Carding Quality: How to Investigate Neps, Waste and Sliver Consistency

A card that delivers fewer neps is not always doing a better job. Its incoming cotton may simply contain fewer neps. A card that makes less waste may be retaining useful fibre. It could also be passing more unwanted matter into the sliver. Even a smooth-looking sliver can hide changes in mass along its length.

Carding quality has to be judged through these separate questions. What entered the card? What changed during carding? How did the sliver perform in the next process and in the yarn?

This guide focuses on cotton processed on revolving-flat cards for staple-yarn production. Roller cards for wool and nonwovens have different layouts and requirements. The worked examples use invented data. Their speeds, yields and quality values are not recommended settings.

Follow the material before choosing a test

Carding opens fibre tufts, removes part of the unwanted matter and forms a continuous sliver. It also affects neps, fibre orientation and the length distribution of the retained material. The Cotton Incorporated guide shows where carding fits. It follows opening and comes before drawing and spinning.

In a typical flat-card layout, the feed system presents material to the licker-in, also called the taker-in. The main cylinder carries fibres through the working zones. The cylinder and clothed flats help separate the fibres. The doffer takes the web from the cylinder, and the delivery arrangement gathers it into sliver for a can. An integrated drawing unit, where fitted, adds a further stage before the final delivery point.

That layout matters when results are compared. A sample taken before an attached drawing module is not the same process output as a sample taken after it. Likewise, cotton sampled at the bale does not describe the exact feed presented to the card after opening and cleaning.

Mark the two sampling points on the actual machine layout. Use approved sample points that can be reached safely. Follow the machine's inspection procedure. Do not take material from moving clothing or defeat guards to obtain a sample.

The test plan should follow the suspected fault. For a nep complaint, compare feed and delivered sliver. For unevenness, check the feed and delivery mass. Then find where the change first appears. For fibre loss, collect the relevant waste streams and check their composition. Testing only the final sliver leaves too many possible causes unresolved.

Distinguish a nep from a yarn imperfection

A fibre nep is a small entanglement of fibres. A nep count describes how many the chosen method detects in a stated mass of material. Nep size and seed-coat-related defects may also matter. A yarn tester counts thick events at a chosen threshold. That is a different measurement from fibre-nep counts.

ASTM D5866-12(2020) covers nep quantity and size in cotton from ginning and processing stages. Its public guidance supports comparing samples before and after a process. It also makes an important limit clear: the test itself does not identify the origin of every nep.

Use the same instrument method for both sides of a comparison. The reports must also use the same definitions. Keep the reference checks current. Record whether the result concerns fibre neps, seed-coat neps or another reported category. A combined total can conceal a rise in one type while another falls.

Methods can also disagree. A 2001 cotton study by Jacobsen and colleagues compared microscopy with AFIS. The two methods gave different results. Their experiment used a particular cotton variety and processing routes. The methods cannot be treated as identical. The study does not give a correction factor for today's instruments.

Plan an overlap period when the mill changes its test system. This includes changes to instruments, software classification or sample preparation. Test comparable samples by both arrangements before joining their results into one trend. Otherwise, the apparent change in the card may be a change in measurement.

Read nep reduction against the matched feed

A common calculation describes the change in nep concentration:

Nep reduction (%) = (feed neps per gram − sliver neps per gram) ÷ feed neps per gram × 100

This calculation requires a positive feed count and comparable samples. If the feed count is very low, a small counting difference can produce a large percentage swing. Keep the actual counts beside the percentage.

Consider two fictional production periods.

Period B delivers the lower sliver count, but its reduction relative to its own feed is smaller. It would be misleading to call the change from 60 to 45 a 25% improvement in the card's removal performance. That 25% describes only the decrease in the delivered concentration.

Nor does 80% nep reduction mean that 80% of the cotton was removed. The units are counts per gram, not kilograms of fibre. Some entanglements may open during processing, while mechanical action can also create new ones. A count difference does not show how many neps left in the waste.

To see another denominator effect, suppose 1,000 g of feed contains 300 neps/g and delivers 950 g of sliver at 60 neps/g. Multiplying concentration by mass gives 300,000 counted neps at entry and 57,000 at exit. The numerical decrease is 81%, compared with the 80% decrease in concentration.

That calculation illustrates the different denominators. It is not a conservation law for neps. More evidence is needed to show physical removal. The nep population can change during carding. Keep this count-based result separate from the mass balance and the measured waste.

Check the feed before blaming the cylinder

Compare incoming cotton from the same mixing and the same stage. A new laydown can change the starting nep level, short-fibre content and trash load. Split the observations when the feed changes during a trial. Each sliver sample then needs a matched input sample.

The feed system also determines how material reaches the working surfaces. A stable average feed weight does not rule out short thick and thin sections. Check when the sliver fault appears. Does it follow a feed stop, a change in chute level or one supply line? Use the machine's records and approved checks to test those possibilities.

Look across cards receiving the same feed. If several machines change together, an upstream cause deserves attention. If only one card changes, investigate that machine and its local feed path. Neither pattern proves the cause, but each narrows the next useful check.

