One compressor is easy: it loads when pressure falls, unloads when pressure rises. The moment you install a second one, you have a new problem that didn't exist before: who runs, when, and how do they decide?

Get that wrong and the compressors quietly fight each other. Plants with multiple compressors and no coordinated control routinely waste 10–25% of their compressed air energy on nothing but bad sequencing — several part-loaded machines doing the work of one, wide pressure bands, and unload running around the clock.

Why multiple compressors at all?

Good reasons to split capacity across machines:

  • Redundancy — one machine down doesn't stop the plant.
  • Demand range — production needs 40 m³/min on the day shift and 6 m³/min at night; no single fixed-speed machine covers both efficiently.
  • Growth — capacity got added over the years, one compressor at a time.

That last one is the most common — and it's why so many multi-compressor rooms are a collection of individuals rather than a system.

The classic failure modes

Everyone part-loaded. Three compressors all running at 50–60% load. A fixed-speed compressor running unloaded still draws roughly 20–30% of its full-load power while delivering no air — so three part-loaded machines waste up to twice the unload losses of one. The efficient state is always: as few machines as possible fully loaded, one machine trimming, everything else off. (Unload Running Explained covers what unload time costs.)

The cascade creep. With simple cascaded setpoints (see below), each compressor needs its own slice of the pressure band. More compressors → wider total band → the highest setpoint drifts upward. Every extra 0.1 bar costs about 0.7% in energy, so a band that grew to 1.5 bar is an invisible surcharge on every operating hour.

The trim war. Two machines with overlapping setpoints alternately loading and unloading against each other, cycling far more than demand requires. Each compressor's own controller is doing its job perfectly; nobody is doing the system's job. (Your Dumb Air Compressor explains why individual compressor controls can't see the system.)

The control strategies

Cascading setpoints (simple, fine for 2–3 machines)

Each compressor gets its own load/unload band, staggered: compressor 1 at 6.8–7.0 bar, compressor 2 at 6.5–6.7, compressor 3 at 6.2–6.4. Falling demand pressure brings machines in one by one.

It needs no extra hardware and it works — at the price of a wide total pressure band and a fixed starting order. Review it twice a year: cascades drift as setpoints get "temporarily" adjusted.

Master controller / pressure-band control

A central controller holds the whole system inside one narrow pressure band (often ±0.1–0.2 bar) and decides which machines run. This is the standard answer for 3+ compressors:

  • One band instead of a staircase → the average pressure drops, often by 0.3–0.7 bar.
  • The controller picks the best machine combination for the current demand, rotates duty so wear spreads evenly, and steps in machines only when genuinely needed.

Base load + VSD trim

The refinement that makes the biggest difference in mixed rooms: fixed-speed machines do what they're good at (running flat out, fully loaded) and a variable-speed compressor handles the trim — the fluctuating remainder. Sized right, the VSD absorbs the demand swings and nothing cycles.

Sized wrong, it's the opposite: a VSD that's too small rides at its limits while fixed machines cycle around it. The trim machine must cover the gap between demand steps, and its regulation range has to span the fluctuation.

What good control looks like

You can judge a compressor room in ten minutes:

  • One narrow pressure band, not a staircase of setpoints.
  • Running machines are loaded. Nothing idles unloaded for long stretches — machines either work or stop.
  • Enough storage. A generous air receiver slows the system down and gives the control strategy room to work — storage and sequencing succeed or fail together.
  • Night and weekend behaviour matches demand — small demand, small machine, not the 90 kW unit trimming for a handful of leaks.

If your room fails that inspection, the fix is rarely a new compressor. Measure first (DIY assessment), then fix the control strategy — it's usually the cheapest large saving in a multi-compressor plant. If a controls problem is acute (pressure swinging, machines hunting), start at troubleshooting: poor control strategy.