Rotor Spinning Productivity: Calculate Output and Test the Quality Trade-Off

Start with yarn delivered, not rotor revolutions

A rotor can turn faster while the mill gains very little saleable yarn. More breaks, longer restarts or rejected packages can absorb the extra delivery. Rotor spinning productivity therefore needs two answers: how much yarn the settings could produce, and how much accepted yarn actually leaves the process.

The first answer begins with units. Rotor speed is measured in revolutions per minute. Yarn delivery is measured in metres per minute. They are different quantities. A production formula that uses rotor rpm directly as metres per minute omits the twist relationship and can give a wildly inflated result.

The second answer needs a clear time period and product basis. A shift total for coarse yarn cannot be compared fairly with a shift total for fine yarn. Nor can a machine's gross output be compared with another machine's accepted output without explaining the difference.

The worked cases below use invented values. They demonstrate calculations and trial decisions; they are not recommended speeds, guaranteed gains or limits for a particular machine.

Understand what the spinning position does

Cotton Incorporated's CottonWorks description explains the central route: sliver feeds a rotor process, fibres collect under centrifugal action, and yarn is withdrawn. The incoming fibre supply and the outgoing yarn do not remain one continuous drafted strand as they do in ring spinning.

The opening roller separates fibres from the feed sliver. Air carries them toward the rotor, where they collect in its groove. The forming yarn draws fibres from this collection and leaves through the take-off path. Delivery and winding then build the package.

This sequence gives several places to lose output. Poor opening can disturb the fibre supply. Deposits can affect yarn formation. A failed piecing can keep one position idle. An unsuitable package can create trouble after apparently successful spinning.

Separate those causes during a trial. A count of stopped positions tells the mill that capacity was lost. It does not say whether the cause was sliver supply, a natural yarn break, a quality cut, package handling or a failed restart.

Automation varies by machine

Rotor spinning does not mean that every task is fully automatic. Rieter's R 37 information describes a semi-automated system. Its piecing uses computer control with manual yarn-end preparation and insertion. Automatic package change is an available option.

This matters when estimating the response to breaks. A labour plan must reflect the fitted machine and its options. It cannot rely on the broad label “rotor spinning.” Record who or what performs each step from a stop to stable yarn delivery.

Two machines may also distribute their work differently. A shared service unit has travel and waiting time. A position with its own restart system has a different pattern of delays. Measure the actual arrangement rather than assigning one standard efficiency to all rotor frames.

Derive production from twist, length and count

For a practical nominal calculation, the basic relationship is:

Delivery speed, m/min = rotor speed, rpm ÷ nominal twist, turns/m.

Rieter's rotor-component guide relates yarn twist to rotor speed and take-off speed. Use the ratio as a planning relationship. Confirm the machine's actual delivery and twist convention when checking a real recipe.

Take an illustrative rotor speed of 90,000 rpm and a nominal twist of 750 turns/m:

Delivery = 90,000 ÷ 750 = 120 m/min.

In one hour, one continuously running position delivers 120 × 60 = 7,200 m. If the yarn is 30 tex, each 1,000 m has a mass of 30 g:

Mass per position-hour = 7,200 × 30 ÷ 1,000,000 = 0.216 kg.

The divisor includes both the tex length basis and the conversion from grams to kilograms. Writing these steps once makes a spreadsheet much easier to audit.

For 400 positions running continuously:

Ideal machine output = 400 × 0.216 = 86.4 kg/h.

For an eight-hour shift, ideal output is 86.4 × 8 = 691.2 kg. This assumes that every position delivers at the stated speed for the full shift. It includes no stopped time or rejected mass.

The full expression is:

Ideal mass, kg = positions × delivery in m/min × minutes × tex ÷ 1,000,000.

A count in Ne cannot be placed in the tex position in this formula. Convert it first. ApparelDiary's yarn count conversion guide explains the different count systems and their direction of change.

Use measured count when investigating a gain

At the same yarn length, a heavier count produces more kilograms. That is arithmetic, not proof that spinning became more efficient. Check whether the product remains within the required count range.

ISO 2060:1994 covers yarn linear density by the skein method. Its public scope identifies different preparation and conditioning options. Use the agreed full method and keep that basis consistent across the trial.

