How OEM Customers Work With a Micro Speaker Manufacturer on a Custom Design
1. Why Custom Development Is a Joint Activity
Short answer: Neither side can specify a custom speaker alone. The customer holds the constraints the driver cannot see; the manufacturer holds the behaviour the customer cannot predict.
The division of knowledge is the reason collaboration is necessary rather than merely convenient. The customer's engineering team knows things no component datasheet can carry: the total thickness the product cannot exceed, which internal volumes are already committed to a battery or an antenna, what the amplifier can actually deliver, how far the listener will be, and what else makes noise at that position. The manufacturer knows things the customer cannot derive: how a given compliance interacts with a given cavity volume, which parameter changes need new tooling and which do not, and how much resonance a diaphragm or winding adjustment will actually buy.
Most stalled programmes are stalled because one side made a decision that belonged to the other. A customer who specifies a resonance frequency without stating the cavity volume has made a decision about a system they have not yet defined. A manufacturer who proposes a driver without asking what else shares the back volume has made a decision about a structure they have not seen. Both are recoverable, but only if the missing input is identified as missing rather than assumed.
2. Who Owns Which Decision
Short answer: Four decisions belong to the customer, four to the manufacturer, and three are genuinely joint. Writing down which is which before the first meeting removes most of the later friction.
Table 1 assigns the decisions that come up in a custom speaker programme. The pattern is that the customer owns everything about the product and the manufacturer owns everything about the component, and the genuinely joint decisions are the ones where those two meet.
Table 1: Decision ownership across a custom micro speaker programme.
Decision | Owner | Why it sits there | What happens when it is assumed instead |
Total thickness and mounting height | Customer | Only the product structure knows, including tolerance stack and cover | A driver that fits the drawing but not the assembly |
Cavity volume the structure can seal | Customer | Depends on keep-out volumes and sealing, not on the driver | An installed resonance above target, discovered after tooling |
Which surface carries the acoustic opening | Customer | An industrial design and environmental protection decision | Level loss from a path nobody designed |
Amplifier voltage and impedance | Customer | Set by the power supply and the amplifier selection | A configuration the rail cannot drive to its intended level |
Frequency band the product uses | Customer | Depends on content, not on component properties | A driver optimised for the wrong band; substitution fails repeatedly |
Driver platform and parameter selection | Manufacturer | Requires knowledge of compliance, magnetic circuit and tooling | Over-specified parts, or parts that need tooling that could be avoided |
Whether a change needs new tooling | Manufacturer | Depends on the platform's parameter range | Tooling cost booked for a change that needed none |
Measurement basis for published figures | Manufacturer | The manufacturer's test fixture defines it | Figures that cannot be compared against the customer's own tests |
Test points for acceptance | Joint | The customer knows the band; the manufacturer knows what is measurable | Acceptance measured where the product does not operate |
Installed resonance target | Joint | The customer sets the requirement; the manufacturer states what the cavity gives | A target met on paper and missed in the product |
Cost and volume trade-offs | Joint | The manufacturer knows which parameter costs what | Cost reductions that quietly remove the acoustic result |
3. What Each Side Needs From the Other
Short answer: The customer needs to supply constraints, not solutions. The manufacturer needs to supply behaviour, not just parts. Each is the harder of the two to give.
1. From the customer: the constraints the driver cannot see. Thickness, keep-out volumes, the surface available for the opening, the amplifier's voltage and impedance, the listening distance and the noise floor. These are facts, and they are cheap to state.
2. From the customer: the band the product uses. This is the single most useful piece of information a customer can give, because it determines which driver is appropriate and where the result should be measured.
3. From the customer: a reference unit if one exists. A subjective requirement becomes measurable when there is something to measure it against, and this is usually what converts a vague target into a specification.
4. From the manufacturer: which parameter changes need tooling and which do not. This is what makes the difference between an inexpensive custom part and an expensive one, and the customer cannot determine it.
5. From the manufacturer: the installed behaviour, not the free-air behaviour. A driver's published figures describe it under stated conditions, and the customer needs to know what the cavity does to them.
6. From the manufacturer: an honest statement of what cannot be met. A programme that is told early that a target is unreachable in the available volume can re-target; one that is told late has to redesign.
4. Where the Handoffs Actually Are
Short answer: Four handoffs carry most of the risk: requirement to design, design to sample, sample to tooling, and tooling to production. Each one loses information unless something specific is transferred.
7. Requirement to design. What transfers is the band, the distance and the constraints. What usually transfers instead is a diameter and a power, which is a solution looking for a problem.
8. Design to sample. What transfers is the cavity volume and the mounting condition. If the sample is built without the cavity the product will have, it validates a different system.
