Capturing quality data on a shop floor with dead wifi zones
Every plant has them. The corner behind the press line, the cell inside a steel enclosure, the spot where the tablet loses signal. If your quality capture dies with the wifi, the data ends up on paper. It does not have to.
Walk a stamping plant with a signal meter and you learn fast that a factory is built to fight radio waves. Steel decking overhead, structural columns every few metres, press frames the size of a small house, machine enclosures wrapped in sheet metal, and an overhead crane that rolls through and swings the dead spot to a new aisle every ten minutes. You can survey the floor, hang more access points, and still leave pockets where a tablet shows one bar or none. An operator standing at the machine hits save, the spinner turns, and nothing happens. So the number goes onto a scrap of paper to be typed in later, and the digital capture system you bought to make records trustworthy has quietly made this one worse than a paper log ever was.
Paper fallback looks harmless and is not
Everyone treats the paper fallback as a small, sensible workaround. Write it down now, key it in when you get back to a good signal. The trouble is what falls off in that gap. A reading taken at 09:40 and typed at 15:00 carries the wrong time, so when an out of spec result shows up you cannot honestly reconstruct the sequence of measurements around it. The person keying the backlog is often not the person who held the caliper, so the record now says someone measured who never did. And a value copied off paper is no longer tied to the released control plan characteristic it belongs to. It is a bare number in a box. That is not a smaller version of good evidence. It is a different thing wearing the same shirt.
What a connection drop actually breaks
- A form that fails to save mid entry, so the measurement the operator just took is gone and has to be redone or guessed.
- Values retyped later that carry the transcription time instead of the true measurement time, so the record is off by hours and nobody can tell.
- In spec and out of spec feedback that needs the server to answer, so the operator is working blind at exactly the moment a bad part matters most.
- Duplicate records when a save times out and gets retried, so one measurement lands twice and the history stops adding up.
- Capture attributed to whoever typed the backlog in, rather than the signed in inspector who did the work.
Continuity beats a second offline system
The wrong fix is to bolt on a separate offline app that stores readings its own way and reconciles them into the real records later. Now you have two capture flows, two write paths, and a reconciliation step that is its own source of errors. What a shop floor actually needs is continuity. One capture flow, one write path, one record, whether the wifi is solid or gone. The kiosk loads everything it needs while it has a connection, the parts, operations, characteristics and specification limits, and then keeps working from what it holds. Entries made during a dead zone stay on the device and submit through the exact same records path the moment the connection returns. There is no second system to trust, because there is only ever one.
How VoraSuite rides out a dead zone
VoraControl Inspector Capture floor mode is a big touch kiosk built for a tablet bolted next to the machine. Because the plan specification limits travel with the working set, in spec and out of spec evaluation happens instantly right on the device during a drop, with no round trip to a server. Every entry keeps its true measurement time and the signed in inspector who took it, not the time and name of whoever caught up on typing afterward. When signal comes back, the entries sync through the same capture engine every other reading uses, protected against duplicates so a retried save never doubles a record, and out of spec results raise their signals and hand off to Resolve as though the wifi never blinked. The VoraTool toolmaker jobs kiosk rides out the same drops for job card work, timers, photos and hand offs, so the toolroom keeps moving through the same dead aisles.
Continuity, not magic
It is worth being straight about the boundary here. This is continuity, not a system that runs on nothing. The kiosk needs a live connection to sign the inspector in and pull its working set, the parts and characteristics and specification limits it will evaluate against. Once it holds that set it carries you through the dead zones that no site survey ever fully clears. But it is not a licence to ignore the network. If a whole line sits dark for a shift, that is a wiring or coverage problem that deserves fixing at the access point level, and no capture tool should be quietly papering over it. Continuity buys you the honest record through the normal drops. It does not excuse leaving a plant half covered.
The record you are left holding
When the drops are handled this way, the capture history for a characteristic reads as one continuous, honestly timed stream, with no gap where the wifi died and no cluster of readings all stamped with the same afternoon catch up time. SPC trends and capability come out of measurements that carry their real times, so a shifting process shows up when it actually shifted. And when an auditor asks for the inspection evidence on a part, what you hand over was made at the machine, by the person who did the work, at the moment they did it. VoraAudit reads that straight from the source records for you to review on screen or print. That is the difference between a system that survives a dead zone and one that fails quietly every time the signal drops.
Released control plans, inspection capture and SPC trends in one governed flow.
Talk to someone who understands manufacturing control and audit evidence. We do not do generic demos.