Determining Spectrogram Signature Thresholds for on Line Autoleveler Draft Adjustment in Flax Slivers

Signal

Flax sliver profiles display severe mass variance because unretted pectin bonds bundle individual bast ultimates into irregular clusters. These composite strands vary in length from 25 millimetres in carded tow to over 600 millimetres in hackled line stricks. When an online sensor scans incoming sliver at the draw frame, capacitive plates or displacement discs record this composite mass variation as an electrical voltage.

The raw output contains both random fibre arrangement and repeating mechanical imperfections. Attenuation demands precise control.

Fourier transformation converts this spatial displacement profile into an amplitude spectrogram. The horizontal axis represents the wavelength of mass variation on a logarithmic scale, extending from one centimetre to fifty metres. The vertical axis represents the amplitude of variation at each discrete wavelength.

In staple yarns composed of uniform synthetic filaments, the baseline follows an ideal Poisson distribution where fibre ends generate predictable, random mass fluctuations. Bast fibres deviate sharply from this baseline. Bundle cohesion varies constantly.

Technical flax bundles possess variable transverse dimensions and split irregularly during drafting. This mechanical separation creates a high, broad elevation in the spectrogram designated as the drafting wave. The drafting wave occupies the zone between two and four times the mean bundle length.

In an autoleveled intersecting gill drawing frame running hackled line flax, this broad swelling peaks between 0.6 metres and 1.4 metres. Cotton-oriented autoleveling algorithms misinterpret this natural swelling as an extended sliver defect. Automated servos attempting to eliminate this natural peak force synthetic irregularities into the delivered sliver.

A 20 ktex line flax sliver running at 220 metres per minute produces a drafting wave amplitude of 1.4 squared millimetres between 600 millimetres and 1200 millimetres wavelength under dry drafting conditions.

The operational spectrogram contains three distinct features: the random bundle baseline, the broad drafting wave, and sharp periodic chimneys. Sharp chimneys indicate mechanical faults within the drawing passage. An eccentric fluted roller, an unmeshed gill drive pinion, or a bent faller bar generates a solitary spike at an exact wavelength corresponding to the component circumference multiplied by subsequent draft.

Online autoleveler thresholds discriminate between these narrow chimneys and the broad drafting hump.

Fourier Decomposition in Bast Fibre Drafting

Mathematical extraction of periodic components relies on discrete Fourier transformation applied to finite sample lengths of sliver displacement data. The sampling frequency dictates the shortest detectable defect wavelength. A delivery speed of 250 metres per minute sampled at 2.5 kilohertz yields a spatial data point every 1.67 millimetres along the strand.

This high resolution prevents aliasing in the short wavelength domain below ten centimetres.

Flax behaves unpredictably. Coarse shive fragments and unhackled root ends produce high-amplitude voltage transients that skew standard Fast Fourier Transform algorithms. Windowing algorithms such as the Hanning or Hamming window constrain spectral leakage across adjacent frequency bins.

Without appropriate windowing, a solitary coarse shive calculates as a periodic harmonic spread across three meters of material. The digital processor applies a running median rejection routine to eliminate isolated voltage spikes prior to spectral summation.

Spectrogram Profiles across Line and Tow

Carded tow slivers yield a substantially different spectral landscape than hackled line slivers. Tow processing breaks long technical bundles into shorter segments, shifting the mean bundle length toward 45 to 80 millimetres. The resulting drafting wave contracts toward shorter wavelengths, settling between 150 millimetres and 350 millimetres.

The overall baseline amplitude rises because tow contains higher fractions of cortical tissue, short fibre fragments, and residual woody core.

Hackled line sliver exhibits a suppressed random baseline below 0.3 metres, accompanied by an intense drafting wave. This profile reflects the high cohesion between parallel long bundles in the sliver body. When drafting rollers grip these long strands, inter-fibre friction forces adjacent shorter strands to move forward in collective bunches rather than individual units.

Mill meters register unevenness. The drafting wave represents this collective movement. Leveling systems operating on long line flax restrict their correction bandwidth to exclude these hydrodynamic drafting surges, focusing drive corrections purely on sustained cross-sectional trends.

Filter

Digital algorithms parsing cross-sectional data separate mechanical gear anomalies from organic bundle clump variations. The controller executes a band-pass filtering architecture tailored to flax kinematics. Wavelength classification divides into three distinct correction territories: the short zone below 0.15 metres, the intermediate zone between 0.15 metres and 2.0 metres, and the long zone exceeding 2.0 metres.

