A conveyor that will not run is not a mystery. It is one of four measurable conditions: a permissive that never closed, a drive enable that never asserted, a current or thermal limit that tripped the drive, or a logic branch waiting on a signal edge that never arrived. Every one of those leaves a number somewhere you can read. The number that matters first is the last fault latched in the drive's fault history, and the state of the interlock string measured at the terminal block, not at an indicator LED.
When a commissioning day ends with no progress, it is almost never because the system is too advanced. It is because the search was unordered - a dozen partial checks, none of them written down, none of them ruling anything out. Order is what converts a ten-hour day into a two-hour fix.
Fault Classes and Where the Number Lives
Conveyor systems fail in a small number of repeatable ways. Match the symptom to the mechanism before touching the program.
| Symptom | Mechanism | First measurement |
|---|---|---|
| Motor hums, shaft does not turn | Lost phase, brake not releasing, or mechanical bind | Current on all three legs at the starter; brake coil voltage at the coil |
| Drive trips during acceleration only | Accel ramp shorter than the load inertia allows; drive hits current limit | Drive fault history, accel time parameter, output current during the ramp |
| Drive trips after several minutes of running | Thermal. Motor overload or drive heatsink. This is heat, not logic | Running current against motor nameplate FLA; drive thermal reading |
| Zone accumulates and never releases | Photoeye stuck made or stuck clear, or a jam timer expiring before the downstream eye clears | Force the eye by hand and watch the input bit online |
| Logic looks correct, nothing moves | Output coil energized but no field power on the output group | Voltage at the output terminal and the load side of the group fuse |
| Random stops with no operator action | Safety relay or interlock chatter; loose wiring on a series string | Monitor the safety contact input while performing a wiggle test on the string |
| Section runs the wrong direction | Phase rotation at the motor leads or a reversed direction parameter | Phase rotation meter, or the drive direction command bit online |
An indicator LED on an output card tells you what the processor wrote. It tells you nothing about whether 24 V reached the field device. When those two disagree, the fault is between the card and the load, and a meter at the terminal ends the argument in thirty seconds.
Four Ways Out of a Stalled Day
Once a fault has consumed a full shift, the decision is no longer technical. Four paths exist, and they carry very different costs.
| Approach | Time to resolution | Commercial exposure | Best when |
|---|---|---|---|
| Keep grinding solo | Unbounded. Repeats checks already done | High - the schedule slips silently | Fault is narrowed to one subsystem and a measurement is pending |
| Structured remote session with a senior engineer | Hours, if a fault packet is prepared first | Low | The fault is in logic, configuration, or drive parameters |
| Escalate for a second engineer on site | Travel time plus hours | Low if raised early, high if raised after the deadline | Fault requires two people - one at the panel, one at the field device |
| Sectional handover and re-baselined schedule | Immediate relief on the deadline | Managed, and owned by the project manager | Working sections can be proven and signed while one section stays open |
Grinding solo is the default and the worst of the four, because it burns the only asset that matters on a commissioning job: daylight with the machine powered and the contractor's people available to move product through it.
The Recommended Path: Fault Packet, Then Escalation
Escalate with data, not with a request for rescue. A senior engineer on a phone call has no meter, no eyes on the panel, and no feel for the machine. Everything they can contribute depends on what you hand them in the first five minutes. That is why unstructured remote help produces nothing while a prepared session produces an answer.
Set a hard cutoff before the day starts - a fixed number of hours after which you stop troubleshooting and start escalating. Escalation raised before the deadline is a schedule decision. Escalation raised after the deadline is a failure report. The technical content is identical; the commercial consequence is not.
Field Procedure
- Back up the program before anything else. Upload the running program from the processor and the parameter set from every drive on the section, and store copies off the laptop. Nothing else on this list is recoverable if that step is skipped.
- Read the drive fault history. Record the fault code, the order of faults, and whether the trip repeats at the same point in the cycle.
- Confirm control power and field power on every output group and every 24 V distribution block feeding the section. Measure at the terminal.
- Go online and force nothing yet. Watch the interlock string, the safety inputs, and the zone photoeyes while an operator walks a carton through the section. Note the first bit that does not change when it should.
- Isolate the section. Bypass upstream and downstream handshakes only if the design permits it, and jog the drive locally to separate a mechanical or electrical fault from a logic fault.
- Write down every check as pass, fail, or unknown. A check that is not written down will be repeated tomorrow.
- At the cutoff hour, stop and assemble the fault packet.