Preserve the timing. Material at the input takes time to reach the output and the can. Samples taken at the same clock time may come from different portions of cotton. This matters most just after a feed or setting change. Allow for material held inside the machine. The trial condition must reach the chosen output point before it is sampled.

Avoid preparing an unusually clean sample by hand. Removing visible particles before testing changes the material being evaluated. Where a method requires a preparation step, follow it consistently and record it. A laboratory result should represent the process sample, not a hand-selected version of it.

Treat clothing, gaps and speed as a system

The carding surfaces act together. Flat clothing retains fibres for work by the cylinder wire. Check its condition and fit for the job alongside the cylinder. Graf's flexible-flat information describes this relationship and offers different designs for different applications.

Before a speed trial, establish whether the clothing is in a serviceable state. Review the service record and processed tonnage. Check for uneven wear or damage. Ask whether the problem started after service, reclothing or a material change. Those events can be more informative than the current speed alone.

The Graf service brochure links flat-clothing condition with sliver-nep performance. Its maintenance equipment is intended for defined service tasks. It does not justify applying one grinding interval or procedure to every wire design. Use the instructions for the installed components and qualified maintenance support.

Gap control is another part of the system. Rieter describes automatic gap control on the C 81. The machine also has features for trash monitoring and clothing care. These show how a particular machine manages process conditions. They do not make an old manual setting valid on a different card.

Changing several settings together makes a failed trial hard to explain. This applies to cylinder speed, flat speed, feed rate and gaps. For an initial fault investigation, test a clear hypothesis with a controlled change. Settings can also interact. To find the best combination, use a planned multi-factor experiment. Keep it within the maker's approved range.

One 2024 C 60 RSB study examined several settings. It used DP-90 cotton to make sliver for rotor yarn. Its scope is specific to that material, machine and route. The published optimum is not a universal recipe for ring-yarn preparation or another card model.

More waste can contain less lost fibre

Waste mass alone cannot show how selectively a card cleans. Licker-in waste, flat waste and other streams may differ in composition. Collecting them separately helps reveal those differences. Trützschler's TC 30i fact sheet describes separate waste extraction on that machine. The maker's recovery claims apply to its stated case. They do not guarantee a trial result.

In an invented comparison, each setting receives 1,000 kg of feed on the same mass basis. Representative waste analysis gives the following results.

Setting B makes 10 kg more waste but loses 4 kg less fibre in the measured stream. Its waste contains 14 kg more other matter. This is why reducing the total waste percentage is not always the right immediate aim.

The fibre fraction still needs interpretation. It can include short, damaged or otherwise unsuitable fibres. Calling all of it good fibre would overstate its value. Assess length and condition with a method suited to that waste, and check the retained sliver as well. A fibre-testing instrument is not automatically approved for every collected waste type.

Confirm that the collection covers the same trial interval and all intended outlets. A full bag at the start, shared suction collection or material held in a duct can distort a short trial. If streams cannot be kept separate, report their combined result. Do not assign it to one machine without evidence.

Use a common moisture basis for the mass comparison. There is no need to repeat a whole-mill yield account here. The trial must still compare like masses. Any returned waste should be identified so that it is not mistaken for new external feed.

For the distinction between regain and moisture content, see ApparelDiary's receiving-weight guide. The same mass definitions apply to the card trial.

Read short-fibre results with their definition

Nep reduction does not show whether the retained fibres kept their length. A trial may change fibre damage or which fibres enter the waste. Use a suitable length test to check those effects. Compare the whole relevant result, rather than infer length from a nep count.

Short-fibre content needs both a length cutoff and a basis. A percentage by number counts fibres. A percentage by weight describes their share of mass. The AFIS 6 technical sheet lists short-fibre results by number and by weight, with stated cutoff options. Match those definitions before comparing reports.

A simplified example shows why the two percentages differ. Imagine 100 fibres of equal fineness: 80 are 30 mm long and 20 are 8 mm long. Choose a cutoff that classifies the 8 mm group as short and the 30 mm group as long. The short share by number is 20%.

At equal fineness, fibre mass is proportional to length. The short group contributes 20 × 8 = 160 length units. The long group contributes 80 × 30 = 2,400. The short share by mass is therefore:

160 ÷ (160 + 2,400) × 100 = 6.25%

Both results describe the same invented population. They disagree because they answer different questions. Actual cotton has a spread of lengths and fineness values. This two-length example cannot replace measurement.

A lower short-fibre percentage after carding also needs context. The card may have removed short fibres, while other fibres may have broken. The retained sample shows the net result of those effects. Review waste and retained-fibre evidence together before attributing the change to gentler treatment.

Separate average sliver weight from unevenness

Sliver linear density is mass per unit length. A sliver can have the correct average weight yet vary badly along its length. Conversely, it can be quite even but consistently heavier than the intended count.

For sliver, 1 ktex equals 1 g/m. Suppose five illustrative 10 m samples weigh 49, 50, 51, 50 and 50 g. Their mean is 50 g per 10 m, or 5 g/m = 5 ktex. The sample standard deviation of those five masses is about 0.707 g.