Each zone commands a completely different autoleveler response.

The short zone falls below the mechanical correction capability of commercial servo drives. Infeed measuring rollers positioned at the gill box detect thickness variations, but the physical inertia of the motor, gear linkages, and drafting cylinders prevents instantaneous speed compensation over a few centimetres of sliver. Mechanical slippage creates ghosts.

Any attempt to correct mass defects below 0.15 metres introduces anti-leveling, where the drafting change arrives out of phase with the defect, doubling the local mass variation. The filter rejects all spectral energy in this short zone.

The intermediate zone contains the natural flax drafting wave. The autoleveler filter treats this zone as a protected exclusion band. When a spectrogram registers elevated amplitude between 0.5 metres and 1.5 metres, the controller evaluates the spectral width of the peak.

A broad, smooth curve indicates natural bundle slippage. The leveling drive ignores this broad feature. Conversely, a sharp, narrow peak rising out of this intermediate zone indicates an operational defect.

Shives blind optical pickups.

Bandwidth Discrimination in Sliver Autoleveling

Mathematical filtering defines the boundary between correctable mass drift and uncorrectable fibre noise. The central processing unit applies an infinite impulse response band-pass filter combined with an adaptive threshold envelope. This dynamic envelope floats at a calculated margin above the natural Poisson baseline of the specific flax lot.

When the detected variance exceeds this floating threshold, the autoleveler engages servo correction.

The drafting wave amplitude shifts with retting degree and moisture content. Wet flax exhibits higher inter-bundle friction, increasing drafting wave amplitude by forty percent relative to dry sliver. An effective threshold algorithm tracks the ambient sliver conductivity and relative humidity in the drawing hall.

The controller broadens the exclusion band whenever sliver moisture rises above eleven percent, preventing the autoleveler from fighting humidity-induced drafting waves.

• Phase Distortion Failures occur when digital low-pass filters introduce transmission delays that shift the correction pulse past the measured thin spot.
• Resonance Amplification Modes appear when cyclic leveling interventions match the rotational period of the intermediate drafting cylinders.
• Baseline Clipping Defects emerge when static threshold settings classify natural bundle diameter shifts as incoming raw material contamination.
• Harmonic Mirroring Errors develop when unshielded high-amplitude shive signals generate mathematical sidebands in the Fast Fourier output.

Distinguishing Mechanical Faults from Fibre Clustering

Mechanical failures produce geometric spectrogram signatures with distinct structural qualities. A defective bearing on a bottom steel drafting roller creates a recurring chimney whose width at half-maximum amplitude spans less than five percent of the central wavelength. Bast fibre clusters create broad, bell-shaped curves spanning thirty to eighty percent of the central wavelength.

The digital signal processor calculates the quality factor of every detected peak to verify its origin.

Purchasers invoking ISO 2370 bind spinners to bundle linear density testing before sliver weight claims settle.

Peaks exhibiting a quality factor above twelve indicate mechanical failure requiring immediate technician intervention rather than autoleveler compensation. If an autoleveler attempts to level out a mechanical chimney caused by an eccentric delivery roller, it introduces a secondary mass fluctuation into the sliver. The autoleveler draft responds to the roller runout, creating a wave in the sliver that matches the roller defect but sits opposite in phase.

When this sliver enters the next drawing passage, the two defects decouple, yielding twice the original variation. How processing software can separate an eccentric front roller peak from an overlapping fibre grouping peak at high throughput speeds remains unresolved under current digital signal processing methods.

Roller

Mechanical displacement sensors mounted directly on the draw frame infeed channel register incoming flax mass through spring-loaded fluted discs. The sliver passes between a fixed anvil and a movable scanning wheel held under pneumatic loading. As sliver thickness fluctuates, the scanning wheel moves vertically, driving an inductive linear variable differential transformer.

The draw box chatters.

The physical geometry of the measuring wheel sets the physical resolution of the autoleveler. A scanning wheel with a diameter of 100 millimetres maintains a contact footprint of approximately twelve millimetres along the compressed flax strand. Mass variations occurring over distances shorter than this contact footprint undergo physical averaging inside the nip.

The sensor acts as a mechanical low-pass filter before any digital conversion occurs. Heavy slivers damp response.

Flax slivers generate high lateral pressures inside the measuring groove. Unlike resilient wool or crimped synthetic tow, unretted flax bundles lack internal spring recovery. Under a sensing nip load of 3.5 bar, woody core remnants flatten permanently.

