Runs When Jumpered But Not In Service Decision Tree
Why this matters
You jumper across a control input and the system runs. You remove the jumper and the system will not run in normal service. This is one of the cleanest diagnostic signatures in field work: it tells you the load and the wiring are healthy, and the fault is in the input you bypassed or upstream of it. This decision tree gives you the isolation order so you can name the failed element instead of just observing the symptom.
This is fundamentally the same isolation method described in NFPA 70B for control troubleshooting and reflects the systematic-substitution approach to fault diagnosis.
Symptom presentation
- Jumper the thermostat terminals at the air handler, blower runs. Reconnect thermostat, no blower.
- Jumper across a pressure switch, burner fires. Reconnect switch, no fire.
- Jumper the float, sump runs. Reconnect, no run on rising water.
- Jumper a remote sensor input on a controller, equipment runs. Reconnect, lockout.
- Jumper a dry-contact demand from a building management system, output operates. Reconnect, no response.
In every case the jumper replaces a control input with a forced closure, and the equipment behaves as if that input were calling. The fault is in why the real input is not calling.
Quick checks
- Confirm the jumper is making at the same two points the real input would make. A jumper at the wrong terminals proves nothing.
- Identify what the bypassed input is supposed to do. Is it a demand input (calling for run) or a permissive input (allowing run)?
- Identify the type of input: dry contact, voltage signal, current signal, sensor with a value range, network message, switch with a polarity.
- At the input device itself, measure what the system controller actually sees. The contact may be making mechanically and reading as open electrically if there is corrosion, broken solder, or a wire issue.
- Read any panel diagnostics that tell you which input the controller is waiting for.
Isolation tree
Stage 1: Is the real input device closing/sending at all?
At the input device, force its physical condition (push the float, satisfy the pressure, raise the temperature, send the network message) and check whether it transitions. Examples:
- Thermostat: drop setpoint well below room temperature; watch terminals for the expected continuity or millivolt change.
- Pressure switch: vary the system pressure across the switch's trip point; watch for continuity transition.
- Float switch: lift or lower the float manually; check for transition.
- Sensor: warm or cool with a heat gun or ice pack; check reading on the controller display.
- Network input: send the expected command and read the controller's input register.
If the device does not transition: the device is failed or out of adjustment. Address the device.
If the device does transition correctly: proceed to Stage 2.
Stage 2: Is the transition reaching the controller?
Read the same signal at the controller input terminals (or input register for networked devices). If the device transitions but the controller sees nothing, the fault is in the wiring or signal path: broken conductor, loose terminal, shield grounded incorrectly, transformer or signal isolator failed in the path, network address conflict, baud-rate mismatch.
For dry contacts, ohm the conductor with the device disconnected at both ends. For voltage signals, check at intermediate junction points. For 4-20 mA loops, check loop power and burden resistor. For network signals, check at the bus level.
Stage 3: Is the controller acting on the input?
If the device transitions and the controller sees the transition but the load still does not run, the controller's logic is rejecting the input. Possible reasons:
- Lockout flag still set from a prior fault, masking the current call
- Parameter set wrong (input is configured as a different type)
- Permissive chain has another input still in the wrong state
- Schedule or time-of-day clock has the equipment in a no-run window
- Network priority override is holding the equipment off
- Firmware bug or corrupted parameter
Read every input the controller is waiting on, not just the one you bypassed. The jumper passed your input; the controller may have been waiting on a second one you did not bypass.
Stage 4: Is the jumper actually the only thing changed?
Be honest about what the jumper did. If your jumper bypassed a safety chain element along with the demand input, your test forced a different condition than just satisfying the demand. Re-bypass only the actual demand input and confirm.
Confirming the diagnosis
The proof is reversibility: identified fault element addressed, real input now operates the system without the jumper. If you addressed the input device and the system still requires the jumper, you have not isolated to the right fault. Common second faults: the original device is fine but its wiring is broken; the device and wiring are fine but the controller has a stuck lockout; a second permissive input is also open.
Always remove the diagnostic jumper before final verification. A system that "works" with the jumper still in place is not a confirmed repair.
Do not leave a jumper installed in service. Diagnostic jumpers across safety chains, demand inputs, or permissive interlocks must be removed before the equipment is returned to operation. A field unit running with a safety-chain jumper is a liability event waiting for an inspector or a failure. Visual inspection of every wire you touched is the last step before you leave.
Next steps
Document the input device that failed, why it failed (mechanical, electrical, corrosion, broken wire, misadjustment, configuration), and the parameter or setting that may need adjustment going forward. Inputs that fail because of an environmental condition (corrosion from chemical atmosphere, vibration loosening a terminal) will fail again unless the environmental cause is also addressed.
If the customer or owner expects continued service before parts arrive, do not leave with a jumper. Find a different temporary arrangement (manual control under supervision, partial shutdown, alternate equipment) and write it up.
References
- NFPA 70B Recommended Practice for Electrical Equipment Maintenance, control circuit isolation
- ISO 13379-1 Condition monitoring and diagnostics of machines, data interpretation and diagnostics techniques
- OSHA 29 CFR 1910.147 Control of hazardous energy
- OSHA 29 CFR 1910 Subpart S, Electrical safety standards
- ACCA Standard 4 Maintenance of Residential HVAC Systems