When a speaker fails a reliability life test and the report reads "tinsel wire broken, no sound", the wire itself is rarely the real cause. Rule out a genuine material defect and nine times out of ten the fault sits in another process step. Where the conductor broke — and how far it sits from the solder joint — already narrows it down: soldering process, terminal design, arc allowance, or conductor selection. This note sets out the evidence for all four failure zones and what to change in each.

1. The break location is the diagnosis

Before you look at the surface of the wire, measure how far the break sits from the solder joint. On the same batch and the same construction, the distribution of break locations pins the cause down on its own.

2. Zone B: solder where it does not belong

Why the back of the joint, of all places

Solder holes in terminal plates are usually much larger than the wire they take. A generous hole threads quickly and wets through, but it also demands more solder — and filling it needs higher temperature and faster thermal recovery.

At the same time, lead-free solder melts far above the 63/37 eutectic at roughly 183 °C: SAC305 runs 217–220 °C, SnCu0.7 around 227 °C. The line takt does not pause for the iron. Squeezed from both ends, the floor typically ends up turning the temperature up, running an underpowered iron dry, and picking a high-activity solder wire because it wets more easily.

There is only one outcome: the wire becomes too easy to tin.

How the failure develops

1. Wick. Solder travels along the gaps between foils toward the back of the joint. Copper and tin have a strong affinity — the more easily the wire tins, the further the solder travels.
2. Freeze. As it solidifies, solder wraps the back in a hard collar. Its edge is not rounded but close to a blade: a thin solder fillet solidifies into a knife section.
3. Cut. The conductor vibrates tens to hundreds of times per second. The hard edge presses repeatedly on adjacent foils; one strand breaks first, arcing at the fracture erodes the neighbouring strands, and the conductor finally separates.

This failure has a signature you can read at a glance: flat fracture, no necking, solder on the fracture face. Ductile overload pulls a neck; fatigue leaves striations. A wicking-hardened break shows neither — that short length is no longer copper carrying load, it is a block of solidified solder doing the cutting.

Why silver-plated wire is more exposed

Silver has an even stronger affinity for tin than copper does. At identical settings, silver-plated wire wicks further and hardens more visibly, and the solder collar behind the joint forms more readily.

Some customers eventually switch from silver-plated to bare copper to suppress the risk, trading away high-frequency behaviour and appearance. That is a passive compromise, not a solution. The right lever is controlling solderability, not downgrading the material.

Four checks on the floor, no equipment needed

1. Look at the back. A full fillet on the joint face is normal; the back should be clean. Solder extending more than 1 mm past the joint edge is a fault.
2. Bend it. A healthy length bends softly and springs back. A hardened length feels brittle and returns poorly.
3. Magnify it. At 15x, a knife-edged or serrated solder rim is abnormal; a rounded, full fillet is normal.
4. Read the fracture. Flat, solder on the face, no necking — all three together essentially confirm wicking hardening.

Countermeasure: write "full on the face, clean on the back" into acceptance

This is the single most practical line in this note. It is easier to execute and to judge than any temperature profile. Actions, in order of return on effort:

1. Cap the iron at 350–370 °C. If the back shows hardening, drop 10–20 °C and retry — do not compensate with a longer dwell.
2. Use a higher-power temperature-controlled iron. Poor thermal recovery is the root of "turning the temperature up"; an oxidised tip transfers heat worse and closes the loop.
3. Solder wire with chlorine content ≤ 0.6%. Acidic residue left on the conductor corrodes the foil over time and pits it — a second, hidden failure source.
4. Do not chase "instant wetting". Moderate solderability is what you want. A wire that tins too easily will wick to the back; that is the most direct material-side cause and effect in this whole discussion.
5. Structural fallback. A soft protective adhesive on the back of the joint. Effective, but it adds a process step and a material cost.

Solving it on the material side

CHRYPOO post-treatment holds solderability in the moderate band and, during soldering, lets the penetrating agent form a protective film on the back of the joint. That film does two jobs at once:

• It blocks further wicking, confining the hardened zone close to the joint instead of letting it run down the conductor;
• It bonds the foils into one body, cutting strand-to-strand movement, rubbing and cutting — the second wear source after the knife edge.

Verified condition and result: lead-free solder wire, temperature-controlled iron at 370 °C, inspection under a 15x optical microscope. On 20-strand silver braided, 12-strand silver stranded, 20-strand copper braided and 3-strand copper stranded constructions, the back of the joint showed no wicking and no hardening in any of them. Because the film already protects that area, the back of the joint no longer needs a silicone dab — one process step and one material cost removed.

3. Zone D: wire slap is not "not hard enough"

When the lead slaps the cone above and the spider below, the first instinct is to ask the wire maker for a harder wire. One concept has to be settled first: hardness and toughness trade off against each other. Raise hardness and toughness falls. The traditional route adjusts the rosin ratio — rosin is mainly a shape-setting agent — to "harden" the wire. Hardness does rise, but the wire turns brittle; worse, the rosin-wax mixture itself degrades under vibration, so service life often drops instead.

