Install procedure · A2L (R-454B) split systems

Start-up & board programming

The commissioning order, start to signature: circuit integrity, safeties, the board, airflow, every call proven, charge at high stage, and the sheet. Set the machine below — the steps adapt. This is the procedure the start-up drills on the board practice against.

0/32 steps

This guide teaches the company process on a reference machine. The unit in front of you outranks it: the install manual, the data plate, and the chart on the blower door govern every actual value. When this page and the door disagree, the door wins — every time.

A2L rules — in force for the whole job

R-454B is class A2L — mildly flammable. Real, and manageable: it takes a concentrated leak plus a strong ignition source.
No open flame near the circuit. Nitrogen purge while brazing, and braze BEFORE the charge is in the lineset.
A2L-rated leak detector on the job, and rated recovery equipment for any removal.
Blends charge as LIQUID — throttle it in; vapor-charging a blend fractionates it.
The refrigerant detection system (RDS) is a life-safety device. It gets TESTED at start-up. It never, ever gets jumpered.

Before any power — circuit integrity

0/8

Everything in this phase is cheaper than everything after it. Nothing gets energized until the refrigerant circuit is proven tight, dry, and documented.

Photograph the data plates (outdoor + indoor) and the blower-door chart into the job file. Note the measured lineset length.The plate drives the trim math, the breaker check, and the charge targets; the door chart drives the board programming. Ten seconds of photos ends every later argument.Braze with dry nitrogen flowing through the circuit. No exceptions on an A2L machine.Scale inside the lineset eats the metering device on any refrigerant — and open flame plus a flammable-class circuit is a rule you never bend.Pressure-test with dry nitrogen to the MANUAL’S test pressure. Log start psi, temperature, and time; hold and re-check corrected for temperature.A pass is a flat line once temperature is accounted for. On A2L equipment "close enough" is not a category — a leak you charge into becomes a fuel drop plus a warranty fight.Evacuate to 500 microns or below, then isolate the pump and decay-test per the manual. If it keeps climbing, hunt moisture or a leak — do not charge into a rising vacuum.The decay test is the honesty check: a deep number with the pump running proves the pump, not the system. Isolated-and-holding proves the system.Open the service valves — liquid first, then suction — releasing the holding charge into the proven-tight, evacuated lineset. Full back-seat per the manual; torque the stems and caps and leak-check both.The factory charge lives behind these valves, and a compressor started against closed valves is a scroll running on a sealed circuit in deep vacuum. Every OEM start-up sheet carries this line for a reason — and the caps are the final seal, not decoration.Seat the RDS sensor harness at the coil and route it clear of panel edges BEFORE the panel goes on.The unplugged RDS harness is the #1 first-day lockout: blower runs, cooling refuses, board flashes, homeowner calls. Thirty seconds now beats a truck roll Friday.Slope the drain, trap it per the manual, and wire the float switch into the Y circuit. It gets function-tested after power-up.The float is the ceiling’s only friend. Wired-but-never-tested is the same as not wired.Torque lugs to spec, strain-relieve the whip, verify breaker size against the nameplate MCA/MOCP, land Y1/Y2/O/C per the diagram, filter in, registers open.MCA sets the wire, MOCP caps the breaker — the nameplate is the law here, not the old breaker that happens to be in the panel. Y2 unlanded is the “can’t keep up” callback; O unlanded is heat in July.

First power — safeties before comfort

0/4

The machine earns the right to cool by proving its protections first.

Energize the air handler first. Verify 24V control power and let the board finish its self-check.Control power proven before line-voltage loads means every later surprise is a load problem, not a mystery.Energize the outdoor unit now — no cooling call yet — and give the crankcase heater the manual’s soak time before the first compressor start. Plan the rest of the phase around it.Refrigerant migrates into cold compressor oil; the first start slugs it out. The heater needs hours, not minutes — energizing it early means the soak finishes while you program the board and prove the calls.Run the RDS self-test per the manual and confirm the board reports SENSOR OK. Know the lockout behavior: on a sensor fault or a real detection, the board drops cooling and runs the blower to disperse.You are proving the safety exists and answers. If it ever locks out in service, that is the system working. The RDS is never jumpered — a bypassed leak sensor on a flammable-class charge is how a callback becomes a news story.Lift the float. Confirm the cooling call actually drops. Put it back.Ten seconds. The alternative is discovering the float is decorative during a drain backup over a finished ceiling.

Program the board

0/7

The board ships with SHIPPING settings, not YOUR design. This phase is where "installed" becomes "engineered." Chart in one hand, settings in the other.