9. Sample to tooling. What transfers is which parameters are frozen and which are still open. A tooling decision taken while the diaphragm or winding is still moving will be revisited.
10. Tooling to production. What transfers is the measurement basis and the acceptance criteria. A part accepted on one basis and re-checked on another will show differences that are not defects.
The pattern across these four is that each handoff transfers a specification, and the specification is only as good as the conditions attached to it. A cavity volume without its sealing method, a sensitivity without its measurement basis, an acceptance level without its test point — each of these is a handoff that looks complete and is not.
5. Reading a Manufacturer's Range Against Your Own Constraints
Short answer: Published configurations state their own conditions, which makes them comparable against a product's constraints before any conversation starts.
Table 2 lists configurations from one published sample catalogue grouped by termination, which is one of the interface decisions a joint programme has to settle early. The termination method determines how the driver attaches to the product's circuit, and it is one of the parameters that can usually be changed without new tooling — which makes it a joint decision with a cheap outcome.
Table 2: Published configurations grouped by termination method.
Termination | What it suits | Representative configurations | Design consequence |
Solder type | Automated assembly with a defined footprint | HS001342H42, HS001534H39, HS002038H, HS002850H50, HS003050H, HS003650H, HS004550H, HS151125H, HS151130H, HS160930H, HS203045H45, HS204130H30, HS251233H, HS253540H, HS284011H | Requires a solder pad in the allocated area; the footprint is fixed by the driver |
Leaf spring or spring contact | Assembly where soldering is undesirable, and where the contact must tolerate some positional variation | HS002045H, HS080923H, HS150827H, HS151125H, HS151130H, HS352052H, HS402055H | Adds a mechanical interface that has to be positioned; removes the solder step |
Lead wire | Products where the driver is mounted away from the board, or where routing is needed | HS001846H, HS003050H50 | Adds a wire length that has to be routed and secured; changes the assembly sequence |
Track magnetic with secondary magnet | Applications needing higher sensitivity from the same outline | HS220942H42, HS280935H35, HS402055H | Adds magnetic circuit content, which affects depth and cost |
Project Case Study (Hongsheng) — The Constraint Nobody Owned
The programme was a compact projector, where the requirement combined audible level with reduced wind noise from the cooling path. The customer's original solution delivered insufficient level and lacked low-frequency impact, and repeated driver substitutions had not resolved it.
The root cause was a decision that had been made without an owner. The bass reflex path is tuned by its port, and the port tuning frequency depends on its cross-sectional area. That cross-section was shared with the product's airflow requirement, so the acoustic design could only ask for about a 10% increase — a limit that existed in the mechanical design and was invisible in the acoustic specification. The tuning frequency therefore stopped at 150 Hz against an ideal of 180 Hz. Until the two sides treated the port cross-section as a shared decision, the acoustic target was being set against a constraint nobody had claimed.
Once ownership was clear, the work moved inside the constraint rather than against it. A 45 mm driver with a polyurethane surround was selected to raise midrange level, and a copper-clad aluminium voice coil reduced moving mass. A surround thickness of 0.12 mm brought the driver's own free-air resonance to 160 Hz, preserving low-frequency impact despite the port tuning being lower than ideal. At the port outlet, acoustic damping material was added, which reduced the wind noise by roughly two thirds without noticeably reducing bass output — a treatment that acts on the flow noise at the exit rather than on the tuning frequency, and therefore complementary to enlarging the port rather than an alternative to it.
Three distinct frequencies apply in this design and were kept separate throughout, because conflating them is the most common specification error in a joint programme. The driver's own free-air resonance is 160 Hz. The installed system resonance, measured after assembly with the cavity and port, is 250 Hz ±15%. The port tuning frequency is the 120 Hz value before the cross-section increase and 150 Hz after it. Each side knew which of the three it owned.
Measurement confirmed the electrical side of the shared decision as well. The published sensitivity is 105 ± 3 dB at 5.65 Vrms at 10 cm, and 5.65² / 4 = 7.98 W, which matches the 8.0 W rating of the configuration, so the driver is operating at its rated point rather than above it. The unit is rated 8.0 W with a 10 W maximum in an 80 cc cavity.
The programme is in production with 20,000 units shipped. The engineering work also reduced unit cost by 0.6 RMB in total, from three sources: a 0.2 RMB saving from the polyurethane surround, 0.1 RMB from replacing screws with adhesive, and 0.3 RMB from production efficiency, which amounts to 12,000 RMB across the shipped quantity. The transferable conclusion is that the expensive part of a joint programme is rarely the driver; it is the decision that nobody owned, and identifying it is usually the fastest route to a result.