CV of the cut-length masses = 0.707 ÷ 50 × 100 ≈ 1.41%

This small example explains the arithmetic. Five pieces are not a general acceptance plan. More importantly, that CV describes variation between 10 m sample masses. It is not the same as a capacitance tester's short-zone CVm result. The measurement length changes what variation can be seen.

ASTM D1425/D1425M-14(2020) covers capacitance measurement of textile-strand evenness, including sliver. Its public guidance explains what makes results comparable. Factors include test length, sensing zone, speed and test conditions. The blend also matters. If its fibres have different electrical properties, uneven blending can alter the signal.

Keep the measure consistent when comparing evenness reports. CV% and U% are different statistics. Use the same sample preparation and instrument settings. Review the trace and changes over relevant lengths. One headline value cannot describe them all.

Where an autoleveller is fitted, check its role in the result. It may correct certain mass changes without removing the source of a nep or trash problem. A lower delivery CV does not prove that the cylinder and flats are cleaning better.

Verify what the claimed production rate means

Delivery speed becomes meaningful only when the sliver weight is known. For a steady delivery:

Output (kg/h) = sliver mass (g/m) × delivery speed (m/min) × 60 ÷ 1,000

A fictional card delivering 5 ktex at 300 m/min produces 90 kg/h while running. If it actually runs for eight hours in a ten-hour period, it delivers 720 kg. Its average output over the full period is 72 kg/h.

The 90 and 72 figures answer different questions. Keep can changes, stops and any material held out of acceptable production visible. Do not compare one trial's running rate with another trial's output per scheduled hour.

An apparent speed gain can also come from a heavier sliver. At the same delivery speed, changing sliver count changes kilograms per hour. That may be useful for the line, but the next process must be set and assessed for the changed feed. Check both production and downstream quality before calling it a free capacity increase.

ApparelDiary's yarn count conversion guide helps translate count systems. Write the sliver unit in full on the trial sheet. Confusing tex with ktex creates a thousandfold error.

Design a trial that can survive normal variation

Start with a specific question. Does the proposed setting reduce sliver neps at the required output? Does it avoid extra fibre loss and harm to yarn quality? State the acceptance criteria before seeing the trial results.

Use several independently collected feed-and-sliver sample pairs. Repeat tests on one handful check part of laboratory repeatability. They do not show variation across a production period. Sample across the conditions that the proposed setting will need to handle.

For a comparison across several cotton lots, test both settings within each lot where practical. The NIST guide to blocked experiments explains this approach. Grouping comparable conditions helps separate the setting effect from other variation. Here, the original application is to avoid testing one setting only on the easiest cotton.

Vary trial order where feasible, and account for the time needed to establish each setting. Avoid always testing the baseline first and the proposed setting later. That order can mix changes over time with the setting effect. A change in ambient conditions or feed quality could then appear to be the treatment effect.

Record each result before averaging. If one setting works on two lots and fails on a third, the mean should not erase that limitation. Investigate whether the failed lot differs in a way that is likely to recur in production.

For example, three matched sample groups might give baseline sliver-nep counts of 50, 60 and 55 per gram. The proposed setting gives 48, 58 and 53 for the corresponding groups. The group differences are each two neps/g, even though counts vary by ten across the groups. The mean changes from 55 to 53, about 3.64% relative to the baseline.

This small invented set explains paired comparison. It cannot establish statistical significance or long-term capability. It also does not prove that a two-nep change matters to the yarn. Collect enough independent production evidence to judge the result. Assess both the uncertainty and the practical effect. A larger percentage change can still be unhelpful. The defect category may already meet the product's requirement.

Keep a reference sample and a link from the trial sliver to the yarn made from it. Mixing trial sliver with other sliver can dilute the effect in the final yarn result. Plan the material route before starting. You must be able to trace the yarn back to the trial.

Use the fault pattern to choose the next action

The following table is a starting point for investigation, not a list of confirmed causes.

Finish the investigation with a decision that names its limits. A setting may be accepted for one cotton group and output range while remaining unproven for another. A maintenance repair may restore the old baseline without supporting a higher production target.

The strongest carding result can be repeated. The retained fibre and sliver are suitable, waste is controlled, and the yarn meets its specification. Keep those outcomes together. A single low number cannot stand in for all of them.

References

• Cotton Incorporated: Textile Processing, David M. Clapp.

• ASTM D5866-12(2020): Neps in Cotton Fibers.

• Jacobsen and colleagues, 2001: Neps, Seed-Coat Fragments, and Non-Seed Impurities in Processed Cotton.

• Graf/Rieter: Flexible Flats.

• Graf: Service Carding and Nonwoven.

• Rieter: Card C 81.

• Alamirew and Tesema, 2024: Optimizing Carding Speed Parameters and C-60 RSB Draft to Improve Sliver Quality.

• Trützschler: TC 30i Fact Sheet.

• ASTM D1425/D1425M-14(2020): Evenness of Textile Strands Using Capacitance Testing Equipment.

• Uster: AFIS 6 Technical Data, February 2025.

• NIST Engineering Statistics Handbook: Randomized Block Designs.