Warehouse 5 GHz Coverage Collapse Behind Loaded Racking
Ridgeway Distribution operates an 80,000 sq ft (7,400 m²) warehouse with 12 m ceilings, steel racking in 2.8 m aisles, and dense pallet storage. Handheld barcode scanners run a 5 GHz-only WLAN. Six APs are mounted at 10 m on rack uprights with 6 dBi omnidirectional antennas, 20 dBm transmit power, and 3 dB cable/connector loss. A pre-installation walk in an empty aisle measured -67 dBm at 15 m; after the racking was filled, the same scanner reads -79 dBm at 15 m and disconnects at 25 m. The customer's SLA requires -72 dBm or better at scanner height (1.2 m) in every aisle. A vendor proposes adding three more APs at the same height and increasing power to 23 dBm. The site has 5 GHz UNII-1 and UNII-3 channels available, and the client wants to keep 5 GHz for scanner throughput.
1. The vendor calls the extra loss 'warehouse attenuation' and proposes 23 dBm. Quantify how much of the loss is unexplained by free-space path loss, and evaluate the power increase.
Read the worked answer
Free-space path loss at 15 m and 5.5 GHz is 32.44 + 20log10(0.015) + 20log10(5500) = 70.8 dB. With a 20 dBm radio, 6 dBi antenna and 3 dB loss, EIRP is 23 dBm, so predicted RSSI is 23 - 70.8 = -47.8 dBm. The measured -79 dBm is 31.2 dB worse, not 12 dB; the empty-aisle baseline of -67 dBm was already 19 dB worse than free space because of antenna pattern and near-field racking. Raising power to 23 dBm adds only 3 dB, leaving about -76 dBm, still below the -72 dBm SLA; it also produces 26 dBm EIRP, which is not legal on UNII-1 (23 dBm cap). The dominant loss is blockage and absorption by palletized stock plus mounting at 10 m, where the omni's main lobe is above the scanner and rack metal creates deep nulls. Power cannot fix a geometry and obstruction problem.
2. Propose a design change that meets the -72 dBm SLA at 1.2 m without simply adding power. Include mounting, antenna, and channel considerations.
Read the worked answer
Move the APs off the rack tops and down into the aisles at 5-6 m, side-mounted on rack uprights with 8-10 dBi 60-90° patch or sector antennas aimed along the aisle. That puts RF where the scanners are, reduces the amount of stock in the path, and cuts multipath from rack faces. Plan cells of about 15-18 m, roughly one AP per two to three aisles, and use 20 MHz channels from UNII-1 and UNII-3 with a repeating non-overlapping plan; 5 GHz has enough channels that a six-AP replacement can avoid co-channel reuse. Disable 802.11b rates and set 12/24 Mbps as basic rates so scanners do not cling to low rates at long range. Keep 5 GHz only. Verify with AP-on-a-stick while racking is fully loaded, because the empty-warehouse measurement is not predictive.
3. What verification steps and acceptance criteria would you write into the installation contract?
Read the worked answer
Write the acceptance test around the loaded condition: with racking at least 80% full, measure RSSI at 1.2 m every 10 m down each aisle on the scanner channel, requiring -72 dBm or better and SNR of at least 25 dB in 90% of samples, with no sample worse than -75 dBm. Run a 10-minute throughput test at each aisle end and midpoint using the scanner vendor's tool, with packet retry below 10% and no disconnects during a walk the length of the aisle. Test roaming between APs with an active scanner session. Record channel, width, power, and firmware in an as-built. Include a post-installation validation 30 days after go-live, because stock levels change seasonally and the RF environment will not stay as tested.