Open the blower-door chart (you photographed it in phase 1) and find the row for THIS outdoor/indoor match and tonnage.Boards are programmed from the chart for the match — never from memory, never from the last job, never from the sim’s reference table.Set the airflow profile the chart calls for — dip switches or config menu per the manual. Write the profile letter and target CFM on the start-up sheet.Factory default is a shipping setting (often a smaller-tonnage profile). A 4-ton match left on a 3-ton profile runs quiet, soft, and one humid week from an iced coil.Apply CFM trim only if the design or the static reading (next phase) calls for it — per the manual’s trim table.Trim is a scalpel for real duct conditions, not a volume knob for noise complaints.Configure dehumidification behavior if the thermostat and design use it.Comfort in humid climates is a latent job; the board can slow the blower to wring the coil — but only if someone turns it on.Thermostat setup menu: set the equipment type — heat pump with O energized in cooling — and the correct stage count. Then verify it SENDS what it claims: R-to-O and R-to-Y2 measured at the sub-base.Wires prove nothing about configuration. A stat menu still set “furnace” on a heat pump blows heat in July with every conductor landed beautifully. Jumper tests and sub-base voltage are the truth.Two-stage: enable second-stage cooling (Y1/Y2) in the stat and confirm the staging strategy (temperature/time-based) per design.Two-stageA two-stage machine on a single-stage brain runs at 65% forever — the "brand new and can’t keep up" callback in its natural habitat.Heat pump: set the defrost interval/curve on the defrost board per the manual for this climate.Heat pumpWrong defrost config is invisible in July and a no-heat complaint in January. Set it while the panel is open and you are already here.

Reference airflow chart — the sim’s machine

Profile A300 CFM/tonshipping default on many boards — NOT a design choice

Profile B350 CFM/tonhumid-climate / high-latent designs

Profile C400 CFM/tonstandard design point for most matches

Profile D450 CFM/tondry-climate / high-sensible designs

This table belongs to the simulator’s reference machine (it is the same chart its board panel shows). On the job, the chart on the blower door of the actual unit — for the actual match — is the only chart that exists.

Prove the airflow

0/2

Charge numbers taken at wrong airflow are fiction. Airflow gets proven BEFORE the gauges matter.

Measure total external static pressure and compare against the design and the chart’s CFM-at-static table.Static is the duct system confessing. High static means the profile you programmed is not the CFM you are delivering — fix restrictions (filter, registers, duct) before touching charge.Confirm delivered CFM is within the design window for the match (400/ton standard unless the design says otherwise). Record CFM and static on the sheet.This number underwrites every SH/SC reading you are about to take and every capacity promise the company just sold.

Prove every call the stat can make

0/4

Every output the thermostat owns gets demonstrated once, with your meter, before charge work.

Call first-stage cooling. Contactor in, compressor running, condenser fan running — measured, not assumed.The baseline call. Everything else is proven relative to this.Step to high stage. Hear/feel the unloader shift and watch amps and pressures step up. Verify 24V at the unloader solenoid on the Y2 call.Two-stageTHE two-stage start-up moment. If Y2 dies anywhere — stat config, unlanded wire, dead staging coil — the machine "works" at 65% until the first design-temperature afternoon, and then it is your callback.Heat pump: cycle O — confirm the reversing valve clicks and the supply flips cold-to-hot and back. Verify 24V at the RV solenoid in cooling.Heat pumpThe valve rests in heat; O is what makes summer work. Proving the click and the flip now is what keeps “brand new heat pump blows heat” off next week’s board.Set the stat to OFF. Confirm everything actually stops.Welded contactors and stuck relays announce themselves right here, on your schedule, instead of at 2 a.m. on the homeowner’s.

Charge — high stage, by the plate

0/4

The factory charge covers the base lineset only. The rest is arithmetic, a scale, and the plate’s targets — at full capacity and proven airflow.

Compute the trim: (measured lineset − base length on the plate) × oz/ft from the plate. Write the math on the sheet.Fifteen base feet is typical; a forty-foot run is a real deficit. “Close enough from the factory” is a system that spends its first summer running long and its first callback blamed on the equipment.Weigh the trim in as LIQUID, throttled, with the scale doing the judging. A2L rules in force: detector on the job, no ignition sources, rated gear.R-454B is a blend — vapor charging fractionates it. The scale is the only instrument that cannot be argued with.Run at HIGH stage and design airflow; verify subcooling to the plate target (TXV) and sanity-check superheat. Then drop to low stage and confirm the numbers still make sense there.Two-stage charge targets live in HIGH stage — a light charge reads almost textbook on low and confesses on high. Charging on low stage is the trap; this step is the cure.Clamp compressor and fan amps against RLA/FLA on the plate. Record them.Day-one amp draws are the baseline every future tech will wish they had.

Prove & paper — the start-up sheet

0/3

The sheet is not paperwork. It is the order of operations, the callback shield, and the baseline the next tech inherits.

Record: model/serials, lineset length + trim added, airflow profile + CFM + static, RDS self-test PASS, every call proven (Y1 / Y2 / O / idle), pressures + SH/SC with indoor/outdoor temps, amp draws.A start-up sheet with real numbers turns every future "it never worked right" into a fifteen-second lookup.Walk the owner: filter location and schedule, thermostat basics, and what an RDS lockout looks like (cooling stops, blower runs — call us; nobody bypasses it).The sixty-second orientation prevents the three most common first-month calls and teaches the household that the safety behaving IS the system working.Sign the sheet. Both stages proven, valve proven both ways, safeties tested — or it does not get signed today.Sign it like you mean it: the signature says the machine was PROVEN, not that the air felt cold on the way out.

Practice these exact traps on the board’s start-up wing — it drills the unplugged RDS, the factory-default profile, the unlanded Y2, and the charged-on-low-stage callback against live physics.