Voice Coil Bonding: Where Heat Kills the Adhesive First

Voice coil bonding is the one place in a loudspeaker driver where the adhesive sits inside the heat source, and the heat resistance of the bonds at the former and collar — not their room-temperature strength — is what decides how much continuous power the driver can actually take. Every other bond in the assembly is warmed by the coil. The bonds in and around the coil are the coil.

This is the selection criterion buyers most often miss, and the reason is understandable. Adhesive data sheets lead with a shear strength figure measured at 25 °C, because that is the number that looks impressive and is cheap to generate. A driver rated for 100 W continuous does not spend any of its working life at 25 °C. It spends it at a winding temperature that a data sheet from a general-purpose supplier was never asked about.

The coil is a heater with a magnet around it

A direct-radiator loudspeaker converts a small fraction of its electrical input into sound. The rest becomes heat, and almost all of it becomes heat in the winding. Under a continuous programme the coil reaches a steady temperature set by the balance between that dissipation and the conduction path out through the former, the gap air and the magnet structure. On a pro-audio or subwoofer driver run hard for hours, coil temperatures of a couple of hundred degrees Celsius are routine, and the adhesive within a millimetre of the wire sees every degree of it.

Several bonds sit in that zone:

• The material holding the turns of the winding to each other and to the former — either a self-bonding coating on the wire, activated during winding, or an applied impregnating adhesive.
• The bond between the coil and the former, whether the former is polyimide film, paper, glass-fibre composite or aluminium.
• The collar — the reinforcing ring at the junction of former, cone neck and often the spider — which carries the mechanical load path from the coil into the moving assembly.
• The anchoring of the tinsel lead where it exits the former, which is hot as well as flexing.

Aosibo Adhesives' stated range is lead-wire adhesive, damper (spider) adhesive, water-based foam and rubber surround adhesive, magnet-circuit adhesive and centring adhesive in three versions. A dedicated voice-coil or collar grade is not listed separately. That is a question, not a conclusion: the collar bond is often made with the same grade used for the centring or damper joint, and the honest way to open the conversation is to ask which grade the supplier intends for that joint and what its thermal rating is, rather than assuming either that a special product exists or that one does not.

Why room-temperature shear strength is the wrong number

Structural adhesives for this duty are thermosets, and a thermoset has a glass transition temperature — Tg — where its behaviour changes character. Below Tg the polymer is glassy and stiff. Above Tg it becomes rubbery. It does not melt and it does not turn to liquid, which is why suppliers can honestly say the bond "still holds" above Tg. What collapses is stiffness: the storage modulus can drop by two to three orders of magnitude across the transition.

For a bond holding a slab of metal to another slab of metal, a soft matrix is survivable. For a voice coil it is not. The winding is held in shape by that matrix. Once the matrix goes rubbery, the turns can move under the magnetic force acting on them, the coil goes out of round, it contacts the pole or the top plate, and then it rubs, chars and opens. The failure sequence is entirely a stiffness failure, and no room-temperature shear number predicts it.

So the selection criterion is Tg, plus a continuous service temperature that accounts for long-term ageing, and shear strength measured at temperature rather than at ambient. Three details make that request precise.

Ask which method produced the Tg. Differential scanning calorimetry, dynamic mechanical analysis and thermomechanical analysis give different numbers for the same material, often 10–25 °C apart. Within DMA alone, the onset of the storage-modulus drop and the peak of tan delta can differ by a similar margin, with the tan-delta peak reading higher. A quoted "Tg 130 °C" with no method attached cannot be compared between two suppliers, and the safest interpretation is the lowest one — the storage-modulus onset — because that is where stiffness starts leaving.

Ask for retained strength, not just a rating. The useful figure is shear strength at 120 °C, at 150 °C, at 180 °C, expressed as a percentage of the room-temperature value on the same substrates. A grade holding 60% at 150 °C and one holding 15% at 150 °C can carry identical data-sheet headlines.

Ask what cure state the Tg was measured in. Tg depends on the degree of cure, and an undercured thermoset has a substantially lower Tg than the same material fully cured. A line that runs a 30-minute oven pass on an adhesive whose schedule wants 60 minutes at temperature can be building drivers with a Tg tens of degrees below the specification, and nothing on the production line will detect it. If the supplier's cure schedule is longer than your available oven window, the answer is a different adhesive or a different oven — not a shortened bake.

Find your own number, then demand theirs

Specifying Tg without knowing your coil temperature is guesswork. The measurement is straightforward and worth doing before any adhesive conversation.

The winding is copper or aluminium, and the resistance of both rises predictably with temperature — for copper, by approximately 0.393% per degree Celsius around room temperature. So: measure the coil's DC resistance cold, with the driver at a known ambient. Drive it with your rated continuous programme until the resistance stops climbing, which is thermal equilibrium. Measure the resistance hot. The ratio of hot to cold resistance, divided by the temperature coefficient, gives you the temperature rise. This is the basis of the standard industry power-handling test methods, and the number it produces is the number your adhesive has to survive, continuously, for years.

Then add margin. Two kinds:

• Ageing margin. Tg tells you where stiffness goes. It does not tell you about oxidation and embrittlement over thousands of hours. A continuous service temperature rating is a separate figure and is usually well below Tg.
• Assembly margin. The former has its own limit. Polyimide film tolerates far more than paper; an aluminium former conducts heat away from the winding but expands more than the adhesive does. The assembly is limited by whichever component gives first, so there is no value in an adhesive rated 40 °C above the former it is bonded to. The heat also travels outward: the magnet-circuit bond sits in the conduction path and gets a share of the same thermal duty.

