may only need to move the access point a meter (three feet) or so. For higher data rate networks, reflection is a bigger problem. Ceiling sprinklers can cause a problem if the access point antenna is less than 60 centimeters (two feet) away. For example in a 2.4 GHz environment, total test time should be less than 13 seconds for 1 Mbit or 12 seconds for 2 Mbit.
NOTE: For large surveys, consider the site survey boundary areas for multiple buildings. Determine whether the network requires a LAN bridge, wireless access points or both.
  1. Mark the location of each access point on the blueprint and move the access point to a different trial location. Continue the process until the entire site has radio coverage.
  2. Take a photograph of at least one access point location for each coverage area. If each coverage area has unique features, take a picture of racks, stock level and hub locations. The pictures go into the report and help establish the condition of the site as it was surveyed.
  3. Consult with customer technical personnel and document any additional hubs that could be required. Fourteen or more access points could require multiple hubs. Consider exact conditions that exist in the final installation.
  4. Complete the Site Survey Report. The report indicates the number and location of the access points. Assign the access points the same channel for each coverage area.

Back to Contents


Copyright (c) 2003 Intel Corporation.