Wire slap usually comes from three causes stacking up:

• Insufficient support. The conductor cannot stand up through the stroke, the arc collapses, and the wire touches the cone or the spider;
• Insufficient internal damping. Springback is too fast, excursion is not absorbed, and the swing grows;
• Length and arc mismatch. Moving from a cone-hole feed to a voice-coil lead adds a third or more to the length and changes the exit angle substantially — the old arc-height settings simply no longer apply.

How to confirm it: strip the driver and look for kinks in the conductor, or a collapsed arc. Acoustically, a regular "tick" in the low-mid band that follows excursion rather than a continuous noise — the classic signature of a lead striking the cone or spider.

What to change: choose a construction with good shape retention (the cross-bracing of a braided wire provides this well); recalculate lead length and arc height — slightly long with a visible arc, never pulled straight and tensioned; raise internal damping through post-treatment to cut the swing. Verified condition: 20 °C, 20–30 mm length, ±45° flexing for 20 cycles, with no loss of toughness and no powder shedding.

4. The fourth type: the conductor is too thin for the power

This type is easy to identify because it tracks power and excursion, and the signs cluster tightly:

• Necking at the fracture — a ductile pull, not a flat cut;
• Break in the middle of the free length, not near the joint;
• Overall discolouration of the conductor (blue or black from overheating);
• Concentrated in high-power, long-throw, low-F0 drivers.

The countermeasure is direct: go up a gauge, or move to a higher-strength alloy. On the TG series, life is roughly 1.5–2x a standard alloy in the same construction.

This is the type most often mislabelled as "bad wire". In practice it is experience from small and mid-power drivers applied unchanged to high-power, long-throw ones. Since Class-D amplification became standard, that mismatch has become common.

5. A winter problem: low-temperature embrittlement and shelf life

The flexibility of conventional tinsel wire comes from a rosin-wax mixture on its surface. That system has two weak points, and neither is visible on the production line:

• Temperature. Tack falls away below 4 °C; 15–30 days at −18 °C causes loss of adhesion, embrittlement and powder shedding.
• Time. Freshly made wire is fine, but rosin wax loses tack in storage and sheds the same way. The same lot number can behave differently at three months and at six.

The symptom is usually not a break but dirt: powder on the voice-coil winding, foreign matter in the gap causing noise, contamination on production equipment.

How to verify: hold the same lot at 4 °C and at −18 °C for 15–30 days each, then flex a 5 mm length through ±180° repeatedly and check for embrittlement and powder. This is more reliable than any visual inspection. CHRYPOO post-treatment shows no embrittlement and no shedding in either range; twist the foil open and the fibre core does not shed either.

6. Does fuzz actually matter?

This is the question customers ask most often, and the answer has two halves — it cannot be settled in one sentence.

The half that does not affect life: tinsel cores are mostly staple chemical fibre (aramid 1313 or 1414 staple), and spinning inherently produces fuzz. The foil is wound helically — the standard is 22 ± 2 turns per 10 mm, about 75% coverage — so it cannot fully cover the fibre. Fuzz is therefore normal and does not shorten service life. Eliminating it entirely means switching to filament fibre (commonly called ballistic yarn), at a considerably higher price.

The half that affects solderability: fibre has a lotus effect and repels solder. Heavy fuzz causes poor wetting — incomplete tinning, slow tinning, cold joints — which directly undermines the stability of automated termination and leaves hidden risk at the joint.

The industry's traditional remedy is to singe the fuzz with an open flame. The process window is narrow: fuzz that is under-burnt passes into good product, fuzz that is over-burnt anneals and softens the foil. Both are latent defects, essentially undetectable afterwards. CHRYPOO solves it in post-treatment instead: both stranded and braided wire come out smooth and fuzz-free with no singeing step — which matters far more on an automated line than a clean appearance.

7. A cycle count means nothing without the test condition

Buyers often compare suppliers on "we pass tens of thousands of cycles". Flex life is extremely sensitive to test conditions: bend radius, angle, speed, load, temperature and failure criterion. Change any one of them and the numbers are simply not comparable.

We use two conditions for two loading modes, plus a low-temperature pre-condition:

• Terminal root: 5 mm length, ±180° repeated, for the limit bend near the solder joint;
• Free length: 20–30 mm, ±45°, 20 cycles, for mid-span swing;
• Low-temperature pre-condition: 15–30 days at 4 °C / −18 °C before testing, covering winter climate and long-distance shipping;
• Failure criterion: DCR rise ≥ 10% or intermittent continuity — either condition fails the sample. Checking only for a complete break misses samples that have lost strands but not yet separated.

Conditions have to be matched item by item. Otherwise what you are comparing is test methods, not products.

8. Soldering settings at a glance

9. Where to start

If you have a suspect batch in hand, work in this order — a conclusion in a day or two:

1. Classify. Take the most recent 20 failures — life-test failures or customer returns — and sort them into the four zones of Table 1. Zone B dominant means the problem is soldering; Zone D dominant means it is mechanical design.
2. Sample the back. Pull 10 soldered sub-assemblies at random and inspect the back of the joint at 15x, combined with a manual bend to judge the hardened proportion.
3. Run a comparison. Log iron temperature and solder wire type for one week and read them against your defect rate. Many plants discover at this step that 10 °C less halves the defect rate.

After those three steps, whether it is a process problem or a selection problem is no longer ambiguous.