Key takeaways
- A four-cylinder set at 1,500 rpm fires at 50 Hz; a six-cylinder at 75 Hz. A panel whose first mode sits within about 15% of that figure will drum no matter how thick the lining is.
- A simply supported 600 × 400 mm panel in 2 mm CRCA has a first mode near 44 Hz, which is why that panel size drums on a four-cylinder set and is quiet on a six.
- Doubling panel thickness raises stiffness eightfold but mass only twice, so the first mode rises by a factor of two; a 20 mm swage or a folded edge gets most of that for far less weight.
- A 5 mm gap at a panel joint passes more low-frequency sound than 2 mm of steel blocks, so seal the joints before adding mass.
- Keep the first panel mode above 1.3 times the firing frequency, or add constrained-layer damping; stiffening alone cannot fix a mode that has already been excited.
A canopy can meet its noise limit at 1 m and still sound wrong. The overall reading sits at 74 dB(A), the specification is 75, and the panel facing the alternator booms on every load step. The number passes and the customer rejects the set, because the sound is a tone rather than a level.
Canopy drumming is panel resonance. The engine fires at a fixed frequency, the panel has a natural frequency of its own, and when the two line up the panel moves further than the small pressure wave that drives it. The radiated sound rises by 10–20 dB at that frequency, the panel fatigues at its fixings, and the lining does nothing to stop it because the lining is on the inside of the moving surface.
The fix is a mode calculation, a stiffness decision and a sealing check, in that order. This guide covers the numbers for each, the failure modes that look like drumming but are not, and what to put on the canopy drawing. The wider set of canopy decisions sits in the genset and canopy hub.
What frequency does a canopy panel drum at?
It drums at the engine firing frequency or one of its harmonics, when the panel's first bending mode sits close to it. For a four-stroke engine, firing frequency is f = (rpm ÷ 60) × (cylinders ÷ 2).
At 1,500 rpm, a four-cylinder set fires at 50 Hz and a six-cylinder at 75 Hz. At 1,800 rpm, the same engines fire at 60 Hz and 90 Hz. The generator end of the set runs at those speeds because the alternator needs 50 Hz or 60 Hz electrical output, so the firing frequency is fixed by the mains, not by the canopy.
The second excitation is the cooling fan. Blade-pass frequency is (rpm ÷ 60) × blade count. A fan at 1,500 rpm with seven blades passes at 175 Hz, which is high enough that it usually excites a higher panel mode rather than the first.
The third is structure-borne vibration through the base frame. A panel bolted to a frame that transmits engine order vibration behaves as a driven plate, and the mode shape tells you where it moves most. That is why a stiffener placed in the middle of a panel does more than one placed near an edge: it interrupts the mode shape where the displacement is largest.
How do you estimate a panel's first mode?
Use the simply supported rectangular plate formula as a first check. It is conservative for a bolted canopy panel, because real edges are partly clamped and clamping raises the frequency. The formula is:
f₁ = (π/2) × √(D ÷ (ρh)) × [(1 ÷ a)² + (1 ÷ b)²]
where D = Eh³ ÷ (12(1 − ν²)), E is Young's modulus, h is thickness, ρ is density, ν is Poisson's ratio, and a and b are the panel dimensions.
For steel, E = 210 GPa, ρ = 7,850 kg/m³ and ν = 0.3. Take a 600 × 400 mm panel in 2 mm CRCA. D works out at 154 N·m and ρh at 15.7 kg/m², so √(D/ρh) = 3.13 m²/s. The bracket term is (1/0.6)² + (1/0.4)² = 2.78 + 6.25 = 9.03, and f₁ = 1.571 × 3.13 × 9.03 = 44 Hz.
That panel sits less than 15% below the 50 Hz firing frequency of a four-cylinder set, and it will drum. The same panel on a six-cylinder set at 75 Hz is 41% away and behaves.
Now take a larger panel: 1,000 × 800 mm in 1.5 mm. D = 64.9 N·m, ρh = 11.8 kg/m², √(D/ρh) = 2.35, bracket = 1.00 + 1.56 = 2.56, and f₁ = 9.4 Hz. That is well below 50 Hz, so it is not directly excited by the firing frequency. It is still a problem, because the second and third modes land near 30–40 Hz and the panel responds to road and transport vibration as well as to the engine.
