Weft Insertion Systems: Choose a Loom for the Yarn and Fabric

Start with the fabric the loom must deliver

A loom demonstration can be impressive and still leave the buying question unanswered. A fast machine may be running an easy yarn at a narrow width. Another may be making a difficult fabric at a lower speed. The mill still needs to know how each system will handle its own product.

Weft insertion is the way the crosswise yarn passes through the open warp shed. Shuttle, projectile, rapier, air-jet and water-jet systems move that yarn in different ways. This choice affects how the yarn is handled and how the edge is made. It also affects utility needs and the range of fabrics.

Begin with the intended yarns, fabric construction and accepted output. Then compare a defined machine configuration against those needs. The numerical cases below are invented teaching examples. They are not machine ratings, supplier quotations or results from physical weaving trials.

Understand what carries the weft

First ask what moves the yarn: a mechanical carrier or a fluid stream. Then look within each family. Its machines may use different guides, transfer parts, feeders and controls.

Shuttle: the yarn package travels too

A shuttle carries a quill or pirn through the shed, feeding yarn as it moves. The package must be replenished. Check how the package is changed. Manual work and stops can differ between shuttle looms.

CottonWorks' weaving overview describes this route. It remains in use for some heritage and specialist fabrics. The back-and-forth yarn path can provide a conventional woven selvedge. Check whether the product needs that edge. The insertion method alone is not a quality grade.

For each order, check how packages are prepared and how often they need changing. Include repairs and the fabric result. Keeping an installed loom and buying a new one are also different decisions. The method's age does not settle either case.

Rapier: control the pickup and transfer

Rapiers may use rigid rods or flexible tapes. In a common double-rapier arrangement, the giver takes yarn towards the centre. The taker receives it there and carries it onwards. Single-rapier systems also exist. Centre transfer therefore does not define every rapier loom.

The DORNIER P2 brochure describes a positively controlled transfer. The receiving clamp secures the yarn before the giving clamp releases it. That is a feature of the described design, not proof that any yarn will run at any setting.

Ask which grippers, guides and tension controls are included in the offer. Itema's R9500EVO description lists both guided and unguided transfer options. It also lists different cutters. A broad label such as “rapier” leaves these choices unresolved.

Run the most demanding yarn changes in the trial. A smooth single-colour sample tests only a narrow part of that range. Include the order's slubs, mixed yarn types and frequent selection changes where needed.

Projectile: separate the carrier from the supply

The projectile grips the weft while the supply package remains outside the shed. The carrier follows a guided path. The machine then returns it for later use. It does not take a full weft package across with each insertion.

Itema's P7300HP V8 description identifies insertion without a central handover. It lists spun yarns, filaments and tapes among its intended materials. These materials may make a projectile option worth testing. The trial must still use the intended yarn and width.

Examine clamp marks, damage, delivery reliability and edge waste in that trial. Confirm the actual guide and carrier configuration. A broad application range is useful background. It cannot guarantee the result for this fabric.

Air jet: match the yarn to the airflow

Air-jet insertion uses airflow to move the weft. Common systems combine a main nozzle, relay nozzles and a profiled reed. Yarn behaviour depends on the complete insertion path, not on a pressure number alone.

It is incorrect to reject every filament yarn as unsuitable. CottonWorks includes textured filament wefts in its air-jet applications. It treats slick flat filaments as a different case. Yarn surface, bulk and construction therefore deserve attention alongside the fibre name. A polyester label alone does not describe those features.

Toyota's JAT910 brochure describes sensing insertion timing and recommending timing adjustments. It also describes nozzle and valve changes. These controls belong to the described machine. The brochure gives no best pressure for another machine or product.

The Belforte and colleagues study of sub-nozzles examined airflow and drag. It also looked at interaction with the profiled reed. Its public abstract supports treating the nozzle and passage together. It does not supply a universal operating recipe.

Water jet: assess the complete wet route

Water-jet weaving is commonly associated with suitable low-absorbency synthetic filament yarns. Check the actual material and its finish. Synthetic yarns need not behave alike. Approve both warp and weft for the wet conditions. The fabric construction and the later process route must also suit them.

