Pro Audio and Subwoofer Drivers: Bonds Under Thermal Load

A 15-inch pro audio woofer running a two-hour power test does not fail because the copper gave up. Copper melts at 1,085 °C and the coil never gets close. What gives up is a polymer — the enamel on the wire, the former it is wound on, or the adhesive holding the two together. That is the uncomfortable part of pro audio subwoofer driver adhesive thermal load: on a high-power driver, the adhesive is usually the lowest-temperature material in the motor, which means adhesive selection, not wire gauge, sets your published power-handling number. If you buy adhesives for a driver line and your technical data sheet has no temperature column, you are specifying the ceiling by accident.

Where the heat actually goes

Nearly all the electrical power into a moving-coil driver becomes heat in the voice coil. Efficiency for a typical pro woofer sits in the low single digits as a percentage, so a driver rated at several hundred watts continuous is a several-hundred-watt heater with a small acoustic side effect. That heat leaves the coil by three routes: conduction into the former and up the leads, radiation and convection across the magnetic gap into the pole piece and top plate, and forced convection if the design has vents or a phase plug.

Two consequences matter when you are writing an adhesive spec.

First, the coil is the hot end and everything bonded to it lives at coil temperature. The bond between coil wire and former, the collar bond where the former meets the cone, and the lead exit are all in the hot zone.

Second, the motor structure itself heats up. Steady-state top-plate and pole-piece temperatures on a hard-driven pro driver run far above ambient, which puts the magnet-circuit bond — the one holding top plate, magnet and yoke in alignment — under sustained warm load. It never sees coil temperature, but it sees a temperature it was probably never asked about.

You can back out coil temperature yourself without a thermocouple. Measure DC resistance cold at a known ambient, run the driver, then measure DC resistance again within seconds of switching off. Copper's resistivity has a temperature coefficient of roughly 0.00393 per °C, so the resistance ratio gives you the average coil temperature rise directly. That is also the mechanism behind power compression: as resistance climbs, the driver draws less power from a constant-voltage amplifier and gets quieter. A few decibels of compression at rated power is unremarkable on a pro driver, and the number your customer complains about is the same number that tells you how hot your adhesive is.

The bond that sets the ceiling, not the wire

Wire enamel systems are specified by thermal class, and the good ones are rated well above what most adhesives tolerate. Formers follow the same logic: a polyimide former survives temperatures that will soften a general-purpose epoxy, and an aluminium former conducts heat away faster than either but needs an insulating treatment and behaves differently on thermal expansion. Line those materials up against a structural adhesive with an unstated glass transition temperature and it is clear which one runs out first.

Understand what actually happens above Tg, because the failure is rarely dramatic. The bond does not melt. It softens — modulus can drop by more than an order of magnitude across the transition — and then it creeps under the load it is carrying. On a coil-to-former bond, creep means turns move, the winding loosens, and the coil starts to buzz or rub before it ever comes apart. On a magnet-circuit bond, creep means the gap concentricity shifts by a fraction of a millimetre, which is enough on a driver with a tight gap. Both present to your customer as a mechanical noise complaint, not a thermal one, so the root cause gets attributed to the coil — the voice coil bonding walkthrough covers that misdiagnosis in more detail.

Three specifics worth knowing before you read anyone's data sheet:

Tg depends on how it was measured. A Tg from differential scanning calorimetry (DSC, midpoint) and a Tg from dynamic mechanical analysis (DMA, tan-delta peak) can differ by 10–20 °C on the same cured sample, with DMA reading higher. A supplier quoting the flattering number is not lying, but you are not comparing quotes on equal terms. Ask which method and which point on the curve.

Tg is not continuous service temperature. Continuous service temperature is set by thermal ageing — how much strength the bond keeps after hundreds or thousands of hours at temperature — and it usually sits below Tg for a structural joint under sustained load. A single Tg number with no ageing data tells you where the modulus falls off, not how long the bond lives there.

Cure state changes Tg. An epoxy pulled from the oven early is under-cured, and an under-cured epoxy has a lower Tg than the same chemistry fully cured. If your line is running a shortened cure to hit takt time, the thermal rating on the data sheet no longer applies to the parts you shipped. That is why cure schedule and thermal specification have to be negotiated together, not in separate conversations.

What each bond has to survive

One line on that table deserves a paragraph. Adhesive squeeze-out into the magnetic gap is a genuinely common pro-audio defect and it is almost never diagnosed correctly. Gap clearance on a large-format driver is a fraction of a millimetre per side. A bead that migrated during cure sits there harmlessly at room temperature, then everything expands under two hours of power and the driver rubs. The warranty return looks like a coil failure. The cause was rheology — a grade too thin, or applied at a viscosity that dropped when the oven ramped. When you evaluate a magnet-circuit adhesive or a centring grade, ask about viscosity at cure temperature, not just at 25 °C.

Test the way your customer will test

If your driver ships with a published continuous power rating, someone will eventually test it against a standard. AES2 is the usual reference for professional drivers and defines a long-duration test with band-limited pink noise at a specified crest factor, run for two hours; IEC 60268-5 covers loudspeaker measurement more broadly, including input-power definitions. Check the current edition of whichever one your customer names, because these documents are revised and the revision changes what you are agreeing to.