Making a Remote Session Produce an Answer
Bandwidth between the site and the remote engineer is the bottleneck, not their skill. Compress it before dialing.
- A one-page written statement: what the machine should do, what it does instead, and at what point in the cycle it diverges.
- The ruled-out list. Every check from the procedure above with its result. This alone cuts a remote call in half, because the first thing any senior engineer does is re-ask what you already tested.
- Screen share of the online program at the rung where the sequence stops, with the tag names visible. Rung numbers and tag names give the remote engineer a map; a verbal description does not.
- Drive fault history and the parameters that were changed, if any, with the original values.
- Photos of the panel interior, the field device in question, and the terminal numbers involved.
If two calls with the same engineer do not converge, the fault needs eyes on site. Say so on the call rather than booking a third.
Schedule Pressure Belongs to the Project Manager
The contractor's deadline is real and contractual, and it is not yours to absorb. Your company holds the contract; the project manager owns the schedule commitment against it. A field engineer who accepts direct schedule pressure from a contractor is negotiating a contract they have no authority over.
The mechanics of keeping that boundary clean:
- Send an end-of-day report to your line manager and project manager every day, including days with no progress. Two lines on what was ruled out and one line on what you need is enough. A day with no report reads as a day with no work.
- When the contractor asks for a completion commitment, give them the technical status and route the date to your project manager by name. Contractors accept this readily - they deal with it on every job - and it removes you from a conversation you cannot win.
- Keep the issue list in a file, not in your head. When a second engineer arrives, the list is what determines whether their first hour is useful or spent re-discovering the problem.
- Being visible while stuck is more valuable than being quiet while stuck. Silence from site is read as everything is fine, and that is exactly when the schedule breaks without warning.
Verifying the Fix Before You Claim It
A conveyor that moves once has not been commissioned. Prove it with numbers before the section is offered for sign-off.
- Run the section loaded, at design rate, for a continuous period long enough to reach thermal steady state. Thermal faults appear minutes in, not seconds in.
- Measure running current on each motor against nameplate FLA. A motor pulling near or above nameplate on a lightly loaded conveyor is a mechanical problem waiting to become a fault code.
- Clear the drive fault history, run the qualification cycle, and confirm the history is still empty afterward.
- Check photoeye margin - block and clear each eye at the actual product position, not by hand at the sensor face - and confirm the zone timers do not fire at production rate.
- Run the failure cases the deadline pressure encourages skipping: E-stop and reset, jam recovery, power cycle and restart from a loaded conveyor.
- Record the final parameter set and program version, and store the backup with the as-commissioned values.
Working Alone Early Is Normal; Being Left Alone Is Not
Site work three weeks in is routine in integration - the field is where the learning is, and there is no substitute for it. Being scheduled solo for seven of ten commissioning days on a system you have not seen before is a resourcing decision, not a test of your competence, and it should be named as one to your manager in writing. Ask directly for coverage on the days that need two people. That request is a normal part of project execution, and the engineers who look competent from the outside made the same call at the same point in their careers.
Frequently Asked Questions
Why does a conveyor drive trip during acceleration but run fine once it reaches speed?
The load inertia demands more torque during the ramp than at steady state, and the drive hits its current limit before the belt reaches commanded speed. Read the accel time parameter and the output current logged in the drive's fault history, then lengthen the ramp and retest before assuming a motor or mechanical fault.
Why does the PLC output LED light up while the motor stays off?
The LED reflects the state the processor wrote to the output, not the presence of field power downstream of it. Measure voltage at the output terminal and across the group fuse; a blown fuse or a missing field supply on that output group produces exactly this symptom.
Why does remote phone support usually fail to resolve a commissioning fault?
The remote engineer has no meter and no view of the machine, so the call spends its time re-asking what has already been checked. Prepare a written ruled-out list, the drive fault history, and a screen share of the online logic at the stopping rung before dialing, and the same call converges in a fraction of the time.
Why should escalation happen before the contractor's deadline rather than after it?
Before the deadline, escalation is a schedule decision your project manager can act on with travel time, resources, or a sectional handover. After it, the same information is a failure report with no remaining options.
Stop troubleshooting and escalate when a fixed cutoff hour passes with the fault unisolated, when two remote sessions with the same engineer fail to converge, or when the next step needs a second person at the field device while you watch the panel. For a drive or controller behaving outside its documented function, open a case with the manufacturer's technical support with the model number, firmware revision, serial number, and the captured fault history in hand. Route the schedule consequence to your project manager the same day - not to the contractor.