What a Compressor Control Strategy Is Choosing Between

Why this matters

Two identical compressors, same frame, same motor, same rated delivery, can differ by a large margin in what they cost to run on the same plant, and the only difference is how the control decides what to do when the plant wants less air than the machine makes. Nobody sells a compressor on that. The nameplate carries a full-load figure, and full load is where a compressor spends the smallest part of its life. Understanding what a control strategy is trading lets you tell a customer why their machine is expensive without touching it, and it tells you which strategies their storage volume will even permit.

The three things every strategy trades against each other

Every control scheme is a position on the same three-way trade, and no scheme wins all three.

• How tightly it holds pressure. A narrow band means the plant sees a steadier supply and the average pressure can be set lower, which matters because every fixed restriction in the system, including leaks and unregulated tools, passes more flow at higher pressure.
• What it draws at reduced output. The power a machine takes when the plant wants half its air varies enormously between schemes. This is the term that dominates the annual bill, and it is covered in depth under part-load behaviour in this library.
• What it costs the machine in starts and cycles. Motor starts are thermally and mechanically expensive, manufacturers publish a maximum starts per hour, and unloading has its own minimum time before the machine reaches its low unloaded power.

The third of those is the one that constrains the other two, because it is set by the storage volume on the system rather than by anything inside the compressor. That is the point most people miss: your available storage decides which trades you are allowed to make.

The schemes, and what each one is actually doing

Start and stop. The motor stops when pressure reaches the top of the band and restarts at the bottom. Zero power at zero output, which is unbeatable, and it is the correct choice for small machines with modest duty. The limit is the motor's published maximum starts per hour, and that limit converts directly into a minimum cycle time and therefore a minimum storage volume for a given band.

Load and no-load, also called on-line and off-line. The motor runs continuously; the inlet valve closes and the machine blows down its internal pressure so it is turning against very little. Power falls to a fraction of full load once blowdown is complete, and manufacturers publish that fraction. The catch is that blowdown takes time, and if the unloaded period is shorter than the blowdown period the machine never reaches its low power. A load/no-load machine on inadequate storage draws close to modulation power while calling itself load/no-load.

Inlet modulation, sometimes called throttling. A valve progressively closes the inlet so the machine takes in less air per revolution. It holds pressure very tightly and it is the worst scheme for power at reduced output, because the machine is still turning at full speed against a compression ratio that has gone up, not down. At zero delivered flow a modulating machine without blowdown still draws a large majority of its full-load power; read the number off the machine's published curve rather than assuming.

Variable displacement. Turn valves, spiral valves and stepped porting change the effective compression volume, so the machine does less work rather than throttling. Better part-load power than modulation, usually available only over part of the range, and then it hands off to another scheme below that.

Variable speed. The drive changes shaft speed so displacement per minute follows demand. Power tracks flow closely inside the drive's turndown range, and below the turndown floor the machine has to stop or blow off. Covered separately in this library, including what it does not fix.

Inlet guide vanes with blow-off, on dynamic machines rather than positive displacement ones, hold pressure by vane position down to a surge limit and then vent the excess. Below that limit the machine is making air it throws away, and the power does not fall.

The comparison

Read the pressure-band column together with the power column. Modulation gets specified because it holds the tightest band on the least storage, and that is a real benefit, but it is bought in the column that dominates the annual bill.

The measurement that decides which schemes are available to you

You do not need a storage calculation to find out how much storage a system has. You need a stopwatch and one pressure gauge.

With the plant isolated and the compressor loaded, time the header rising through the control band. With the compressor unloaded and the plant drawing its normal load, time it falling through the same band. Those two times are the system's storage expressed in the only units that matter for control, and they let you compute the cycle rate directly.

Two conditions on that measurement, both of which change the answer if you get them wrong. Take both times across the same two pressure marks, or the comparison is meaningless. And take the falling time at a stated plant demand, because the fall time is a function of what the plant is drawing, not a property of the system alone.

