Quality defects on a ceramic fiber blanket line are not random. The same seven account for nearly every failed acceptance test, and each one has a short, learnable diagnostic chain: the symptom's shape tells you the mechanism, the mechanism tells you which of five control variables moved, and the variable tells you the fix. A quality inspector who knows the chain can localize a problem in one sample; a process engineer who knows it can close it in one shift.
This guide lays out all seven defects in the same format: mechanism first, then a symptom–cause–fix table you can take to the line. Acceptance limits are cited from the standards the industry actually tests to. It is written for the inspector making the call, the process engineer owning the root cause, and the buyer writing acceptance criteria into a contract.
The Seven Defects at a Glance
| Defect | Symptom | Test method | Acceptance limit | Priority |
|---|---|---|---|---|
| Excess shot | Gritty, hard hand feel; visible particles | Water-wash sieve, φ ≥ 0.212 mm | ≤ 15% (premium ≤ 10%) | High |
| Coarse fiber | Average diameter > 5 µm | Microscope | Spun fiber 3–5 µm | High |
| Wide diameter spread | Broad histogram, high CV | Diameter histogram / std dev | Single tight peak | Medium |
| Broken / short fiber | Short fiber, dust, low strength | Length observation, tensile test | ≥ 0.04 MPa | Medium |
| Delamination | Layers peel apart after needling | Visual, peel test | No separation | High |
| Excess linear shrinkage | Shrinkage > 4% at classification temperature | 24 h soak (GB/T 17911) | ≤ 4% | High |
| Density variation | Thickness/weight fluctuates across or along the roll | Sectional weighing | Within ±5% | Medium |
Limits per Chinese national standards GB/T 16400 and GB/T 17911, with industry reference values where standards give none. Export contracts typically map these to the buyer's national equivalents.
Defect 1: Excess Shot
Mechanism. Shot is melt that never became fiber. Three mechanisms produce it, and each leaves a different shape. Melt that is too fluid (furnace too hot, viscosity too low) is thrown off the roller surface as intact droplets — spherical shot. Melt that is too viscous (furnace too cold) tears instead of stretching, leaving crescent-shaped fragments. And melt whose composition has drifted — alumina above ~60% raises surface tension sharply — can only ball up regardless of temperature. The full surface-tension mechanism is covered in the physics of ceramic fiber formation.
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Mostly spherical shot | Furnace too hot; viscosity too low | Lower furnace temperature back into the spinning window (±10–20 °C band) |
| Mostly crescent shot | Furnace too cold; viscosity too high | Raise temperature; check orifice for partial blockage |
| Shot concentrated in a band | Local hot/cold spot at the melt orifice | Inspect orifice temperature distribution |
| Whole batch over limit | Composition off the 45/55–50/50 window | Re-check batching and raw material analysis |
| Random scattered shot | Deposits flaking off the roller surface | Inspect roller cleanliness and wear — see roller maintenance |
Acceptance limit: shot content ≤ 15%, premium grade ≤ 10% (GB/T 16400, water-wash sieve at φ ≥ 0.212 mm — national standard). The deeper control strategy is in shot content in spun fiber.
Defect 2: Coarse Fiber
Mechanism. Fiber solidifies before it is stretched thin enough — either because the stretch ratio was insufficient (spinning roller too slow for the melt flow) or because the melt was too fluid to hold a stable drawing film.
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Average diameter too large | Spinning roller speed low; stretch ratio insufficient | Raise roller speed; recalibrate the speed ratio |
| Diameter coarse across the board | Furnace hot; viscosity low | Pull furnace temperature back; hold the ±10–20 °C band |
| Coarse sections intermittently | Melt flow surging | Close the loop with laser flow monitoring — see fiber diameter control |
| Coarse fiber plus shot together | The whole spinning window has drifted | Full parameter regression check, not a single-dial fix |
Defect 3: Wide Diameter Distribution
Mechanism. The average can be on target while the spread is not. Roller speed matching errors above 5% visibly widen the distribution; every 0.1% of additional impurity adds roughly 0.3 µm to the standard deviation; and melt flow fluctuation is the dominant noise source (manufacturer data).
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Bimodal or broad histogram | Roller speed matching error > 5% | Reset speed ratios at the VFDs; check roller dynamic balance |
| Long-tailed distribution | Melt flow fluctuating | Closed-loop flow control; hold std dev within ±0.3 µm |
| Std dev persistently high | Elevated impurities | Raw material analysis; change batch |
| Large batch-to-batch variation | Incoming material variability | Tighten incoming inspection — the cheapest control point on the line |
Defect 4: Broken and Short Fiber
Mechanism. Three routes to short fiber: quenching too aggressively, which cracks the fiber surface under tensile stress; a damaged roller surface, which hooks and cuts fibers mechanically; and melt starvation, which forms fiber in segments instead of continuously.
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Short fiber and powdery dust | Cooling too aggressive | Adjust roller water flow and the cooling gradient |
| Clean, square fracture ends | Roller surface burr or damage | Inspect and replace the spinning roller |
| Fiber formed in segments | Melt supply insufficient | Check orifice and flow monitoring |
| Tensile strength failing | Grain coarsening from cooling too slowly | Restore the gradient cooling setpoint (quench to 800 °C within ~0.1 s) |
Acceptance limit: tensile strength ≥ 0.04 MPa for a 25 mm blanket; 0.08–0.12 MPa is typical at 128 kg/m³ (national standard / manufacturer data).