Panel geometry, mode and the Ganesh limit
The table collects the values that decide whether a canopy panel drums, and the limit we work to when a drawing asks for a quiet panel.
| Parameter | Typical value | Ganesh limit or note |
|---|---|---|
| Sheet thickness | 1.2–2.0 mm CRCA | 1.5–2.0 mm for acoustic panels; 2.0 mm on the alternator face |
| Panel size, unsupported | 600 × 400 mm to 1,000 × 800 mm | Keep the smaller dimension under 700 mm where a 50 Hz set is specified |
| First mode target | Above 1.3 × firing frequency | 65 Hz minimum on a 1,500 rpm four-cylinder set |
| Stiffener | 20 mm swage or folded rib | One swage on the long centreline; two on a panel above 800 mm |
| Stiffener depth | 15–25 mm | 20 mm formed on the 200 T brake; deeper needs a different fold |
| Fixing pitch | 200–300 mm | 250 mm on acoustic panels; gasketed joint at every fixing |
| Gasket compression | 25–35% of free thickness | Closed-cell EPDM, 3–5 mm free thickness |
| Lining | 25–50 mm acoustic foam or mineral wool | 50 mm on the alternator face, 25 mm elsewhere; retention by the OEM spec |
| Damping | Constrained layer where a mode cannot move | Loss factor 0.1–0.3 against 0.001–0.01 for bare steel |
| Leak area | Zero at joints | A 5 mm gap over 1 m is 5,000 mm² of leak; seal it before adding mass |
| Coating | 60–90 µm powder | Coating mass on a panel shifts the mode by under 1% |
The mode target is the row that decides the design. A first mode at 1.3 times the firing frequency keeps the panel out of the resonant peak even with some clamping uncertainty and some build variation. Below about 1.15 times, the mode is inside the half-power band of the excitation and the panel responds as if it were tuned to it.
Doubling thickness is the obvious lever and the wrong one. D scales with h³, so 2 mm sheet is eight times stiffer than 1 mm, but ρh doubles, and the first mode rises by √(8/2) = 2. So 2 mm raises the 44 Hz panel to 88 Hz, clear of a 50 Hz set. The cost is 100% more steel and a heavier panel that needs two people to handle. A 20 mm swage on 1.5 mm sheet can raise the same mode by 30–60% for a fraction of the weight, which is usually enough to clear 50 Hz.
Which failure modes look like drumming?
Five, and only the first is the panel mode itself.
Panel resonance
The panel moves at its own frequency and radiates a tone. The cure is a mode shift or damping. A mode shift comes from a swage, a folded edge, a deeper pan or a shorter unsupported span. Damping comes from a constrained layer: a viscoelastic film between the steel and a second thin constraining sheet, which converts the bending motion into heat. A bare steel panel has a loss factor around 0.001–0.01; a constrained layer takes it to 0.1–0.3, cutting the resonant amplitude by an order of magnitude.
Air leaks at joints
A gap passes low-frequency sound almost without attenuation. A 5 mm gap running 1 m along a panel joint has 5,000 mm² of open area, and at 50 Hz the wavelength is 6.9 m, so the gap is acoustically small and radiates efficiently. Sealing that joint with a compressed gasket does more for the firing-frequency band than doubling the lining thickness.
Loose fixings and fretting
A panel held on M6 bolts at 500 mm pitch deflects between the bolts, and the bolt holes fret. The panel is then both a drum and a rattle, and the rattle excites the mode. Fixing pitch at 250 mm and a gasket at every fixing removes the lost motion.
Lining that has come away
Acoustic foam bonded to a panel inside a canopy sees 80–100 °C and constant vibration. An adhesive that was adequate at 20 °C softens, the foam drops, and the panel behind it is now bare and free to resonate. Mechanical retention, not adhesive alone, is the specification.
Base frame and engine mount transmission
If the engine mounts are hard or the base frame is flexible, the canopy sees structure-borne vibration that no panel stiffness will fix. A set that drums on a skid but not on a test bed is showing a mount or frame problem, and the canopy is the amplifier rather than the source.
How do stiffness and damping fix canopy drumming?
They solve different halves of the problem, and the choice between them is a decision rule.
Move the mode when the panel is lightly damped and the target frequency is known. A swage or rib is cheap, adds little weight, and is formed on the same brake as the panel. Aim for a first mode at least 1.3 times the highest firing frequency the set will see.
Add damping when the mode cannot move. A panel that is already at the maximum size, or a set that runs at variable speed and sweeps through the mode, cannot be tuned out. A constrained layer works across a band, and it also reduces the structure-borne noise the panel radiates at other frequencies.
Do both when the panel is large and the target is tight. A 1,000 × 800 mm alternator-side panel on a 50 Hz set needs a swage to get near 50 Hz and a constrained layer to take the peak down. The weight penalty is about 1.5 kg/m² for a 0.5 mm constraining sheet and a 0.5 mm viscoelastic core.
The cost order is worth stating: a formed swage costs almost nothing at the brake, a constrained layer adds material and labour, and a thicker panel adds weight to every handling and freight calculation for the life of the product. Try them in that order.