The TSUDAKOMA ZW8200 brochure describes the water-jet nozzle. It also shows devices that measure and store the weft. Water carries the yarn, but the yarn supply still needs control. Assess those parts together.

Include water supply, filtration, drainage and removal of residual water in the proposal. Ask for requirements tied to the offered machine and site conditions. A generic pH value cannot define the installation. Nor can a headline water-saving percentage.

Turn the product range into a machine brief

Give each supplier the same description of the target fabrics. For warp and weft, state the material and linear density. Add yarn structure, twist, surface finish and package form. Identify deliberately irregular yarns and any low-strength or damage-sensitive material. Provide representative packages, not only a small hand sample.

Use one count system consistently when comparing offers. A larger tex number means more mass per unit length. A larger indirect count has a different meaning. ApparelDiary's yarn-count guide explains these conversions. The number still does not describe the yarn's bulk or handling by itself.

Specify the weave, warp density, pick density, width and required fabric condition. Define the state in which the sample will be judged. It may be off the loom, after relaxation or after full finishing. A strong loom-state sample does not answer every later appearance or handling question.

Weft selection is another practical constraint. Different colours may use the same yarn construction. Two yarns with similar colours may still need different tension settings. Ask the supplier to run the actual selection sequence. Include the changes between yarn types. Confirm which feeders, selectors and controls the quoted configuration includes.

Separate current demand from possible future styles. A machine may meet today's narrow range but need extra equipment for a later fabric. Record that gap explicitly. An option may be worth buying for later use. Keep that unused capability separate from current output.

Follow the yarn from package to fabric

The insertion device is only part of the yarn's journey. The package, guides, feeder, brake, cutter and arrival control form a connected route. Identify that route for each proposal. First check whether yarn reaches the insertion point. Failure there differs from failure to carry it across the shed.

Ask the trial team to record where each confirmed weft interruption occurs. Keep an unresolved stop separate until its cause is known. A single total labelled “weft trouble” hides this distinction. It gives little basis for choosing a different feeder or insertion option.

Check more than a newly mounted package. A short demonstration may miss the package change and yarn tail. It may also miss conditions later in the package. Include the conditions that matter to the order and retain their output separately. Record any swap from a difficult package to an easier one. The later run then tests a different supply condition.

The same principle applies to yarn selection. Test the transition into and out of the demanding weft, not just a long run of each yarn on its own. Check the cut end, first inserted pick and following fabric against the agreed requirement. Then ask what limits the result. Is the product beyond the proposed range, or does a setting still need work?

Check width before comparing speed

Start with nominal machine width and maximum reeded fabric width. Neither is the same as loom-state cloth width or finished usable width. Edge allowances and width changes occur between them. A machine nameplate cannot supply the whole width plan.

For example, DORNIER's brochure marked July 2024 lists a 180 cm nominal P2 width. Its corresponding maximum fabric reeded width is 1,770 mm. This is a dated catalogue example, not a current offer for every configuration. Obtain the dimensions for the exact machine being considered.

Now take an independent planning case. The buyer needs 1.68 m of finished usable width. A validated route retains 96% of the usable loom-state width. The required loom-state usable width is:

1.68 ÷ 0.96 = 1.75 m.

Suppose the plan also needs a total of 0.04 m for edges outside that usable width. The simplified loom-state fabric width becomes 1.79 m, or 1,790 mm.

That already exceeds the catalogue example's 1,770 mm reeded width. It also has not allowed for further draw-in between the reed and the woven cloth. The “180 cm” label cannot justify the proposal on these assumptions.

The 96% retention and edge allowance belong only to this invented case. Establish real values through the intended fabric route. ApparelDiary's guide to usable and nominal fabric width explains the later department’s width needs. State clearly which width that department can accept.

Convert picks into accepted cloth

Picks per minute describe insertion frequency. They do not describe metres of cloth without a pick density. They also omit stoppages and the fabric lost to quality problems.

