Tech Notes: The Value of a Good WiFi Survey

If you are considering deploying a new wireless network, you should be budgeting for a wireless site survey as well. Replacing your existing wireless access points with new ones will not necessarily provide the level of coverage you need. In fact, it could make your network worse than it is today. Even if you had one when your original system was designed, anytime you change the system, you should conduct both a pre-deployment design survey as well as a post-install survey.

Without a wireless site survey, you are just guessing where to place your access points. If you have an older 802.11n only network and want to deploy more modern 802.11ac or ax access points, their coverage will be much different than what you had. 

But wait… you had a someone walk around with a phone looking to see where the signal was weak. So that covers it right? Not hardly. That only showed the signal strength of the phone in those locations tested. It did not predict the performance of other devices, it did not show channel overlap, signal-to-noise-ratio, interference, throughput, and latency. The phone with a WiFi performance application is not wireless site survey. 

There are several types of surveys that can be conducted. 

Passive Survey of an Existing Network

During this type of survey, the surveyor is reading the existing network and building heat maps that show things like signal strength, signal-to-noise-ratio, channel overlap, and interference. If conducting a more advanced survey, they would also be gathering spectrum data to show sources of interference and potentially locate those sources. The results of a passive survey with a good survey tool should show fairly accurate access point locations (which can be adjusted to be more accurate). The survey tool should allow for detailed reports and heat maps. 

Active Survey of an Existing Network

If the surveyor has the right tools, they should be able to conduct an active survey while doing the passive survey. These additional tests will check for jitter, latency, and throughput to actual network devices. 

AP on a Stick Survey (APOS)

The passive and active surveys are designed for an existing environment, where you want to find the weaknesses in the current system and improve them. The APOS survey is great for determining where to place access points in a greenfield environment (no network). With this method, the surveyor carries an access point (preferably the one you intend on deploying). This AP is placed in a location the surveyor determines to be suitable. The surveyor then walks the area with their test gear gathering the passive survey data from the new AP. When they finish, they move the AP to the next location and repeat the test. This goes on until the entire facility has been covered. 

This is by far the best planning tool available. You will know exactly where every AP should be placed. You will know how to align antennas. This is also the most time consuming of the surveys.

Predictive Design Survey

The predictive survey does not look at an existing network. With this type of survey, the engineer doing the design might not visit the site. They use the advanced capabilities of tools like those previously mentioned to place access points with fairly accurate antenna patterns on a copy of the floor plans. They make adjustments to the settings of each AP such as power, height, and alignment. The software shows expected coverage. 

This type of design requires all the walls to be drawn in accurately as it relates to the RF loss they cause. For example, typical drywall over 4” wide studs can cause about 3dBm loss, whereas a concrete wall can cause 6-12dBm loss. The design needs to account for floors, ceilings, and any other obstructions in the facility such as elevators, cabling closets, crowds, wiring, plumbing, etc. It can actually take more time to prepare the survey then it does to do the actual design. There is a great risk of selecting an inaccurate wall type when doing a pure predictive design.

Hybrid Predictive with Testing RF Loss Rates

With this type of survey, the design is going to be based on a predictive design, but with the RF loss rates of the obstructions being determined by an onsite inspection and testing. With this type of survey, you gain much more accuracy over the pure predictive design. The surveyor conducts these tests according to industry accepted methods to determine the RF loss rate of a given obstacle. They inspect the facility and might gather photos for a designer to build it out properly.

Choosing the Right Survey Method

If you have an existing network and need to beef it up, go with the passive and active survey. These can be conducted throughout the entire facility or focused in the areas where users are complaining. These will paint the picture of what you have today and will allow the design engineer the chance to place test APs to see how their added coverage will improve service. 

If you are building a new network from scratch, the AP on a Stick, Predictive, or Hybrid are the best options. It might make sense to also conduct a passive survey of the area to see what your neighbors are doing so you can accurately design around them as well. The AP on a Stick survey is going to cost the most, but will produce the most accurate results. It will likely prevent having to move access points post-install. The pure predictive survey is the least accurate and will likely result in having to move some access points or adding more after the cables have been run. The hybrid is a good compromise between cost and chance for error. 

Your environment will also determine the best type of survey to conduct. In buildings where the floor plans tend to be repetitive and the walls the same throughout, the hybrid or pure predictive might be suitable. When you have a varying content / or wall type situation, the AP on a Stick is best. Also consider your mounting and cabling situation when choosing the right survey. If you have a warehouse, cold storage, or other areas that are difficult to access (such as hard ceilings or assembly/production lines), AP on a Stick reduces the risks associated with post install movement. If the content of your warehouse changes frequently, AP on a Stick will allow testing to those worst-case situations better. 

Choosing the Right Survey Vendor

The right survey vendor has experience conducting surveys and designing wireless networks. They have invested in the tools and education required. Tools like Ekahau and AirMagnet can easily end up costing $8,000 – $10,000 per surveyor (software, test antennas, laptop/tablet). My survey kit for full AP on a Stick surveys is well over $15,000 which includes Ekahau Pro, Ekahau Sidekick, iPad Pro, laptop, battery packs, cones, tripods, test access points from various vendors, RF Explorer to locate rogue devices, camera, laser measuring tool, ladders, and carts/crates to move it all. 

A good surveyor has a solid wireless background and understands RF fundamentals as well as building construction methods and materials. They know proper design fundamentals. They also have experience working with different vendors and understand that when a vendor says their AP covers x feet out, that is marketing fluff and only under the perfect conditions. 

Signs of a Bad Survey Vendor

These are just some signs that the vendor you are considering might not be a good choice…

  • They offer a formula like “Place 1 AP per X number of square feet.”
  • When you ask them what tools they use, they tell you a phone or tablet and a free or low-cost app that only reads signal strength and channel assignments. If they don’t mention Ekahau, Airmagnet, or ibWave, they are most likely using an inferior product that will not provide solid analysis. 
  • If you ask to see samples of their design work and every design shows the access points lined up in the hallway. There are only a few cases where this is appropriate. 
  • Their deliverable only shows signal strength heat maps.
  • They don’t ask probing questions like:
    • What are your intended uses for WiFi? What types of data are you accessing?
    • How many devices will be using wireless and what types are they?
    • Any areas where folks concentrate to work like conference rooms, nurses’ stations, etc?

What is the Value of a Good Wireless Survey?

If you obtain a good wireless site survey, you should be able to deploy a solid wireless network that meets your needs, whether they are just routine access to WiFi, support for VOIP or Bluetooth systems, location services, or movement of large amounts of data. You should not have to move access points much post-install. You should be able to improve the user experience by reducing drops or slow connections. 

If you are deploying complex systems like VOIP, Bluetooth, or location services, they should work as designed. For example, VOIP needs low-latency and solid hand-off between access points. Bluetooth and location services require multiple access points to be heard (3 or more usually) so they can triangulate a location. If you have complex environments like warehouses with high mounting points, your clients moving about the warehouse should have solid connections. Without a good survey, you might get the luck of the draw. 

Concluding Remarks

A good survey can save you time, aggravation, and money. It is a good investment in your infrastructure design and it will make sure you get what you need. A wireless site survey should be completed by a competent surveyor educated and experienced with wireless network design and RF principles. They should be using advance tools to do it right. A good survey could take several days to weeks depending on the size of your facility. 

Do it right. Do it once. 

The author:

Alex Wilson is the co-owner of Collabsion, a technology consulting firm that specializes in wireless networking solutions. These solutions include WiFi, point-to-point, and point-to-multipoint networks that can span great distances. Collabsion can also provide consulting services related to process improvement, project management, business continuity planning, and office support services.