What the canopy build does at Ganesh
Acoustic canopy panels are laser cut on the 12 kW fibre laser on the 40 ft × 16 ft bed, which lets a full side panel with its swage reliefs and fixing pattern come off in one piece. Panels are formed on the 200 T, 5-axis brake with the 4,450 mm bed. A 20 mm swage runs along the long centreline, and the panel edges get a return fold that acts as a stiffener and as a gasket land.
Frames and panel stiffeners are welded in fixtures built in our tool room, on the three CLOOS robotic MIG cells or the manual stations. Weld sequence matters on a large panel: a continuous run along one edge pulls the panel out of flat, and a flat panel that has been pulled will not seal against a gasket. Stitch welds on alternating sides hold the shape, and the panel is checked on a surface plate before coating.
Coating is seven-tank pretreatment, then powder at 60–90 µm, measured on the face and at the edges. The lining is fitted to the OEM's acoustic specification, because the insertion loss figure is certified against that build and not against a generic foam. Mahindra Powerol is the canopy reference on our customer list, and the panel families are on the genset parts page.
The manufacturing route for the enclosure as a whole, including the frame, doors and louvres, is covered in how a genset acoustic enclosure is manufactured. Panel bending and long-panel flatness sit in bending for genset canopies and panels.
What the canopy drawing has to settle
- Engine speed and cylinder count, so the firing frequency is on the record and the panel modes can be checked against it.
- Panel size and unsupported span, with the stiffener position and depth.
- Fixing pitch and gasket type, including the compression range and the fastener isolation.
- Lining specification, with the retention method and the temperature rating.
- Damping, where a constrained layer is required, with the core thickness and the constraining sheet.
- Sealing details at every joint, door and cable entry, because the leak path is part of the acoustic design.
- Third-octave noise target, not only an overall dB(A) figure, so a tone cannot hide inside an average.
When the fix is not a panel change
Not every drumming complaint is a canopy problem. If the set drums at one speed and not another, the excitation is narrowband and the panel is the amplifier; check the mode first. If it drums at all speeds, look at the engine mounts, the base frame stiffness and the alignment of the alternator, because a misaligned coupling puts a once-per-revolution force into the frame that the canopy cannot absorb.
If the set is installed indoors with a ducted inlet and outlet, the duct can act as an organ pipe and the canopy panel is not the source at all. Measure at the panel surface and at the duct mouth before changing the panel.
We do not measure sound power or certify a canopy to a noise standard. That is the OEM's test to ISO 8528-10 or ISO 3744, and the acoustic consultant owns the lining specification. What we do is build the panel to the mode and stiffness the design asks for, hold the flatness and sealing tolerances, and say when a panel size cannot meet the mode target with the thickness drawn.
Send the panel drawing, the engine data and the lining specification through request a quote, and the mode check comes back with the quotation rather than after the first set is rejected on tone.
Frequently asked questions
At what frequency does a genset canopy drum?
At the engine firing frequency and its harmonics. A four-cylinder four-stroke at 1,500 rpm fires at 50 Hz, and the second order at 100 Hz. Panels with a first mode between 40 and 60 Hz are the ones that boom.
Does thicker sheet stop canopy drumming?
It raises the panel mode, but not as much as stiffening. Doubling thickness doubles the first mode. A 20 mm swage or a folded edge can raise it by more, at a fraction of the weight.
Will acoustic foam fix a drumming panel?
Foam absorbs airborne sound inside the canopy but does not damp the panel itself unless it is bonded as a constrained layer. A panel driven at resonance needs damping or a mode shift, not more absorption.
Why does the canopy drum only at certain engine speeds?
Because the excitation is narrowband. At idle the firing frequency is below the panel mode; at rated speed it coincides. A set that drums at 1,500 rpm and is quiet at 1,200 rpm is showing a resonance, not a loose panel.
Can a drumming canopy be fixed on site?
Sometimes. Replacing missing gaskets, tightening panel fixings and adding a stiffening rib between existing bolts can move a mode. A panel that needs a new swage or a lining change comes back to the works.
Sources and further reading
- ISO 8528-10 — Generating sets, measurement of airborne noise by the enveloping surface method · reviewed 3 Nov 2026
- ISO 3744 — Determination of sound power levels using an enveloping surface over a reflecting plane · reviewed 3 Nov 2026
- Central Pollution Control Board — noise limits for diesel generator sets · reviewed 3 Nov 2026
Chasing a drumming canopy panel?
Send the panel drawing, the engine speed and cylinder count, and the lining specification. We will check panel modes and stiffener positions before the panel is cut.
Request a quote