For a simple single-phase case, let the running insertion rate be n picks per minute. Let the fabric contain P picks per metre at the measurement state being used. If r is the running-time fraction within the observation window:

Gross cloth length = n × window minutes × r ÷ P.

Use actual inserted-pick counts instead if the speed changes materially. A single advertised maximum is not a time-weighted operating rate. Check the assumed pick-per-cycle relationship too. If the insertion arrangement differs, adjust the formula.

An eight-hour comparison

Consider two fictional proposals, R and J. They represent a rapier and an air-jet configuration for the same fabric. Both make 24 picks/cm, which is 2,400 picks/m, and the observation window is 480 minutes.

The window begins after setup. Its running fractions account for stops during production. They exclude the separate style change considered later. Assume constant speed while running and one inserted pick per stated cycle.

R inserts 480 × 480 × 0.88 picks. Dividing that result by 2,400 gives 84.48 m. Applying its agreed inspection result gives 84.48 × 0.98 = 82.7904 m of accepted cloth.

J follows the same calculation. Its stated running rate is 50% higher, but its accepted length in this case is only 14.45% higher. The lower running fraction and accepted fraction reduce the gain.

These are hypothetical results for two configurations, not typical figures for either technology. Each candidate must first meet the fabric requirements. A percentage cannot make an unsuitable fabric acceptable.

Keep the measurement state consistent

The table reports length at one defined cloth state. Do not insert a finished pick density into the equation and label the answer as loom-state length. If finishing changes length, convert each quantity on its proper basis.

“Accepted fraction” must also have a clear meaning. Here it is accepted metres divided by gross inspected metres. It is not one minus a defect-points rate. Several defects can occur in one length, and some rules permit defects within a limit.

Suppose both candidates have 1.80 m of usable width at this same state. Accepted area is 149.02 m² for R and 170.55 m² for J, using the unrounded lengths. If their widths differ, area and length rankings may differ too. Compare the output the next process can actually use.

Put utilities on the same output basis

A machine power figure does not necessarily include its supporting services. An air-jet comparison needs an agreed compressed-air allocation. A water-jet comparison needs the relevant water-system and drying boundary. Auxiliary equipment can also be needed on mechanically driven insertion systems.

Use a matched observation period and state what each meter includes. Keep allocated shared-service energy separate from directly metered machine energy. Keep cubic metres of air separate from kilowatt-hours. A motor nameplate is not a measurement of consumption.

Extend the fictional comparison. Suppose R uses 176 kWh within the agreed boundary. J uses 136 kWh at the machine, plus 96 kWh allocated to its compressed-air supply. Its comparison total is therefore 232 kWh. These values are invented, and the allocation would need evidence in a real study.

Using the unrounded accepted areas, R uses 1.181 kWh/m² and J uses 1.360 kWh/m². Under these assumptions, J makes more accepted fabric in the window. Its energy use per accepted area is also higher.

This does not establish a universal ranking. A changed yarn, width, pressure, machine option or utility plant can change the result. Energy per area is not a carbon footprint either. That requires more data and a defined assessment boundary.

ApparelDiary's compressed-air cost guide develops the shared-service question. When choosing the loom, request evidence on the same basis for each offer.

Allow for the time available to insert each pick

Width and speed interact through the available insertion time. A simple timing calculation can expose an unrealistic comparison. The machine's motion still needs its own analysis.

Assume uniform shaft rotation at 600 cycles/min, with one pick per cycle. One cycle lasts 100 milliseconds. In this example, allow 200° of the cycle for the relevant insertion window. That gives:

100 × 200 ÷ 360 = 55.56 milliseconds.

If a weft tip must traverse 2.0 m within that window, its required average travel speed is 36 m/s. At 800 cycles/min, the cycle is 75 milliseconds and the same angular window lasts 41.67 milliseconds. The required average becomes 48 m/s.

Alternatively, keep 600 cycles/min but increase the path to 2.4 m. The average needed in the same window becomes 43.2 m/s. A wider fabric has changed the demand even though the cycle rate is unchanged.

