Forum

I was on a site survey last week in which several problems presented that required various tools, hardware mounting rigs, zip ties, and the like. Without these tools, the survey would have been much more difficult, time consuming, and probably less accurate. I'd be interested to see some lists of essential and helpful equipment that most surveyors include in their kit. Forego the obvious, such as laptops, survey software, APs, antennas, batteries, and tripod/poles, unless some specific information about them has been found to be particularly helpful/essential.

You still doing 'AP on a stick' surveys?

Please do a bit of an empirical test. See what kind of measurable differences you get from an AP on top of a file cabinet, and an AP up on a tripod, or zip-tied.

I think you'll be surprised with the changes.

Were you doing a Sample Install? or a survey to find the RF characteristics of the environment?

Did you test the system with at least four Access Points running at the same time? under load? all under load? Test RSSI? Test user per AP density? test Coverage Overlap? test channel interference? test Jitter? test throughput? test MIMO spatial streams? etc...

If you did the 'normal' just with a single AP measuring RSSI values and data rates only, how do you know the network will perform to meet all the Wireless LAN Design parameters?

You won't know until you test them as well.

OK, now for the less 'flippant' answer.

My site RF Analysis kit (checking to see the RF environment prior to designing a Wireless LAN)

The goal of this visit is to determine RF characteristics of the materials, walls, floors, etc. of the site.

Kit includes:
4 Access Points - all same brand, same power, same antennas
4 25' extension cords (to power the APs) - no batteries needed
4 25' Cat5 Cables
4 PoE Injectors

2 laptops with AirMagnet Survey, with integrated AirMagnet (Cognio) Spectrum Analysis cards
(one laptop is a 'spare' in case of any failure)

I model the area in AirMagnet Planner. Placing the four access points on the floor plan, with same power, antenna, and channel settings.

Then I place the four APs in the same positions in the real world. Then run a joint 802.11/RF Spec An survey of the entire area around the 4 APs. Clear to their signal dropping below -90dBm.

Then try to 'match' the two survey's using the DiffView in AirMagnet survey.

I change the 'wall thickness' settings in Planner until I can get them to 'match' those from the real world. Thus 'calibrating' the virtual survey in Planner to better match the real world.

I might repeat this process in every section of the facility that has a difference style or era of construction. Plus a couple of 'spot checks' around to make sure my RF analysis is working.

The first 'passive' survey also sweeps ALL available channels that might be legal in the environment. Plus I re-check the Spectrum Analysis data to make sure there aren't any RF anomalies in either the 2.4GHz or 5GHz spectrum.

Then, and only then, can I open up Planner and continue the design process - the main goal of which is to meet ALL the Wireless LAN design parameters.

Remember, "Coverage is Easy"!!! - Getting just the appropriate RRSI values is the easiest thing in the world! That is where most 'Site Survey' teams stop... it is merely the beginning. The HARD WORK is to make sure your Wireless LAN Design meets all the other criteria in ADDITION to Signal Strength.

By the way, minimizing channel interference (and thus Collision Domains) is the hardest thing to clean up.

If there is going to be an attempt at Location Tracking, there are some other additional steps.

Whew!

Sorry for the long answer to a short question.

By the way, when I teach my Advanced Wireless LAN Design class, answering your question takes about 2 1/2 days!

Hi Keith

WOW!!!

Really opens your eyes when it is spelt out like that. Pardon me if I am just being a bit thick, but is that indepth method of surveying covered in the CWNP curriculum?

CP

the short answer is 'no' - the CWNP curriculum does cover "site survey" - but what I was discussing in the last post was "Wireless Network Design"

I personally think we should stop using the term "Site Survey" - what our clients want is a working wireless network... just like they want a working wired network.

In a wired network, how much time is spent 'specifying' the cabling plant? That is the specs for the physical layer.

A 'Site Survey' is the proof that the wireless network meets your design specs. In a wired network we use something like a Fluke LANmeter to prove the cabling is say, "Cat6" - meaning it meets all the specs for Cat6 (connectivity, near-side cross-talk, far-side cross-talk, pin-outs, twist ratios, etc.)

You test the cable AFTER it is install, not while it is still on the spool on the back of your truck.

So we MUST do a Site Survey AFTER install to prove the wireless network meets it's design specs. These are pretty precise (just like the Cat6 ones) - they include not only RSSI, but double coverage (overlap), channel interference, jitter, throughput, data rates, client density, etc.

What the previous post was talking about is Wireless Network Design. The important word is DESIGN!

The goal of the on-site visit is to learn of the RF Characteristics of the site. NOT to design your network.