Worked example: choosing a scheme for a plant that already has one

A single-machine plant, one shift, a 10 psi control band, and a demand that sits near 60 percent of the machine's rated delivery for most of the shift. The machine currently modulates.

Take the two times. Loaded with the plant drawing its normal 60 percent load, the header rises through the band in 68 seconds. Unloaded at that same 60 percent draw, it falls through the band in 45 seconds.

Cycle rate. Total cycle is 68 plus 45, which is 113 seconds, so 3,600 divided by 113 is 32 cycles per hour.

First check, and it eliminates one scheme immediately. Start and stop at 32 starts per hour is far above the maximum starts per hour published for a motor of this size. Look the number up on the nameplate or in the manual rather than guessing it; the point is that the storage on this system does not support that scheme at this band. Start and stop is out.

Second check, on load and no-load. The unloaded window is 45 seconds. Compare that against the machine's published blowdown time. If blowdown completes well inside 45 seconds, the machine reaches its low unloaded power for most of the unloaded window and the scheme works. If blowdown takes longer than 45 seconds, the machine never gets there, and load/no-load on this system would deliver something close to modulation power while cycling the inlet valve 32 times an hour on top of it. That single comparison is the whole decision.

What the loaded fraction tells you, and it is a useful cross-check. On a load/no-load machine the loaded fraction of the cycle equals demand divided by capacity, because that is the only way the average delivery matches the average draw. Here 68 divided by 113 is 60.2 percent, which matches the stated 60 percent demand. If your measured times do not reproduce your believed demand, one of the two is wrong, and it is usually the believed demand.

The power comparison. Say the machine's published unloaded power is 0.25 of full load; substitute your machine's figure, because this varies by design and by how the blowdown is arranged. Blended power under load/no-load is 0.60 times 1 plus 0.40 times 0.25, which is 0.70 of full load. Under modulation, read the 60 percent point off the machine's published curve; on typical published curves for inlet throttling it lands well above 0.80. The difference between those two figures is what the plant is currently paying for a tighter band than it needs.

What would flip this. Two things, and both are about the unloaded window rather than about power.

If the plant's demand moves toward the middle of the machine's range rather than sitting at 60 percent, the cycle rate rises, because cycle rate is highest when demand is near half of capacity and falls toward zero at both ends. At 50 percent demand the same storage gives a shorter cycle and a shorter unloaded window, which can push a machine that passed the blowdown check into failing it.

And if the band cannot be widened, because the top of it is already at the highest pressure the system is allowed to see, then storage is the only remaining variable. Adding receiver volume lengthens both times proportionally without moving the band, which is the cleanest way to make load/no-load work on a system that is currently forced into modulation.

The failure mode. Converting a machine to load/no-load on paper, without measuring the unloaded window against blowdown time, and then reporting a saving the plant never sees. The customer's bill does not move, the compressor now cycles its inlet valve tens of times an hour, and the next person to look at it has a machine wearing out under a control scheme that was never appropriate for the storage available.

Checking your own conclusion

Re-time the band after any control change, at the same two marks and at a stated demand, and confirm three things: the cycle rate is inside the machine's published limit, the unloaded window exceeds the published blowdown time, and the loaded fraction of the cycle still reproduces the demand you believe the plant has. If the third one has drifted since the last check with no equipment added, you are looking at a loss problem rather than a control problem, and this library covers that trace separately.

References

• Manufacturer documentation for the specific machine: rated delivery, published part-load power curve, unloaded power fraction, blowdown time, and maximum motor starts per hour
• 29 CFR 1910.147, lockout and control of stored energy before any work that opens the compressor or the pressurised side; 29 CFR 1910.333(b)(2) for work inside the starter or drive enclosure
• See related: Why Part Load Behaviour Decides the Energy Bill; How to Sequence Multiple Compressors Without Short Cycling; What Storage Buys You That Horsepower Cannot