Scanners Drop Off in the Far Aisle
A 60,000 sq ft warehouse runs 26 handheld barcode scanners, 9 forklift-mounted tablets, and 40 office laptops. Scanners drop off the WLAN for 5 to 20 seconds in the far aisle. Staff report "the Wi-Fi is down," but office laptops are unaffected. The site has 6 access points set to 2.4 GHz channels 1, 6, and 11 in a repeating pattern, with 5 GHz also enabled. The scanner VLAN is a /24 with a 2-hour DHCP lease; 26 scanners, 9 tablets, and 12 spare test devices currently sit on it. Scanners are locked to a 2.4 GHz-only profile because a vendor once said 5 GHz "doesn't work" near the freezer area. The drops began after holiday peak staffing added 15 more scanners to the fleet.
1. Name the two most plausible causes and describe the evidence that would separate them.
Read the worked answer
Cause A - DHCP address exhaustion on the scanner VLAN. A /24 gives 254 addresses minus the router and reserved ranges; 26 + 9 + 12 = 47 in use, so the pool is not exhausted. That arithmetic actually argues against A as the primary cause - but only if those 12 "test devices" are the only other occupants and if the lease table matches. Verify, do not assume. If the pool is fine, A drops out cheaply.
Cause B - 2.4 GHz coverage, co-channel interference, and roaming failure in the far aisle. Scanners locked to 2.4 GHz cannot use the 5 GHz radios, and a 6-AP pattern for 60,000 sq ft leaves weak coverage at the edges. That produces exactly the described symptom: a device at the cell boundary disconnects, takes 5 to 20 seconds to reassociate, and appears to "come back" on its own. Office laptops are unaffected because they are stationary, close to APs, and free to use 5 GHz.
The separating evidence: do the drops follow the device or the location? If a scanner that drops in the far aisle works perfectly near the office, coverage or roaming is the cause. If it drops everywhere, look at the device profile or the DHCP/authentication layer. Also pull the AP logs for deauthentication and reassociation events at those times - a client bouncing between two APs at similar signal strength is visible there and nowhere else.
2. What is the cheapest first test, and what would make you escalate to a site survey?
Read the worked answer
The cheapest test is a walk test with one known-good scanner and one office laptop: walk the far aisle slowly while watching signal strength, roaming events, and the time of each drop. It costs one person and thirty minutes, requires no new hardware, and directly tests the location-versus-device question. Record signal strength at the point of each drop.
Alongside it, pull two logs that already exist: the DHCP lease table for the scanner VLAN (does it match the 47 devices you expect, and are leases churning?) and the WLAN controller's client event history for one affected scanner over the last 24 hours.
Escalate to a proper site survey when the walk test shows the drop correlates with location and signal strength is weak or two APs are competing at similar levels. That is when you need physical survey data - racking heights, freezer wall attenuation, and AP placement in a 60,000 sq ft space cannot be judged by walking with a laptop. Escalating before that point buys you a report you could have written yourself.
3. Propose a durable fix and explain how you would validate it without disrupting picking.
Read the worked answer
The durable fix has three parts, and only the third is expensive.
- Reclaim the pool. Remove the 12 test devices from the production scanner VLAN onto a dedicated staging SSID. Even if DHCP is not today's cause, it removes a variable and prevents a real outage later.
- Challenge the 2.4 GHz-only rule with evidence, not argument. Test 5 GHz in the freezer area on one scanner. The old vendor statement may have been true of hardware you no longer own. If 5 GHz works, freeing scanners onto the less congested band solves most of the problem without moving a single AP.
- Fix coverage only where the data says coverage is broken. Add one AP or adjust power and channel plan in the far aisle if the walk test shows a genuine hole.
Validate during a shift change or a scheduled non-picking window, with two or three volunteers on a single aisle first, and measure the same metric you started with: the number and duration of drops over one hour of normal scanning. Roll out to the full fleet only after that hour is clean. Any fix that cannot be measured against the original symptom is a change, not a solution.