The port cross-section was shared with the airflow requirement, which the acoustic specification could not see. Once that constraint had an owner, the tuning target became achievable and the cost came down at the same time.
The distinction matters because it decides what a joint programme can achieve. Hongsheng states the termination, cavity condition and measurement basis alongside each published configuration, so a customer's constraints can be matched against a published option before a structure is committed.
6. How to Run the First Meeting
Short answer: Six items make the first conversation productive, and most of them are constraints rather than requirements. The goal is to allocate decisions, not to place an order.
11. State the constraints before the requirements: thickness, keep-out volumes, the surface available for the opening, and the amplifier's voltage and impedance.
12. State the frequency band the product uses, as a range rather than a point.
13. State the listening distance and, if known, the noise floor at that distance.
14. Bring a reference or competitor unit if one exists, so that a subjective requirement can be measured against something.
15. Ask which parameter changes would need new tooling and which would not. This is the question that determines the cost of the programme.
16. Agree which decisions are joint. Writing this down takes a few minutes and prevents the most common failure mode.
7. FAQ: Working With a Manufacturer on a Custom Design
Q1: What should the customer bring to the first meeting?
Constraints rather than solutions. Thickness and mounting height, the volumes already committed inside the product, the surface available for the acoustic opening, the amplifier's voltage and impedance, the frequency band the product uses, and the listening distance. A reference unit if one exists.
Q2: What should the manufacturer bring?
Behaviour rather than a catalogue. Which parameter changes need tooling and which do not, what the cavity does to the published figures, and an early statement of anything that cannot be met in the available volume.
Q3: Who decides the cavity volume?
The customer decides what the structure can seal, and the manufacturer states what that volume does to the installed resonance through FC = Fs × √(1 + Vas / Vb). The target that comes out of that exchange is joint, not one-sided.
Q4: How do we avoid paying for tooling we do not need?
Ask early which changes are parameter changes on an existing platform. Diaphragm thickness, voice-coil layer count and wire gauge, dome material, surround material, magnet grade and termination are all normally adjustable without new tooling. In one programme, 6 dB of level came from three such changes with no change to the driver's exterior.
Q5: What is the most common cause of a stalled programme?
A decision that nobody owned. The frequent example is a constraint that exists in the mechanical design and is invisible in the acoustic specification, such as a port cross-section shared with an airflow requirement. Until it is claimed by one side, the acoustic target is being set against a limit nobody can see.
Q6: Should the customer specify the resonance frequency?
Specify the band the product uses and the installed resonance needed, not a free-air value. The installed resonance depends on the cavity, which is why the target has to come out of a discussion rather than out of a datasheet.
Q7: How is a subjective requirement handled?
By converting it into a comparison. Measure a reference unit under the same conditions on the same day, and state the requirement as a relationship to that measurement. A subjective requirement that is never converted gets renegotiated repeatedly, which is the most expensive form of iteration.
Hongsheng can also state which parameter has to move first when a target cannot be met in the available volume, which is usually the information that decides whether a structure is committed or revised.
8. Summary: Allocate the Decisions, Then Build
A custom micro speaker programme is a sequence of joint decisions, and its cost is determined mostly by how those decisions are allocated. The customer owns everything about the product: thickness, keep-out volumes, the acoustic opening surface, the amplifier's voltage and impedance, and the frequency band the product uses. The manufacturer owns everything about the component: the platform, which changes need tooling, and the measurement basis behind published figures. The installed resonance, the test points and the cost trade-offs are genuinely joint, and they are where programmes either converge or stall.
The handoffs are where the information is lost. Each one transfers a specification, and a specification without its conditions — a cavity volume without its sealing method, a sensitivity without its measurement basis, an acceptance level without its test point — looks complete and is not. Hongsheng's published configurations state their termination, cavity condition and measurement basis, so a customer's constraints can be matched against a published configuration before a structure is committed, and the engineering team can state which parameter has to move first when a target cannot be met as drawn.
Next step If you are planning a custom micro speaker programme, the shortest route to a configuration worth testing is to state five things: the space available for the driver and its cavity, the impedance the amplifier will drive, the power the rail can supply, the level required at the intended listening position, and the frequency band the product actually uses. With those specified, our engineering team can recommend a suitable configuration for evaluation, or state plainly which part of the acoustic design has to change first.
More in This Series
· What Information Should You Provide When Requesting a Micro Speaker Quotation? →https://www.hsdz-spk.com/news/589.html
· Custom Micro Speaker Development: From Product Requirements to Mass Production → https://www.hsdz-spk.com/news/560.html