New Install Won't Start: Commissioning Checklist Tree
Why this matters
A brand-new install that will not start is a different animal from an old unit that quit. The part is good. The factory tested it. So the fault is almost always something you or the installing crew did, did not do, or set wrong during the install. That changes the order you look in. Chase it like a worn-out-component failure and you will pull good parts and waste an afternoon. Walk the commissioning tree instead and most no-starts fall out in the first ten minutes, before you ever open a meter.
Safety first on any energized or pressurized start
Before you troubleshoot a unit that will not start, make it safe to work on.
- Electrical: confirm the disconnect is where you think it is, then de-energize and verify dead at the terminals with a meter before you put hands inside. A new install can be fed from a circuit you did not expect.
- Gas or fuel: if you smell fuel, stop, ventilate, and find the leak before you energize anything. A spark near a fuel-air mix is the hazard, not the no-start.
- Pressure or stored energy: relieve and verify zero on any line, tank, or capacitor before you open it.
Only once the unit is safe do you start the tree below.
Start here: is it actually getting power and a call to run
The most common new-install no-start is not the equipment. It is upstream of it.
- Is the circuit on? Check the breaker or fuse feeding the unit. A new circuit may never have been turned on, or it tripped on the first attempt and nobody reset it.
- Is the disconnect closed? Local disconnects get left open after the install for safety and forgotten.
- Is there a call to run? Confirm the control, thermostat, switch, or float is actually asking for operation. A new control wired but not configured will sit silent and look like a dead unit.
- Did anyone confirm voltage at the unit's terminals? A measured reading beats an assumption every time. No voltage in means the problem is the feed, not the box.
If there is no power or no call, you have found the family. Trace the feed or the control wiring. If power and a valid call are both present and it still will not start, move on.
Next: was it set up, not just installed
Installed and commissioned are not the same thing. A unit can be perfectly mounted and wired and still refuse to run because a setup step was skipped.
- Configuration or address not set. Many modern controls ship in a neutral state and do nothing until configured for the application. An unconfigured board can look identical to a dead one.
- A shipping or service lockout still engaged. Shipping bolts, transit locks, service switches, and protective shrouds get left in place. The unit reads them as "do not run."
- A safety waiting on a precondition. A flow switch with no flow, a pressure switch with no pressure, a float with no level, a door switch with the panel off: each one is doing its job and blocking the start. Confirm the precondition exists before you call the switch bad.
- Polarity, phase, or rotation wrong. On anything direction-sensitive, reversed connections can prevent operation or trip a protection instantly. Verify against the install documentation.
Then: is a built-in protection holding it off
New equipment protects itself aggressively, and a fresh install trips those protections more than a settled one.
- A time delay or anti-short-cycle lockout. Some units enforce a wait after power-up. It is not broken, it is counting. Confirm the delay before you condemn the control.
- A high or low limit tripped on the first attempt. Air or debris in a new line, a closed isolation valve, a blocked intake: any of these can trip a limit the instant you energize, then the unit locks out and looks dead. Find the limit that tripped and ask why.
- A reset that was never pressed. A manual-reset device tripped during install stays tripped until someone clears it.
Read the fault indicator if the unit has one. A blinking code or a latched lockout points you straight at the family instead of guessing.
If it still will not start: now suspect the rare stuff
Only after power, call, setup, and protections all check out do you suspect actual hardware. Even then, a new install fails on the install far more than on the part.
- A connection backed out or never landed. A lug that looks seated but is not, a quick-connect not fully pushed home, a ribbon cable one row off.
- Damage during the install. A pinched wire, a cracked board from over-torqued mounting, a line nicked while being routed.
- A wrong or incompatible accessory. A control, sensor, or kit that does not match the unit and will not let it run.
If you reach a genuine part failure on a new install, document it before you swap, because that is a warranty claim, not a normal repair.
Recap: the order that catches install errors
- Make it safe: de-energize and verify, or relieve pressure, first.
- Power and a valid call to run, both measured, not assumed.
- Setup done: configured, locks removed, preconditions met, polarity and rotation correct.
- No protection holding it off: delays, limits, manual resets.
- Only then suspect a connection, install damage, or a genuinely failed part.
The rule to bank: on a new install, the part is innocent until the install is proven clean.
References
- Manufacturer installation and commissioning documentation for the specific unit
- NFPA 70 (National Electrical Code) for feeder, disconnect, and overcurrent basics
- Trade-standard commissioning and startup practice
- See related: The Startup Sequence That Catches Install Errors; Right Part, Wrong Setting on Startup (decision tree)