A moisture or preparation change can complicate the mass comparison. Retain the test basis beside the count result. For production weighing, subtract tubes and packing and account for unfinished packages at the beginning and end of the period.

Turn theoretical output into accepted kilograms

Use a running fraction for the share of scheduled position-time spent delivering yarn. Then use an accepted fraction for the share of produced yarn mass that passes the product requirements. These factors answer different questions.

In the invented baseline, the running fraction is 92% and the accepted fraction is 98.5%:

Produced mass = 691.2 × 0.92 = 635.904 kg.

Accepted mass = 635.904 × 0.985 = 626.36544 kg.

Thus, the shift plan gives about 626.37 kg of accepted yarn. It does not give 691.2 kg ready to ship.

Now consider a trial at 100,000 rpm while keeping nominal twist at 750 turns/m. Delivery becomes 133.3333 m/min, using the unrounded value in later calculations. Ideal shift output becomes 768 kg, a gain of 11.11%.

Suppose the faster trial runs for 88% of scheduled position-time and accepts 97.5% of produced mass:

Accepted trial mass = 768 × 0.88 × 0.975 = 658.944 kg.

The accepted-output gain is about 5.20%, much less than the ideal delivery gain. This may still be useful, but it is the smaller figure that belongs in the production case.

If the faster trial accepted only 92% of produced mass, accepted output would be 621.7728 kg. It would then produce less acceptable yarn than the baseline despite its higher speed.

Find the break-even quality level

With the faster trial's ideal output and running fraction held fixed, the accepted fraction needed to equal the baseline is:

Required accepted fraction = 626.36544 ÷ (768 × 0.88) = 0.926795.

That is about 92.68%. It is an economic comparison within this invented model, not a permissible defect level. The customer's quality requirements remain unchanged. A trial can exceed this output threshold and still fail the order specification.

Do not apply these loss factors twice. If a measured shift total already records accepted net yarn mass, it is the result. Use the model to explain differences, not to reduce that result again.

If different sections run different counts or speeds, calculate each group separately. Add their masses afterward. A simple average of rpm can give a misleading total when the groups have different numbers of positions or running times.

Calculate two lots before combining them

Consider a separate one-hour case with two equal groups. All positions run continuously, and no quality loss is included yet.

Group A gives 200 × 120 × 60 × 30 ÷ 1,000,000 = 43.2 kg. Group B gives 200 × 100 × 60 × 20 ÷ 1,000,000 = 24.0 kg. Their sum is the correct total for these assumptions.

Using the simple mean delivery of 110 m/min and mean count of 25 tex would give 66.0 kg. The average delivery has been paired with an average count that no group actually runs. Multiplying these averages loses the link between the two inputs.

The same principle applies when one section stops for a lot change. Keep each group's own time and speed before adding its output. A mixed machine total can be valid, but it should be traceable to the production groups that created it.

Reconcile partly filled packages at the shift boundary

Full packages removed during a shift are not always the same as yarn produced during that shift. Some were partly filled before the shift began. Others still contain new yarn on the machine when it ends.

In a separate example, assume a single unchanged product and mass basis. The machine removes 600 kg of net yarn in full packages. Partly filled packages contain 80 kg at the start and 105 kg at the end. With no other yarn transfers or unrecorded losses:

Yarn produced = 600 + 105 − 80 = 625 kg.

The 600 kg figure understates this shift's production by 25 kg. It would be wrong to explain that difference as a speed or running-time loss.

This balance measures produced yarn. It does not establish that every kilogram passes quality checks. Keep held, rejected and accepted mass separate, and retain the same stock boundary in each trial. Estimates of partly filled packages also carry uncertainty. If that uncertainty is close to the proposed gain, weigh more carefully or extend the comparison period.

Make stopped time visible at position level

One machine stop and one yarn break are not equal events. The scale of the lost capacity matters.

The example machine has 400 positions and an eight-hour shift, giving 400 × 480 = 192,000 scheduled position-minutes. A 15-minute stop affecting the whole machine loses 6,000 position-minutes, or 3.125% of the shift's capacity at unchanged delivery.