The practical point is that an adhesive qualification that runs shorter than the standard proves nothing about the standard. A bond that holds through a 20-minute burst is not evidence for a two-hour rating. Neither is a room-temperature lap-shear number, which is the number most adhesive data sheets lead with.

A qualification sequence you can actually run, in order:

1. Fix the cure schedule first. Bond samples using the oven temperature and dwell your line will really use, including the worst case at the cold end of your oven. Every later number is only valid for that cure state.
2. Measure room-temperature strength on the actual substrate pair — coil wire enamel to former material, or plated steel to magnet — not on standard aluminium coupons. Enamel and plating dominate adhesion far more than the bulk chemistry does.
3. Repeat the strength test hot. Take the joint to your expected steady-state bond temperature and test at temperature. A shear number that survives to temperature is worth more than any Tg claim.
4. Age it. Hold bonded samples at temperature for a meaningful number of hours, then retest cold and hot. Retention after ageing is what separates a grade that lives at temperature from one that merely survives arriving there.
5. Run the full driver test. AES2 or whatever your customer specifies, full duration, with DC resistance measured before and immediately after so you know the coil temperature you actually reached.
6. Cut the driver open. Section the motor and inspect the coil, the gap and every bond line. Squeeze-out, discolouration and delamination at the collar all show up here and nowhere else.
7. Repeat on a second production batch of adhesive before you release. Batch-to-batch behaviour is the thing sample testing is worst at predicting.

Step seven has a commercial counterpart. Adhesive is qualified at the driver level, but it is bought as a consumable, and a formulation change under the same product code is invisible to your receiving inspection. Ask in writing for notification before any formulation, raw-material or supply-source change, and put it in the purchase agreement. Buyers who have been through this once usually also read what documents to request from a supplier before the first order with different eyes afterwards.

Where this supplier sits, honestly

Aosibo Adhesives in Huangjiang, Dongguan lists the loudspeaker bond schedule as its core range — lead-wire, damper, surround, magnet-circuit and centring adhesives — with centring offered in one-part, epoxy and halogen-free versions, and it names driver makers and acoustic-module contract manufacturers among its applications. That is the right shape of catalogue for the discussion above, and the company states that viscosity, cure schedule, pot life and substrate compatibility are shared at inquiry rather than published.

What that means for you is simple and worth saying plainly: no thermal data for these grades was published or provided at review, so nothing on this page should be read as a temperature rating for any product. Continuous service temperature, Tg, ageing retention and whether a dedicated high-temperature grade exists at all are open questions you have to ask directly. The listing itself is published with the corporate-registry check still open, which is a separate reason to run your own document checks early rather than late.

Common questions

Does a higher Tg adhesive always give a higher-power driver?

No. Tg tells you where modulus falls away; it says nothing about strength retention after a thousand hours at temperature, nothing about toughness, and nothing about whether the grade wets your particular wire enamel. A high-Tg epoxy that is too brittle for the collar joint will crack under excursion long before heat matters. Specify thermal performance alongside the mechanical duty of each individual bond.

Should I use one adhesive for the whole hot end?

Rarely. The coil bond wants heat resistance and wetting on enamel; the lead exit wants flexibility retained after ageing; the magnet circuit wants gap-filling shear on plated steel. Those are different property sets. Buying them from one supplier as a matched set is reasonable, and is how driver lines usually buy; specifying one product for all of them is not.

My driver is neodymium — does that change the adhesive question?

It changes the constraint order. Neodymium magnets have a maximum operating temperature that varies by grade and by the working point of the circuit, and on many designs that limit sits below what a good structural adhesive can tolerate. If the magnet caps the motor temperature anyway, your magnet-circuit adhesive has an easier job than it would on a ferrite motor that is allowed to run hotter. Ask your magnet supplier and your adhesive supplier the same temperature question and see which answer is lower.

What if the supplier has no thermal data at all?

Then you generate it, or you do not use the grade in a hot bond. Ask for the composition class, the recommended cure, and a written statement of maximum recommended service temperature, and then verify it on your own samples using the sequence above. A supplier who says "send me both substrates and your cure window" is answering the right question; a supplier who quotes a temperature with no method behind it is answering the wrong one.

What to ask the supplier next

Send this as one message, before you request pricing:

• Glass transition temperature for each grade, with the test method and the point on the curve — DSC midpoint or DMA tan-delta peak.
• Maximum continuous service temperature, and the thermal-ageing data behind it: hours, temperature, and percentage strength retained.
• Lap-shear strength measured at elevated temperature on your actual substrate pair, not on aluminium coupons at 25 °C.
• Viscosity at 25 °C and at the recommended cure temperature, plus any thixotropic index, so you can predict squeeze-out.
• The full cure schedule, and what happens to Tg if cure is shortened to fit a faster line.
• Any outgassing or condensable-volatiles measurement, since deposits in the gap are a rub mechanism of their own.
• A written formulation-change notification commitment, and the batch traceability you will receive with each shipment.
• The safety data sheet, the UN number and packing group, and shelf life from date of manufacture — thermal specification is useless if the drum arrives already half-aged.

Aosibo Adhesives is one example of a supplier whose stated range covers the whole driver bond schedule, so the thermal questions above can be asked bond by bond in a single conversation rather than across four vendors; the factory listing sets out what was and was not verified before you make contact. If you have never written a specification tight enough to survive this kind of qualification, the walkthrough on writing a product spec sheet for a Chinese factory is a reasonable place to start.