13.2 Compressor Sequencing
Key Takeaways
- Lead/lag assigns a role: the lead starts first and should stop last; lag machines start only after running machines are truly full and a suction or kW deadband has been violated for a delay.
- Equalize runtime by rotating the lead role on hours—not by starting extra machines or short-cycling a lag package so the hour meters look even.
- Unload the slide toward minimum and prove amps/position before stopping; a loaded stop slams checks, dumps a capacity step into suction, and makes the next start a stall risk on wye-delta or soft-start.
- Winter sequences should shed lag machines earlier so you are not running several screws at low slide; summer sequences keep more machines available but still start the next one only on a sustained deadband.
- Wye-delta and soft-start packages must be proven unloaded before the start contact, and large motors should be staggered so inrush and demand intervals do not stack.
13.2 Compressor Sequencing
A single screw cannot cover every hour of a production year. Sequencing is how the plant decides which packages run, which is lead, which is lag, when the next machine starts, and how a machine leaves the line without wrecking suction, oil, or the electric bill. CIRO items in this area are supervisor decisions: equalize hours without short-cycling, bring the next machine on a suction or kW deadband, and unload before stopping.
Lead and lag
The lead machine is first to start and, in a well-written sequence, last to stop. Lag machines start only after the lead is no longer able to hold suction (or kW) within the deadband. On a falling load, lag machines should unload and stop before the lead is asked to shut down. That keeps one machine in a decent slide range instead of three machines hunting at 25%.
Lead is not the newest compressor and not the one nearest the door. It is a role the controller assigns. Rotate that role so run hours stay comparable. Equalizing runtime spreads oil changes, alignments, and overhauls. It does not mean starting a second 300 HP package for an hour so the hour meters look pretty. If you need to catch up hours, swap lead at a shift change or after a planned stop—not by adding an extra running machine in mid-winter.
When to bring the next machine
Start logic should require a sustained need, not a spike. Typical permission to start the next screw:
- Suction pressure (or suction temperature) has been above setpoint plus deadband for a time delay, and
- Running machines are near full slide (or VFD speed) so there is no unused capacity already online, or
- A kW or remaining-capacity calculation shows the running fleet cannot meet the load (some plants use evaporator counts or a kW target).
The deadband exists so a dock door, a defrost recovery, or a brief pull-down does not start a lag machine that will then stop twelve minutes later. Short-cycling destroys motors, couplings, and oil. Minimum run timers, minimum off timers, and anti-recycle timers are part of sequencing, not optional extras. If the plant hunts, widen the deadband or fix the slide/VFD that is not actually loading before you add another start.
kW-based sequencing is useful when suction is a poor proxy—multiple suction levels, a VFD already holding pressure, or a demand-charge strategy. The idea is the same: a band, a delay, and proof that machines already running are contributing. Do not start the next motor because instantaneous kW flickered during a wye-delta transition on reduced-voltage starting.
Unload before stopping
A screw at 100% slide is a large, oil-filled gas pump. Stopping it loaded dumps a step change into suction (capacity disappears at once), slams the discharge check, and asks the next start to accelerate against a full slot. Unload first: command the slide toward minimum, wait until amps fall and the position feedback agrees, then stop. The same rule applies before a lead/lag swap.
Unloaded stops also protect oil. A sudden stop at full differential can push oil where the separator did not expect it and can foam the sump on the next start. Your mechanical-integrity walk-downs will see the difference as vibration, coupling wear, and mystery oil loss.
Winter versus summer sequences
| Season | Typical load and head | Sequence emphasis | Efficiency trap |
|---|---|---|---|
| Summer | High wet bulb, high product load, floating clamp rarely binding | Keep lag machines available; start the next one only after lead is full plus deadband delay | Treating every defrost-recovery suction bump as a start command |
| Winter | Lower load, head should float, defrost still needs a floor | Shed lag machines earlier; fewer packages at higher slide | Three screws parked at ~30% slide instead of one healthy lead |
Summer: wet bulb is high, the floating-head clamp is rarely the binding constraint, product load is high, and you may run most of the fleet. Sequence on suction/kW; keep lag machines off until the lead is truly full. Defrosts are frequent; do not interpret a defrost-recovery suction bump as a reason to start the last reserve machine without a delay.
Winter: load is lower, head should float, and the efficiency trap is too many machines at low slide. Two screws at 35% often draw more kW/ton than one machine at 80% because of slide bypass (next section). Winter sequences should shed lag machines earlier, raise the stop-lag slide threshold, and keep a hard eye on the floating-head minimum so defrost and oil cooling still work. Some plants use a seasonal schedule: fewer enabled packages, longer min-run, wider deadband.
A winter plant that parks two lag machines and lets the lead float head will almost always beat a sequence that keeps everything in auto for reliability at 20% slide each.
Soft-start, wye-delta, and what sequencing must prove
Starting method changes what start the lag machine is allowed to mean.
Wye-delta: the motor starts in wye at reduced torque and current, then transitions to delta. A loaded screw may not accelerate in wye. The motor sits near locked-rotor current, the transition slams torque, and voltage on a weak feeder dips. Sequencing must prove slide at minimum (limit switch, analog position, or a verified unloaded amp band) before the start contact. Stagger starts so two large packages do not transition in the same few seconds. Open-transition wye-delta is a brief disconnect; closed-transition is kinder to the bus—know which you have.
Soft-start: current is ramped. A loaded start still may time out on stall if the slide is at 100% or the discharge check is stuck closed. The sequence should not treat a failed start as try again in 30 seconds without unloading and investigating.
VFD: can start with more torque control and can be the capacity device (VFDs in refrigeration plants). Sequencing implications: do not across-the-line bypass-start a VFD machine at the same instant as a wye-delta lag; watch harmonics and demand; a VFD machine can often trim load so you delay starting the next across-the-line screw. After a power failure, staged restart—utilities, pumps, condenser, then compressors unloaded, then load—protects both the process and the utility demand interval.
Demand charges bill the highest kW in an interval. Starting two 300 HP motors in the same interval can dwarf a month of floating-head savings. Supervisor-level sequencing includes start staggering and whether this start is needed before the interval rolls.
Equal hours without equal abuse
Hour equalization should compare running hours and, if the controller supports it, starts. A machine with 6,000 hours and 400 starts is not equal to one with 6,000 hours and 80 starts. Sequences that short-cycle the lag machine to share hours make the start problem worse. Prefer: rotate lead weekly; inhibit rotation during pull-down or during a known defrost group; never rotate by starting an extra machine.
When you walk a CIRO-style screen, ask: which package is lead, what is each slide %, what is suction versus deadband, and did the last start happen after a timer or after a blip? If amps on the lag machine never left the unloaded band, you started it for no capacity—fix the slide or the start permissives before you add more logic.
A lag 300 HP screw is about to leave the line on a falling box load. What should the sequence do first?
The lead screw has been at 95–100% slide for 12 minutes and suction is still above setpoint plus deadband. What is the correct reason to start the next machine?
Why is a winter sequence that keeps three screws at about 30% slide usually worse than one machine at high slide with the lag packages parked?
A lag screw with wye-delta starting is called to start while its slide feedback still reads 100%. What is the sequencing implication?