By comparison, 100 separate yarn breaks with a mean total loss of two minutes each lose 200 position-minutes. That is about 0.104% of scheduled capacity. The calculation assumes the two minutes include waiting, repair and return to delivery.

These are independent illustrations. They do not explain every loss in the earlier 92% running fraction. A real loss account must include all relevant stopped periods without counting overlapping events twice.

This way of counting can change the order of improvement work. A mill might spend time reducing frequent short stops while a long restart after a power interruption costs far more yarn. Keep both the number of events and the time that affected positions could not produce.

Separate piecing success from piecing quality

A restart can succeed at making yarn but create a poor joint. The first outcome affects running time; the second affects product quality and downstream use. Both need evidence.

For example, 285 successful first attempts out of 300 piecing events give a first-attempt success rate of 95%. Report the remaining events too. Repeated attempts, service calls and the time until stable delivery determine the actual capacity loss.

Rieter's R 70 brochure describes rotor cleaning at piecing and controlled yarn-end preparation. It also identifies individual position automation and additional cleaning equipment. These are useful examples of how a machine addresses the restart process, not proof of a fixed success rate in every mill.

Test the pieced regions under the agreed quality plan. Do not approve them solely because a package continued winding. Keep the settings and material associated with each result so that poor joints can be traced to the trial condition.

Treat opening, rotor and nozzle as a matched route

The opening roller, rotor and take-off nozzle should suit the feed material and intended yarn. Rieter's component guide describes different tooth designs, rotor grooves and nozzles for different applications. It also warns that opening-roller speed can harm quality when it is too high or too low.

That does not supply one optimum setting for all cotton mixes. Begin within the maker's approved range for the fitted parts, then test the actual material. Changing opening speed and rotor speed together makes it difficult to tell which change caused a result.

Try to answer a specific question. Does the faster trial suffer from poor fibre separation, more contamination in the groove, or a change in yarn formation? Examine the feed and waste as well as the yarn. A clean-looking waste stream alone cannot prove that the process is retaining useful fibres.

Waste quantity needs a denominator and a defined collection point. More material in the trash container can mean stronger cleaning, more contamination in the feed, or more useful fibre lost. Compare the composition of that waste alongside its mass. Keep piecing waste separate where the machine collects it separately.

During a feed-material trial, keep the blend identity attached to the sliver cans. A remaining can from the previous mix can blur the changeover. Mark when the intended material reaches the sampled positions, then collect the quality and production data for that material. This is especially useful when a trial changes a recycled-fibre share as well as the machine recipe.

Compare conditions through the cleaning cycle

A short run on freshly cleaned parts may miss a problem that develops later. Include samples and running records from different points in the normal cycle. Keep the cleaning method and the time since cleaning visible in the comparison.

Suppose breaks rise toward the end of a run but fall after cleaning. That pattern is a clue, not a complete diagnosis. Check whether the raw material, room conditions or feed supply changed at the same time. Repeat the comparison before assigning the whole effect to deposits.

Record part identity and condition too. An older rotor or opening roller should not disappear behind a common machine-model name. If a trial requires a new part, its cost and service needs are part of that proposal.

Use the machine maker's inspection and service instructions. The article's calculation examples do not authorise higher speeds or a change in part combinations. A rotor's size and design belong with its approved operating range.

Do not turn a fibre category into a universal rule

Cotton, regenerated material and man-made fibres need their own trials. A broad claim that pure polyester cannot be rotor spun is incorrect: Rieter's component guide includes a specific 100% polyester application. That example does not approve every polyester grade, blend, finish or count.

For a changed feed, retain its source, preparation and blend basis. Compare the resulting count, quality and usable output. A cheaper input is useful only if the complete route still meets the order and gives a sound cost per accepted kilogram.

Keep the laboratory comparison fair

Rotor speed does not have one simple effect on every yarn property. A 1979 study by Simpson and Patureau tested speed alongside rotor, cotton and take-off conditions. Its public abstract reports different strength responses for the yarn counts studied.

Those historic results are evidence that the variables interact. They are not speed limits or recipes for a current machine. A modern trial still needs matched materials, counts, test methods and sufficient independent production runs.

