Having completed a WLAN design for a distribution centre I have a query:
The client requirement of a -72dBm cell edge has been achieved throughout the premises along with a SNR of at least 20dBm. This has been proven with both an Ekahau and Airmagnet post site survey. The site has not gone live yet as they are still using their legacy WLAN until their upgraded Handheld Terminals (HHTs) are rolled out across the site and the Forklift Truck (FLT) devices have been upgraded also. The client has concerns that in certain areas the HHTs and the new test FLT devices are reporting a poorer RSSI than our software reports. The power has been set at 1 on the controller because the APs have a clear line of sight of one another as they are mounted high above the racking at 12 metres so are naturally reducing their TX power when left to RRM to manage which has a consequence to the clients on the ground. Without going into the model of AP and directional antennas used in this design, my question is because we have met and exceeded the customer requirements should additional equipment to satisfy the customers concerns with their devices be at our cost as in time and hardware or theirs as they have moved the goal posts to suit (understandably so) their device readings throughout.
Our software still shows it passes the requirement even though the stock has changed round since the initial site survey was completed. We are only using the Proxim 8494 adaptors and the initial site survey was completed at 1/2 Tx power taking the antenna gain into account also.
I can add more information regarding hardware but initially just would welcome your thoughts?
re:The client has concerns that in certain areas the HHTs and the new test FLT devices are reporting a poorer RSSI than our software reports.
re:Our software still shows it passes the requirement even though the stock has changed round since the initial site survey was completed.
First get one each or more of the client devices. Set them on a table not close to each other. Note the RSSIs. Rotate, stand-up, move slightly and record the RSSIs. Calculate average. Remove them. Put survey laptop there. Do same to get average. During survey eval, add in the difference. Now you're surveying to customer requirements. Do for active and passive surveys.
Using your survey tool without calibrating to client devices does not mean "you have met and exceeded the customer requirements".
Consider the stock as lead walls to the ceiling all around their striped off locations. Even if there is no stock at all. Therefore deploy APs only in paths of trucks carrying scanners. The active survey is almost superfluous because if you use the survey to deploy APs then you are considering the current product placement. You'll get tickets forever.
BUT you must dictate that scans are not guaranteed while "in" the product.
Thanks for the response but I've now satisfied the customers' concerns. The couple of dB drop at the ends of the aisles has been solved with a small degree of tilt on the antenna' towards either end of the aisles.
It's not the perfect/ideal design in this type of environment because of restrictions in possible mount positions but it works which is the main thing. The customer was asking for more APs for infill but I advised against this as there are enough APs throughout and that I can make slight changes to the antenna positions to satisfy the areas of concern. These areas were at the ends of the isles only and was spot on everywhere else.
I don't agree with you regarding treat each aisle as if it's a lead wall from floor to ceiling because they're obviously not and the different aisles do not have enough isolation to separate them completely irrespective of what product this particular client uses and as a result of putting in APs in each aisle would cause more trouble than not especially at 2.4GHz.
It's a lesson learned and next time I will complete the site survey at 1/4 power to ensure that extra 3dB will be there if required.
Thanks for your input.
As product type changes (density) and quantity changes and stacking changes the RF that you measured in specific areas that previously made its way through the product might not make it anymore. Looks fine in the beginning though!
Much more the case when using 5GHz vs. 2.4
As Forrest already noted, you should design and survey for the least capable client device. Sidekick and AirCheck are premium devices as is your personal laptop if you used it for surveying. If the client uses low-end Android cell phones or hand-held scanners then you need to adjust your design. Calibrating like Forrest described is one way.
RRM in a warehouse may be troublesome. The APs are high up with clear LOS to each other while the users are down below between the aisles. You didn't specify but apparently/hopefully you have used directional antennas. Omni antennas mounted high above is a recipe for trouble especially with RRM.
Aircheck, and most likely Sidekick, have customization controls that will let you "tune" them to more correctly act like your client devices. Try re-running your survey after you have made those adjustments.
It is almost guaranteed that RRM is spoiling your coverage. Try narrowing the power level range-limits it is using to tune itself.
It sounds like the ultra-high AP's was the first big mistake in the design - there are always alternatives.
Thanks for all the responses.
I did use directional antenna at 12m along the aisles, and there was no other option as I’ve already said. I do survey to the weakest device with regard to its max transmit power etc. These things I’ve learned in both experience and whilst passing CWDP and CUWSS exams along the way. It was simply a situation I’d never come across in a decade I’ve been involved in Wi-Fi and wanted your views. The design is actually perfect with the options I had available as every option had been investigated and already tried. I just wanted to avoid excessive amounts of radios throughout the warehouse when they’re not needed. I’m not using RRM because as I said this was causing many issues with coverage because of how strong the APs can hear each other up above the racking. I wouldn’t call 12m ultra high as this is not the roof but is at the top of the racking so that FLTs don’t take them out. It’s on 3m stud from the ceiling to where it needs to be. Having spoken with client the stock is pretty much what it is where it is all the time so there will not be any nasty surprises moving forward. I wouldn’t like to charge the client another 20000 in hardware whilst explaining that this is for contingency only, when there’s enough APs to cover both coverage and capacity as well as AP resiliency. The customer is now very happy with the extra couple of dB achieved where they had concerns.
Sounds like you have it covered. Only time will tell.