Troubleshooting & FAQ

Symptoms first. Each one says what is happening behind it and what to do next. Where the answer is not something you can settle from the portal, reach out to your provider — they set your site up and can see the hardware side.

A zone’s number is not updating

What’s happening. A volume needs three things: a floor to measure up from, enough returns to describe the surface, and enough of that surface in view. When one of them is missing for a zone, that zone stops producing new numbers and keeps showing the last one it produced.

The usual causes, most common first:

No visible floor. The material has grown over the bare ground the zone was measuring up from, so there is nothing left to model the floor with.
Too few returns. Dust, rain, a fouled sensor window, or a sensor that has drifted off aim.
Too little of the surface seen. Something new sits between the sensor and the zone, or the zone has grown past what the sensor reaches.
One sensor of several has gone quiet. Where more than one sensor covers a zone, part of the zone is no longer being measured.

What to check. Compare the zone with its neighbors. If their timestamps are advancing and this one’s is not, the problem is local to this zone — its view, its floor, its footprint. If nothing at the site is advancing, it is capture or connectivity; see A zone is flagged stale.

What to do. Clear the sensor’s line of sight. Widen the footprint so it takes in a ring of bare ground around the material. If the zone is a bay or bunker that rarely runs empty, capture an Empty scan the next time it is cleared — that stores the exact floor and structure, so the zone no longer needs bare ground in view. If none of that applies, reach out to your provider.

A zone reads zero

What’s happening. A published zero is a real measurement: the sensor saw the floor and nothing on it. It is not a failure code and not a placeholder.

What to do. Nothing — it climbs again on the next delivery. If a zone you know has material in it reads zero, the floor baseline is the thing to look at: a baseline captured with material still in the zone, or a footprint drawn so tight it barely contains any material, can both put the floor level with the pile. Recapture the empty scan with the zone genuinely clear, or redraw the footprint wider than the material.

A zone reads low

What’s happening. Usually the floor or the footprint, not the material.

What to check, in this order:

The material itself. Check the history for a matching drawdown before anything else — a low reading is often correct.
The footprint. Drawn too tight, it clips the edges of the pile. Drawn over an adjacent wall or ramp, it gives structure detection something to latch onto that you did not intend. It should cover the storage area with a little room around the material, and stop short of the hardware.
The floor baseline. A baseline captured a while ago may no longer match the bay — the floor was regraded, a divider moved, or the sensor was re-aimed since. Capture a fresh Empty scan when the zone is next clear.
The sensor’s view. If the sensor only reaches part of the pile, only that part is measured. Persistent partial coverage is a mounting question — reach out to your provider.

What to do. Redraw the footprint or recapture the empty scan, whichever the checks point at. The reading recovers on the next capture.

A zone looks stuck

What’s happening. At rest a zone’s number holds steady rather than trembling a few units either way — that jitter is the sensor, not your inventory. The number moves once consecutive captures agree that the pile has changed, so a delivery shows up as a step rather than a ramp. A flat line on a pile nobody has touched is correct.

A flat line on a zone you know has been worked is not. That is the same symptom as A zone’s number is not updating — work that list.

What to do. To see the individual captures behind a step, expand the segment in the history table.

Something was parked in front of my zone

What’s happening. An obstruction hold. Equipment or people inside the zone — a loader working the face, a truck backing in, someone walking through — would otherwise be counted as material, so the volume is held at its last trustworthy value until they leave. The zone carries an Obstructed badge while this is happening, and Stabilizing for the short settling period after the view clears.

What to check. The Perception evidence card on the zone’s detail page. Each row names what was seen, when it started, how long it lasted, and the volume held. Click a row to replay it in the 3D scene. This doubles as a record of what happened in the zone — deliveries, people, equipment.

What to do. Usually nothing. Holding is controlled per zone by Obstruction-aware volume, on by default.

Readings looked frozen while equipment moved through

What’s happening. An occlusion hold. Where an obstruction is in the zone, an occlusion gets between the sensor and the zone — a truck on the haul road, a stacker swung across the aisle. Rather than treat the sliver still in view as the whole pile, the zone holds its last value until the view recovers.