These are simple distance-and-time averages. They are not peak yarn speeds, nozzle-air velocities or recommended machine settings. Acceleration, braking, handover and the actual path all matter. Obtain the relevant timing limits and motion profile from the machine supplier.

Include changeover in a short-order decision

The earlier production window gives accepted rates of 10.3488 m/h for R and 11.8440 m/h for J. Both rates include production stops and quality losses. Setup time remains separate.

Now consider an illustrative 80 m order of accepted cloth at that same state. Suppose changeover takes one hour for R and 2.5 hours for J. Assume the observed accepted rates still apply. Also assume that no other resource delays this order.

R's elapsed time is 1 + 80 ÷ 10.3488 = 8.73 hours. J's is 2.5 + 80 ÷ 11.8440 = 9.25 hours. The candidate with the higher production rate finishes this short order later.

The two time estimates become equal at about 122.96 m. Set 1 + L ÷ 10.3488 equal to 2.5 + L ÷ 11.8440, then solve for length L. Above that point, J's higher rate saves more time than its extra setup takes.

Real changeovers may include reed changes, feeder work, trial cloth and approval delays. Define which tasks occupy the loom and which can run in parallel. Avoid counting the same setup loss again inside the running fraction. The example is a planning comparison, not a complete capacity schedule.

Use a trial that can expose the weak condition

A useful trial includes the fabric features most likely to challenge each proposal. This may mean testing the widest style or the most delicate weft. It may mean a demanding yarn-selection sequence or a specific edge for later processing. A demonstration should answer those questions directly.

Agree the measures before running. Include accepted width and density, fabric appearance and yarn damage. Also record confirmed stop reasons, repair time, utility use and edge waste. Keep rejected cloth and start-up cloth visible in the result. Run the trial long enough to cover the relevant package conditions. Include the operating conditions that matter to the order.

Compare yarn or trial-day groups fairly. The NIST guide to randomized block designs explains how blocks can control selected sources of variation. In this case, use two approved yarn specifications. Each round could test both candidates on each yarn.

Three rounds would give 12 runs: two candidates × two yarn specifications × three rounds. This is an example plan, not a prescribed sample count. Balance or randomize the order where practical and retain the run conditions.

Several swatches from one run provide within-run evidence. They do not create several independent machines. Nor do they prove a result for a whole insertion technology. With one machine per type, the trial compares those specific machines and their tested setups.

Inspect the fabric after the agreed downstream route too. Keep the same sample identities through relaxation, finishing and final checks. Investigate a later fault using its original loom position and conditions. Do not assume that insertion alone caused it.

Make the decision specific enough to verify

Name the machine configuration in the final comparison. State the approved yarn range, width range and fabric requirements. Attach the trial results and the installation needs that support the proposal. Separate demonstrated capability from an option that still needs a trial.

Review edge handling and waste with the intended later process. A selvedge that looks neat may still consume width that the cutting room expected to use. Confirm the waste path and usable width rather than judging edge appearance alone.

Use the actual product mix to compare accepted output and changeovers. Add utility needs and support to that comparison. Include spare-part access and the skills needed for routine settings and repairs. Use the supplier's installation and operating instructions for the specific equipment.

There is no single loom family that wins every yarn and fabric comparison. Choose using the product requirements, the configured machine and the trial evidence together. The mill can then use that record to check the installed machine as it begins making the intended cloth.

References

• CottonWorks: Weaving Basics and filling-insertion methods.

• DORNIER: P2 rapier weaving machine brochure, July 2024, printed pages 14–17 and 30–31.

• Itema: R9500EVO transfer, selector and cutter options.

• Itema: P7300HP V8 projectile insertion and application description.

• Toyota: JAT910 brochure, weft-insertion controls and features, printed pages 03–04.

• Belforte, Mattiazzo, Testore and Visconte: Experimental investigation on air-jet loom sub-nozzles for weft yarn insertion, 2011 issue; public abstract.

• TSUDAKOMA: ZW8200 water-jet loom brochure, nozzle and measuring/storage features.

• NIST/SEMATECH: Randomized block designs.