The thermal fields to demand

No technical data sheet was available for any grade in this range at review, and the supplier's own position is that viscosity, cure schedule, pot life and substrate data are shared at inquiry. That makes the table below a request list rather than a comparison — but it is the request list that separates a serious reply from a brochure.

Sealed micro-drivers add a second failure mode

Earphone and true-wireless drivers turn the problem sideways. The absolute power is small, so peak coil temperature is less of a concern, but the enclosure is sealed and tiny, and anything the adhesive releases has nowhere to go. Volatiles that evaporate during or after cure condense on the diaphragm, in the gap, or on the acoustic mesh — and a few micrograms in a 10 mm driver is a measurable change in response, not a cosmetic one.

Aosibo lists headphone and earphone assembly as an application, describing it as micro-driver bonding where cure temperature and outgassing matter. That framing is right, and it points at two questions rather than one. Ask for outgassing or condensable-volatiles data, in whatever test form the supplier has. And ask separately for the lowest cure temperature available across the range, because the second constraint on a micro-driver is that the parts around the bond are small, thin and easily deformed by an oven pass that a woofer would shrug off.

A test plan that fits a one-week sample

Sampling is stated at 7 and production at 15, given as bare numbers on the company's application and read as days. That is enough time to eliminate a wrong adhesive, which is most of the value. Test the thermal bonds alongside the rest of the schedule rather than in isolation — each adhesive applied later has to cure below the limit of everything already in the assembly, so the centring adhesive's cure window and this one are a single constraint, not two. Run it in this order:

1. Bond coupons in your real materials. Your former stock, your wire with its actual coating, your collar material. Generic substrates tell you about the adhesive; your substrates tell you about your driver.
2. Cure on the schedule you can actually run, not the one the supplier recommends. If those differ, run both and compare — the gap between them is your risk.
3. Test in hot shear. Bring the coupons to your measured continuous coil temperature and test there. Compare against a room-temperature set from the same batch.
4. Soak a set at 1.2× your measured coil temperature for as long as the week allows, then test cold. This finds embrittlement that a short hot test misses.
5. Build a small number of complete drivers and run a power-compression test to equilibrium, taking coil temperature by resistance rise before and after.
6. Measure rub and buzz, and second-harmonic distortion, before and after the power test on the same units. A coil that has started to move shows up here before it shows up as a failure.
7. Retain cured samples so a third-party laboratory can run DMA later if you decide to qualify the grade, and so you have a reference when a future lot behaves differently.

Build this into a broader sample process rather than treating it as a one-off: the general discipline of requesting a sample from a Chinese factory without getting burned applies with extra force to chemicals, where the sample and the production lot can differ invisibly.

Common questions

Is a higher Tg always the better choice?

No. Raising Tg usually means raising crosslink density, which tends to make the cured adhesive stiffer and more brittle, and often raises the cure temperature required. On a collar joint that also has to tolerate flex, or in a micro-driver where the cure has to stay gentle, an adhesive with more thermal headroom than you need can be the wrong answer. Match Tg to your measured coil temperature plus margin, then optimise the other properties inside that constraint.

The supplier quotes a shear strength but no Tg. What does that tell me?

It tells you the grade has probably not been characterised for this duty, which is common for general-purpose products and less common for adhesives sold specifically into speaker assembly. It is not automatically disqualifying — ask directly for Tg, method and hot shear, and see what comes back. A supplier who can answer within a day or two has the data; one who answers with reassurance about quality does not.

Which bond gives first in a voice coil assembly?

It depends on the design, which is why the question is worth asking of your own drivers rather than in general. Common candidates are the matrix holding the winding turns, then the coil-to-former bond, then the collar. The lead exit is a separate case because it combines heat with flex fatigue. Autopsy your own field returns and warranty units before choosing where to spend engineering effort — the answer is usually consistent within a design family.

What to ask the supplier next

For the collar and voice-coil bonds specifically, send this and treat the shape of the reply as information:

1. Which grade in the range is intended for the collar and voice-coil joints, and whether a dedicated grade exists.
2. Tg for that grade, with the measurement method named, and for DMA, which marker.
3. Continuous service temperature, peak temperature and the duration assumed for peak.
4. Shear strength at 120, 150 and 180 °C as a percentage of the 25 °C value, on your former material and wire coating.
5. Shear retained after 1,000 hours at the continuous rating.
6. The full-cure schedule, plus the Tg achieved if cured at your maximum available oven dwell.
7. Outgassing or condensable-volatiles data, and the lowest cure temperature available across the range, if any part of your business is sealed micro-drivers.
8. Shelf life from date of manufacture, storage temperature, and a minimum remaining shelf life on arrival written into the contract.
9. The unit behind the stated 1,000-unit minimum order — bottle, cartridge, syringe or kilogram — with pack size and price basis, since that stated figure means nothing until the unit is defined. Comparing suppliers on minimum order quantities and lead times only works once the units match.

Send both substrates and the cure window you have available, as the supplier's own guidance puts it, and send your measured coil temperature with them — that single number changes the conversation from a catalogue exchange into an engineering one. The Aosibo Adhesives profile records what the company states about its loudspeaker range and its terms, and which product data was not documented at review, so you can see in advance which of the nine items above is still an open question rather than a lookup.