Defect 5: Delamination
Mechanism. The needle loom interlocks layers that the collection chamber formed. If the batt arrived with local density differences, or if punching density or depth was insufficient for that batt weight, the layers never truly lock — and heat setting reveals it as peeling.
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Clear interlayer peeling | Punching density or depth insufficient | Adjust stroke rate and penetration depth — see needling density and tensile strength |
| Localized delamination | Batt area weight uneven | Adjust vacuum plenum zoning and belt speed |
| Edge delamination | Material starvation or buildup at the edges | Adjust edge guides and edge air |
| Warping delamination after heat setting | Furnace temperature or soak time wrong | Recalibrate the heat-setting schedule |
A delaminated sample's thickness profile is the fastest diagnostic: uniform thickness with peeling means needling; thickness variation with peeling means collection.
Defect 6: Excessive Linear Shrinkage
Mechanism. Alkali oxides (Na₂O + K₂O) depress both viscosity and refractoriness; in service the glass phase softens and the blanket shrinks. Low alumina fails the same test but only at the highest temperatures. This is the one defect that chemistry — not machinery — owns almost entirely.
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Shrinkage over 4% across the batch | Na₂O + K₂O elevated | Change raw material lot; hold ≤ 0.5% (high-purity ≤ 0.2%) |
| Collapse at classification temperature | Product grade underspecified for the application | Re-check classification temperature selection |
| Localized shrinkage pits | Shot-rich zones in the blanket | Treat as a fiberizing problem — see Defect 1 |
| Only the hottest samples fail | Alumina content low | Re-verify composition; raise Al₂O₃ |
Acceptance limit: linear shrinkage ≤ 4% after 24 h at the classification temperature (GB/T 17911 — national standard).
Defect 7: Density and Thickness Variation
Mechanism. Belt speed mismatched to fiber output, uneven vacuum across the plenum, and inconsistent squeezing at the needle loom each leave a distinct spatial signature — which is why the pattern of the variation is the diagnostic.
| Symptom | Likely cause | Corrective action |
|---|---|---|
| Left-to-right thickness variation | Collection uneven across the width | Re-zone the vacuum plenum |
| Periodic variation along the length | Belt speed drifting against fiber output | Interlock and recalibrate belt speed |
| Isolated dense or loose patches | Uneven squeezing at the needle loom | Adjust loom feed and needle boards |
| Whole roll off density | Batt weight setpoint wrong | Verify with green-batt weighing before needling |
The Response Flow: From Detection to Closed Loop
Finding a defect is the beginning, not the end. The plants that contain quality escapes follow the same six-step discipline:
- Detect — online visual inspection plus scheduled sampling against the test panel above;
- Act immediately — correct the parameter from the linkage table; slow or stop the line if the defect is structural. Defects propagate — a furnace drift makes every subsequent meter defective;
- Contain — isolate, mark and hold the affected rolls;
- Disposition — judge against GB/T 16400 limits: accept, rework (re-slitting or splicing where legitimate), or scrap;
- Trace — log furnace number, batch, time and parameters so the root cause is findable later;
- Close the loop — fold the corrective action back into the operating procedure and internal control limits, so the same defect costs you once, not monthly.
One rule is absolute: never "downgrade-sell" defective rolls or blend them into passing batches. A refractory defect does not stay a cosmetic issue — at 1200 °C in a furnace lining it amplifies into lining collapse and melt leakage. The quality floor is a safety floor.
Matching Analysis Tools to Root Causes
| Defect | Usual root-cause family | Suggested QC tools |
|---|---|---|
| Excess shot | Temperature / viscosity | Fishbone diagram + control chart |
| Coarse fiber | Roller speed / temperature | Control chart + scatter plot |
| Wide diameter spread | Flow / impurities / roller speed | Histogram + stratification |
| Broken fiber | Cooling / roller surface | Pareto chart + fishbone |
| Delamination | Collection / needling | Cause-chain analysis |
| Shrinkage | Composition | Chemical analysis comparison |
| Density variation | Collection / belt | Stratification + control chart |
What Buyers Should Take From This
If you purchase blanket rather than produce it, this same table inverts into your acceptance checklist. Four questions put these defects under contract:
- Which standards and methods? Name the test methods (sieve size, soak duration, sample size) — a limit without a method is unenforceable.
- Which limits? Shot ≤ 10%, shrinkage ≤ 4%, density within ±5% and tensile ≥ 0.04 MPa are all contractible numbers, not aspirations.
- Which traceability? Can the supplier tie each roll to a furnace batch and parameter log? That is what makes a claim resolvable instead of arguable.
- Which containment? Ask how defective sections are isolated. A supplier with a documented containment flow almost never ships you the problem.
Frequently Asked Questions
What causes shot in ceramic fiber blanket?
What is the acceptable shot content for ceramic fiber blanket?
Why does ceramic fiber blanket delaminate?
What causes excessive shrinkage in ceramic fiber blanket?
How is ceramic fiber blanket quality tested?
What causes uneven density or thickness in needle-punched blanket?
The Bottom Line
Seven defects, five control variables, one discipline. Shot, coarse fiber and wide diameter spread all live in the melt-and-spinner window — furnace stability and roller speed matching. Broken fiber lives in cooling and roller surfaces. Delamination and density variation live in collection and needling. Shrinkage lives in raw material chemistry. Learn to read each symptom's shape — spherical versus crescent shot, uniform versus localized peeling, periodic versus random thickness variation — and the defect names its own cause. Then close the loop: contain, disposition, trace, and fold the fix into the operating standard, because the same defect should never have to be solved twice.
Related Reading
Continue with these guides and equipment pages: