PET spunbond basis weight variation should be investigated by confirming the measurement method, separating cross-direction differences from machine-direction fluctuations, and comparing the results with production records. An acceptable average GSM does not prove that every part of the fabric meets the agreed specification.
On a PET non woven fabric production line, changing throughput or line speed before identifying the variation pattern can shift the average without resolving the underlying problem. A structured investigation helps production and quality teams distinguish measurement errors, persistent widthwise differences, time-related changes, and isolated fabric defects.
Basis weight is the mass of fabric per unit area, usually expressed in grams per square meter, or GSM. It describes how much material is present within the measured area, not how thick, strong, or well bonded the fabric is.
A roll can have an average close to its target while containing lighter and heavier regions. Reporting only the mean can therefore hide local differences that matter during coating, lamination, slitting, or other downstream operations.
Define what is failing before investigating the process: the overall average, individual specimen results, differences across the usable width, or changes along the roll. These are related but distinct quality questions.
Cross direction, or CD, runs across the web width. Machine direction, or MD, follows the direction of fabric travel. A persistent difference between the left edge, center, and right edge is a CD pattern. A change at the same widthwise position over successive production intervals is an MD pattern.
Both can occur together. For example, a persistent light region near one edge may remain present while the average across the whole web also changes over time. Combining every result into one average makes these patterns difficult to distinguish.
A visibly darker area, a thicker region, or a different surface texture does not establish a GSM difference. Thickness also depends on fabric structure and the measurement conditions. Optical appearance can be affected by filament distribution and bonding.
Measure suspected areas before classifying the defect. If mass per unit area is consistent but thickness or appearance differs, the investigation needs to include those properties rather than treating the issue solely as GSM variation.
ISO 9073-1:2023 specifies a method for determining the mass per unit area of nonwoven fabrics. Use the agreed test method and its applicable requirements for formal testing. For troubleshooting, add a documented location map so every result can be traced to a position in the web.
Identify the usable product width, trim boundaries, and operator-side and drive-side orientation. Do not mix trimmed waste with saleable fabric results unless the investigation specifically concerns edge formation.
The following matrix illustrates how to organize diagnostic samples. It is not a prescribed sampling frequency or a substitute for the agreed test method.
| Production interval | Left usable-width position | Center position | Right usable-width position |
|---|---|---|---|
| T1 | L1 | C1 | R1 |
| T2 | L2 | C2 | R2 |
| T3 | L3 | C3 | R3 |
Record the actual distance of each position from a defined edge. Additional positions may be needed to resolve a narrow band or a localized defect. Select spacing and frequency according to the problem being investigated.
A single cross-width sample cannot establish whether a pattern persists over time. Repeat the mapped positions at documented production intervals and record the roll number, sample location, production time, and operating condition.
When samples are removed later from a finished roll, identify their position along the roll as accurately as possible. Do not equate sampling time with production time. Allow for the travel time between upstream equipment, any measuring station, and the sampling point.
Calculate GSM by dividing specimen mass in grams by specimen area in square meters. Retain enough information to check the calculation, including the actual cutting area and balance reading.
A pattern helps prioritize checks, but it is not proof of a specific cause. Use the observations below to select evidence to collect, not as an automatic adjustment guide.
| Observed pattern | Evidence to confirm | Candidate checks | Verification step |
|---|---|---|---|
| Persistent edge-to-center difference | The same widthwise region remains lighter or heavier in repeated samples. | Sampling boundaries, web position, localized laydown conditions, and relevant airflow or suction distribution. | Repeat the widthwise profile at unchanged operating conditions and compare corresponding positions. |
| Whole-width average changes over time | Several widthwise positions move in the same direction during the same interval. | Actual material throughput, web speed, usable width, and transition records. | Align sample positions with time-stamped process trends. |
| Repeating bands along the roll | Similar changes recur at measurable length or time intervals. | Periodic process events, control fluctuations, and relevant rotating equipment records. | Compare repeat spacing with operating records without assuming a mechanical cause in advance. |
| Isolated thin areas or holes | Local defects occur without a consistent full-width or time-related pattern. | Local web disruption, filament interruptions, handling damage, and sample integrity. | Inspect adjacent material and check whether the defect was present before downstream handling. |
| Laboratory and online readings disagree | Results differ for apparently corresponding material. | Location matching, time delay, sensor calibration, measuring area, and averaging settings. | Compare matched material using a documented correlation procedure. |
When thin areas coincide with broken filaments or spinning interruptions, use the PET filament breakage troubleshooting guide for the upstream investigation. Do not assume that every uneven region is caused by filament breakage.
Under steady conditions, average basis weight depends on the mass incorporated into the web divided by the area produced. With other conditions unchanged, increased material delivery per unit area raises the average, while increased web speed at unchanged material delivery lowers it.
This relationship helps check overall mass balance, but it does not explain a persistent CD profile. Increasing throughput to correct one light edge can make the rest of the web heavier while leaving the widthwise difference unresolved.
Compare actual operating values where available, not only controller setpoints. Also account for usable width, trim, waste streams, and whether the recorded throughput represents material incorporated into the sampled fabric.
For a confirmed widthwise pattern, review the conditions associated with filament distribution and web collection. Depending on the equipment design, relevant evidence may include laydown settings, air distribution, suction conditions, forming-belt condition, and recent maintenance.
Do not treat these variables as interchangeable adjustments. Checks and changes should follow the equipment's approved procedures. Inspection around hot, moving, or pressurized equipment requires the appropriate isolation and safety controls.
Start-up, speed changes, product changes, and shutdown material should be identified separately from stable production. If an outlying sample comes from a transition, investigate that event rather than using it to characterize the entire steady-state run.
Likewise, compare like with like: the same sampling stage, usable width definition, specimen size, and test method. Measurements taken at different process stages may not be directly interchangeable.
Confirm cutting, weighing, and sample identification
↓
Map the results by cross-width position and production interval
↓
Classify the issue as CD, MD, combined, or isolated
↓
Compare the pattern with matched production records
↓
Perform one approved check or controlled adjustment
↓
Repeat sampling and verify the result over an agreed period
Record the baseline, the reason for the change, the actual adjustment, and its time. Allow the process to stabilize and account for material travel before collecting verification samples.
Changing several settings together may improve the immediate result without establishing which change mattered. If a coordinated trial is necessary, plan it in advance and document the conditions rather than making unrecorded adjustments.
A single acceptable sample does not demonstrate that the pattern has disappeared. Repeat the same sampling plan over an agreed observation period, and confirm both the overall level and the position-specific results.
Use approved procedures to identify and manage material produced during the investigation. Product release decisions should follow the agreed quality requirements, not an operator's visual impression alone.
The following values are hypothetical and are included only to illustrate data interpretation. They are not Yanpeng production results, customer data, or recommended acceptance limits.
| Production interval | Left GSM | Center GSM | Right GSM | Interval average GSM |
|---|---|---|---|---|
| T1 | 47 | 50 | 53 | 50 |
| T2 | 48 | 50 | 52 | 50 |
| T3 | 47 | 51 | 52 | 50 |
Every interval averages 50 GSM, yet the left position is consistently lighter than the right. The next step is to verify the repeated widthwise pattern and investigate the corresponding conditions, not automatically increase total material delivery.
Whether this material passes depends on the agreed specification. An average target, an individual-result limit, and a cross-width uniformity requirement are not the same acceptance criterion.
A useful report retains the individual readings, their positions, and the test conditions. Summaries may include the mean, minimum, maximum, range, and coefficient of variation where appropriate.
Label the data group used for each calculation. A CV calculated from all positions and times can combine CD and MD effects; it should not replace separate profiles. Comparing CV values from different specimen sizes or sampling plans can also be misleading.
There is no single GSM variation limit suitable for every PET spunbond product. Agree on the nominal basis weight, usable width, sampling method, reporting rules, and acceptance criteria for the intended application.
For filtration-support applications, review the PET filtration support validation requirements alongside the GSM data. Basis weight alone does not establish filtration-support performance or downstream processing suitability.
When the investigation requires equipment-level support, prepare a concise record containing:
For a new project or configuration review, the PET spunbond non woven fabric production line page provides the product-level starting point. Translate the required fabric consistency into agreed operating conditions, test methods, and acceptance criteria rather than relying on nominal capacity or a single sample.
Yes. Lighter and heavier regions can offset each other in the average. Retain position-specific results and assess them against the agreed individual and profile requirements, not only the overall mean.
No. Thickness and mass per unit area are different properties. Measure both under consistent conditions before deciding whether the issue is material distribution, fabric structure, bonding, or a combination of factors.
Use the applicable test method and agreed sampling plan for formal acceptance testing. For troubleshooting, add enough mapped positions to resolve the observed pattern. A narrow defect may need closer spacing than a broad edge-to-center difference.
Only after confirming whether the issue is a low overall average or a localized light area. Increasing throughput may change the average without correcting a CD distribution problem. Follow the approved operating procedure and verify the result by repeat sampling.
Online measurement can provide useful process trends when installed, calibrated, and appropriate for the material. Laboratory checks remain useful for correlation and agreed acceptance testing. Compare matched locations and time intervals, accounting for the different measuring areas and averaging settings.
No single CV should be applied to every product and sampling plan. Specify the application requirement, specimen area, sample count, locations, and calculation basis before using CV as an acceptance criterion.