By using the 'AP on a stick' method, you are designing the Wireless LAN on 'accident' - based on the first placement of an AP - leads to the second, third, fourth, etc. Each AP is placed according to where it's coverage pattern overlaps with the previous Access Point.

This is a very simplistic way to design. You may as well had just stayed home and dropped colored circles on the floorplan.

The easiest thing is to get 'Coverage' - it is merely getting RSSI to a spot in three-dimensional space. "Tim the Toolman Taylor" could do this with his eyes closed. Getting coverage is EASY!

The hard part of the design is making sure your design meets all the other criteria. Specifically making sure you have 'double coverage', making sure you don't have Collision Domain issues (channel interference)

Coverage is easy, just put up more APs with more power and you can ALWAYS get coverage.

But the intelligent Wireless LAN designer works on confirming all the other criteria FIRST - RSSI and coverage will necessarily follow. But starting with RSSI First will leave the other criteria un-fulfilled.

Simple things like NOT putting APs in hallways - hallways are easy to use, give you coverage, but also ALWAYS cause channel interference! You'll never figure that out with an AP on a Stick, since you only have a single AP, you'll never see your channel interference. You'll get your coverage all right, but be causing more problems down the road.

Remember the goal of a Wireless Network Design isn't to just provide coverage, but to meet ALL the design parameters the client needs for a working wireless network!

Also remember, the design parameters for wireless networks can be mutually exclusive! Location tracking and Voice can be on opposite ends of a spectrum - meaning it is really hard to get those two to work together - one wants lots of APs all over, and the other only wants a few strong APs.

Check your device specs BEFORE starting your design to make sure your client knows exactly what they want - first.

Thanks for the good design info, Keith. While I definitely see the merits and agree with your approach to surveying and design, the survey I participated in was for a small facility with only a few APs and the survey was just for coverage. The ceilings were fairly high, which is why we wanted to test the RF environment with real-world AP placement. Surely most survey examples would benefit from your approach, but it would be hard to justify for small, low-cost surveys like this one. No less, I'm comfortable deferring to your expertise in design and surveying. :)

Marcus,

It is actually a faster method of Wireless Network design than the AP on a Stick method.

A 4-AP location would take no more than one hour. Total time. Start to finish.

Not only does it give you much better results, but it save you lots of time as well.

Second, I have *never* had a client who asked for coverage-only... ever end up actually wanting dB's delivered, they want packets delivered - they just don't know how to ask the right questions.

That's why it's up to the WLAN professionals of the world to ask the right questions, so we can design and deliver exactly what they *really* want. Then we all can look better.

Now that you've got a little experience like this one... the very best thing you can do to reinforce the knowledge is to do a Post-Install survey of the exact same location. Then compare the results!

Pre-install survey's have been 'vogue' for awhile - but Post-install surveys should be required of all wireless LAN installations.

I'd be very interested in your pre vs post survey analysis. That is a learning experience that might help others through sharing.

Thanks for the comments,

Keith

Keith, what do you mount your AP's on and at what height?

Thanx,

David

I don't believe a good kit is a one time purchase. There is more often than not, the one extra thing you wish you had for your current work and buy shortly after or during. Then there are specialy tools, software or hardware that you use once or twice but carry just incase. Just like working on cars, when you still could, you accumulate tools as you go and as the technology changes. I agree with Keith that the network design method is a better plan than is used by most survey techniques. The survey should be based upon the intended use of the WLAN and the environment using the actual devices and programs to be deployed if possible. I have seen too many networks having trouble due to poor or no survey being conducted. A good Site Survey/WLAN design course is really needed in our industry.

Thanks for starting this thread Marcus.

Boring as it may sound... I just place my Access Points one whatever works in the environment. I found years ago very little difference between a normal office ceiling, and the top of a filing cabinet or top of a cube wall, etc.

It is sooooo much easier to not carry the tripods, batteries, etc. Especially when I'm NOT trying to have a 'sample install' but to find out RF characteristics. (remember, you MUST have multiple APs under load to even get close to accurate information from the pre-survey) that's lots of equipment.

I now keep it down to a single Pelican case I can travel with as airline luggage, under 50lbs. I can't 'afford' all the tripods, batteries and garbage. And I've found you don't need them.

(sorry to all those who think this is blasphemous - but it's true)

That is for the 'pre-survey' - for a Post Install survey, all the equipment is already in, and I just need a laptop, and perhaps one AP to do some additional RF spot checks.

Caveat - The above is for an office or hospital type environment. Different plans for warehouse, distribution center, airports, or construction sites.