What to do. Check how long it lasted. A hold as long as a machine takes to pass is working correctly. One that lasts a shift means something is parked in the sightline — move it, or talk to your provider about re-siting the sensor.

A zone is flagged stale

What’s happening. No new scan has arrived for that zone in several days, and it is marked on site dashboards with how long — No scan · 5d. The number on screen is still the last one received; stale means it may no longer describe what is physically there.

What to check, in order: whether the zone is paused (paused zones keep their last number and are dimmed in bay layouts); whether the device is online; whether anything has changed about the sensor’s view.

What to do. Restoring capture clears the flag on the next scan. If the device looks healthy and scans still are not landing, reach out to your provider.

My bay has a sloped floor

What’s happening. Sloped floors are measured as they are, at any angle — a drive-through bay that falls two meters end to end is measured as the sloped surface it is. An empty scan captures the slope exactly.

What to do. Draw the zone and let it measure. If a sloped bay does read wrong, work it the same way as any other: the footprint, the sightline, the floor baseline.

My tonnage doesn’t match my scale tickets

What’s happening. Tonnage is derived, not measured: volume × density. If the volume looks right and the tonnage does not, density is the thing to look at.

What to check.

The product on the zone. Density comes from the assigned product, so the wrong product gives a right volume and a wrong tonnage.
The zone’s density override, under Advanced settings. Left blank, the product’s density applies.

What to do. Correct the product or the density and the tonnage follows. If they still disagree, compare volumes rather than weights — a volume that tracks your tickets with a weight that does not is a density problem; a volume that does not track them is a measurement problem.

Days to empty is missing or doesn’t look right

What’s happening. Deliveries and days-to-empty are both inferred from the volume history rather than measured.

Deliveries are detected from increases in volume between readings. A scale ticket attached to a capture is definitive; without one, a change has to be significant both in proportion and in absolute size before it counts.
Days to empty is a projection from recent consumption. It needs enough history to see a drawdown, and it is only as steady as that drawdown. A new zone, or one drawn down in irregular bursts, will not produce a useful one.

What to do. Check that the changes you expected are present as volume changes first — a zone that has stopped updating shows no deliveries at all. Attaching scale tickets to captures makes movement unambiguous. Treat both as decision support rather than a ledger; if a projection consistently misleads, set an alert on fill level instead.

The history table doesn’t have one row per scan

What’s happening. The history folds runs of agreeing captures into segments — one row for the period the reading held steady, annotated confirmed 48 times or similar, instead of hundreds of identical rows. The newest segment is still in progress, so it reads as current.

What to do. Expand a row to see the readings behind it. If you need every reading in another system, ask your provider about the API.

I can’t start an empty scan

What’s happening. A site runs one floor capture at a time, so a second request while the first is pending is turned down until the first finishes.

What to do. Wait for the running capture, then retry. Two other causes look the same: the zone is brand new and has not been saved yet (save it, then reopen it), or you do not have edit rights on that site. Once a capture starts, keep the zone clear until it finishes — it usually takes a minute or two, and a failure usually means the zone was not fully cleared.

I can’t edit a zone

What’s happening. Access is granted per site, at three levels:

You can…	Requires
View sites, zones, and measurements	Viewer
Edit zones, capture and clear floor baselines, record adjustments	Editor
Manage the organization, its members, and their access	Admin

Editing a zone requires editor rights on that specific site. It is normal to be an editor on the sites you run and a viewer on the rest. Editing controls are hidden when you only have view access, so a missing Edit or Add zone button is a permissions answer rather than a missing feature.

What to do. Ask an administrator in your organization to raise your access on that site. If you are the administrator, see Team & Access.

Still stuck?

Not sure what a term means? See the Glossary.
Setting up or fixing a zone’s floor? See Floor Modes & Empty Scans.
How the volume is worked out: Volume Methodology.

Anything that needs an account, a device, or an organization-wide setting changed goes through an administrator in your organization. Anything to do with sensors, mounting, coverage, density references, or how “full” is defined for your bays goes to your provider.