• FR
• NL
• EN

Pitfalls #5 – Limits to speed

The short answer: because the drive is never the first thing to fail.

Why 3,000 rpm is a reference point — not an arbitrary limit

On a 2-pole asynchronous motor running at 50 Hz, synchronous speed is 3,000 rpm. In practice, accounting for motor slip, the fan runs at around 2,900 rpm under load. This is the speed at which the fan manufacturer established its performance curves, optimized its best efficiency point (BEP), and defined its mechanical warranties.

Operating at synchronous speed therefore means operating within the envelope the machine was designed and validated for. That reference speed isn’t necessarily fixed at 3,000 rpm / 50 Hz, either: it can just as well be a higher speed — 60 Hz, for instance — if that’s what the equipment was specifically sized and validated for from the outset, to meet a particular technical requirement. What matters is staying within the design envelope that was actually retained. Beyond that reference speed, you enter a zone that isn’t covered by warranty by default — and several limiting factors start working against you at once.

The three real limiting factors

1. Wheel tip speed

This is most often the first limiting factor — and the most dangerous one to overlook.

Centrifugal stress in a fan wheel increases with the square of rotational speed. Doubling the speed quadruples the stress on the material. Each manufacturer sets a maximum tip speed depending on the wheel material:

• Welded steel: up to ~120–140 m/s depending on grade
• Cast aluminum: typically 80–100 m/s
• Polypropylene or GRP: often ≤ 50–60 m/s

Tip speed is calculated simply as: v = π × D × n / 60 (D in meters, n in rpm).

A 500 mm diameter wheel running at 3,500 rpm reaches a tip speed of ~92 m/s. On an aluminum wheel, that’s right at the limit. On a polypropylene wheel, it’s well beyond it.

Exceeding the manufacturer’s limit isn’t just a matter of degraded performance — although that’s part of it too: increasing rotational speed often pushes the fan toward a degraded point on the curve (running off the end of the curve, for example), with aerodynamic performance suffering as a result. But beyond that degradation, the real risk is wheel failure. The manufacturer may void any warranty and, on certain equipment, require a calculation certificate confirming the wheel’s mechanical integrity at the requested speed — a point to confirm case by case with the manufacturer — for any overspeed request.

2. Absorbed power — the cube law

Fan affinity laws leave no room for argument:

• Flow ∝ n
• Pressure ∝ n²
• Absorbed power ∝ n³

A 10% increase in speed results in a 33% increase in absorbed power. At +20%, that’s +73%.

In practice: an 11 kW fan selected at 2,900 rpm will need ~15 kW at 3,200 rpm, and ~19 kW at 3,480 rpm. If the motor hasn’t been resized accordingly, it will trip — either on thermal protection or on the drive’s own safeguards.

There’s also an often-underestimated effect on top of this: above 50 Hz, the motor operates in flux-weakening mode. Magnetic flux decreases to maintain voltage, and the maximum available torque drops proportionally. At precisely the moment the wheel demands more torque (higher speed, greater power), the motor has structurally less torque available. The safety margin shrinks from both sides at once.

3. Acoustics — the fifth-power law

This is the factor most often underestimated, and frequently the one that settles the debate on projects with noise constraints.

Acoustic power radiated by a fan scales approximately with n⁵. A 10% increase in speed translates into a sound power level increase of roughly +6 to +7 dB(A).

As a reminder, +3 dB(A) corresponds to a doubling of acoustic energy. +6 dB(A) means quadrupling it. On an industrial site with regulatory constraints or third-party requirements, this increase can render an installation non-compliant — even if it was perfectly within limits at nominal speed.

In summary

So the question isn’t “can the drive go up to 55 Hz?” but rather “what’s the first limiting factor on this machine, and how much margin is actually left?”