Keep the main product tests visible even when output is the trial's main aim. These include actual count, tensile behaviour, mass variation, imperfections and the surface properties relevant to the buyer. The chosen limits should come from the product agreement and validated process needs.

Check what an imperfection total contains

The Uster Tester 6 technical data lists several sensitivity levels. Its standard total uses different nep thresholds for ring/air-jet yarn and open-end yarn. A lower total can therefore come from a different counting rule.

Consider an invented report with 12 thin places, 40 thick places and 80 neps per kilometre. Its total is 132. If a different nep threshold counts only 63 neps from the same material, the new total is 115. No spinning improvement has been demonstrated by that change alone.

State all thresholds beside the total. If a customer requests one particular combination, retain it. A machine's default report is not a substitute for the agreed comparison basis.

Averages also need context. Preserve results by package and machine position. A small mean improvement can hide a group of poor positions. ApparelDiary's practical guide to statistical process control explains how process behaviour differs from a product acceptance limit.

Choose a valid twist test

Do not assume that a familiar direct-counting method suits rotor yarn. The public scope of ISO 2061:2015 explicitly states that it is not suitable for products of open-end spinning.

Agree a method suitable for the yarn structure with the testing laboratory and customer. Record whether a figure is a nominal machine setting or a laboratory result. The production ratio above does not remove this need or turn one into the other.

For tensile comparison, ISO 2062:2009 covers single-end breaking force and elongation using a constant-rate-of-extension tester within its stated scope. Its options include different specimen preparation conditions. Use the applicable full procedure rather than combining pieces of different methods.

Keep the actual count beside force and tenacity results. Also examine variation and weak results. An acceptable average does not automatically show that pieced regions or all packages will perform well in the next process.

Check the package and the next operation

A rotor frame can wind yarn directly into a cross-wound package. The R 70 documentation, for example, describes controlled package build, adjustable density and electronic length measurement. A separate cone-winding step should not be inserted into every route by habit.

Decide what the customer actually needs. Package shape, tube, length, density and unwinding behaviour may matter as much as gross package mass. Further processing may still be required for a particular order.

Compare downstream performance using the intended package and fabric route. Keep yarn identity through warping, weaving or knitting. Record whether stops arose at piecings, ordinary yarn sections or package defects. This helps distinguish a spinning fault from a handling or setup problem.

A faster trial is not complete when its laboratory average passes. It must also meet the agreed end use. If a changed yarn requires different downstream settings, retain those changes and their costs in the final comparison.

Compare energy and value on accepted output

Use a common meter boundary and the same shift basis. Include the functions that the comparison claims to cover, such as machine drives and assigned suction. Keep shared-service allocation explicit.

Suppose the baseline shift uses 1,200 kWh for its 626.36544 kg of accepted yarn. The faster trial uses 1,280 kWh for 658.944 kg:

Baseline specific energy = 1,200 ÷ 626.36544 = 1.9158 kWh/kg.

Trial specific energy = 1,280 ÷ 658.944 = 1.9425 kWh/kg.

Accepted output rose by 5.20%, but energy per accepted kilogram rose by about 1.39%. Higher production has not automatically lowered the energy cost of each kilogram. These meter readings are invented inputs, not claims about any named machine.

Add the other costs that changed: raw material, parts, labour, rejected yarn and any extra downstream work. Use realistic values for downgraded output. A second-quality package should not receive the first-quality selling price unless the actual order supports it.

The final choice should identify the product, feed, parts, nominal twist, delivery, running time and accepted output. It should also state the test basis and the downstream result. This makes a speed change a defined production decision that another shift can understand and repeat.

References

• Cotton Incorporated, CottonWorks: open-end spinning.

• Rieter: rotor spinning components, opening rollers, rotors and take-off nozzles.

• Rieter: R 37 semi-automated rotor spinning machine.

• Rieter: R 70 rotor spinning machine brochure.

• Simpson and Patureau, 1979: rotor speed, spinning and yarn properties; public abstract.

• Uster: Tester 6 technical data, April 2026.

• ISO 2060:1994: yarn linear density by the skein method.

• ISO 2061:2015: twist by direct counting; public scope and exclusions.

• ISO 2062:2009: single-end breaking force and elongation.