The measuring unit measures compacted volume rather than absolute mass. Calibration curves convert displacement values into linear density by applying a bulk density coefficient derived from raw fibre retting values.

Displacement Dynamics in Infeed Scanning Units

The scanning roller assembly mounts upstream from the main drafting zone. The physical separation between the measuring nip and the front drafting roller constitutes the correction distance. This distance represents the transport delay of the autoleveling loop.

If the infeed speed runs at 30 metres per minute and the sensor sits 250 millimetres ahead of the front drafting nip, the flax requires 500 milliseconds to transit from measurement to attenuation.

1. Verify zero-point calibration of the displacement sensor using a precision steel gauge block inserted into the scanning disc channel.
2. Adjust pneumatic loading on the measuring disc assembly to 3.2 bar for dew-retted line flax or 4.0 bar for coarse tow slivers.
3. Feed six conditioned sliver ends into the guides while checking that lateral guide aprons prevent edge strand climbing.
4. Measure delivered sliver linear density on a wrapping reel and microbalance to establish the base drafting ratio under zero servo correction.
5. Inject a known mechanical displacement pulse through the calibration lever to measure physical actuator delay on the digital oscilloscope.
6. Set the proportional-integral-derivative gain parameters on the drafting servo drive to match the calculated transport delay window.

Actuator lag must balance transport delay with microsecond precision. Modern intersecting gills deploy low-inertia brushless AC servomotors coupled to planetary gearboxes driving the infeed rollers. The servo motor accelerates or decelerates the infeed roller while the front delivery cylinders maintain constant linear speed.

If the servo responds five milliseconds too late, the drafting compensation lands on the wrong segment of sliver, worsening output regularity.

Kinematic Response Lags and Nip Distance

Faller bars in intersecting gill boxes complicate the drafting kinematics. The pinned bars penetrate the sliver strand to restrain fibre velocity, creating an inter-pin drafting zone. Pins control bundle movement between the back feeding rolls and the front drawing cylinders.

Flax fibres do not accelerate uniformly when passing into the drafting zone. Technical bundles slip along the pin surfaces until front roller nip pressure overcomes pin friction.

Delivery cylinders present constant mechanical wear risks. A single burr on the front delivery cylinder produces an unlevel spot every 236 millimetres on the delivered strand. The machine builder asserts that routine cleaning routines and periodic visual checks by floor operatives entirely eliminate these surface marks without automated monitoring.

Span

Working calculations for draft intervention demand exact boundary limits derived from real mill conditions. To establish operational thresholds, consider a drawing passage handling dew-retted water-hackled line flax graded at Courtrai Mark 4. The process line inputs six sliver strands of 22 ktex each into an intersecting chain gill, yielding a combined input web of 132 ktex.

The machine operates at a nominal draft ratio of 7.33, producing a single leveled delivery sliver of 18.0 ktex at a delivery speed of 240 metres per minute.

Infeed velocity calculates directly from delivery speed and draft. At a draft of 7.33, infeed speed runs at 32.74 metres per minute, or 545.7 millimetres per second. The physical distance from the centre of the measuring disc nip to the gripping line of the front delivery cylinder measures 285 millimetres.

The transport time delay across this span equals 522 milliseconds. The digital servo drive requires 18 milliseconds to calculate speed commands and achieve torque adjustment. The electronic controller stores measuring disc data in a shift register for 504 milliseconds before commanding servo velocity modulation.

Bundle length distributions dictate spectrogram intervention boundaries. Laboratory testing under ISO 6741 gives a mean bundle length of 260 millimetres with a length coefficient of variation of 44 percent. The natural drafting wave calculated for this bundle distribution establishes its lower threshold at 1.8 times the mean length, or 468 millimetres.

The upper boundary extends to 3.8 times the mean length, reaching 988 millimetres. Multiplied by the post-draft ratio of unity at delivery, this drafting wave occupies the 0.47 to 0.99 metre span on the delivery spectrogram.

Leveling drives correcting short staple anomalies introduce periodic thin spots into flax roving.

The autoleveler sets an amplitude ceiling across this drafting span. Normal baseline amplitude within this window registers at 1.15 squared millimetres. The dynamic intervention threshold establishes its trigger point at 1.85 squared millimetres, fifty percent above the normal drafting wave crest.

Any broad swelling below this ceiling passes through without servo draft modulation. A chimney spike exceeding this limit by thirty percent with a quality factor over eight generates an alert and locks out autoleveler compensation.

Worked Draft Adjustment Calibration Scenario

Draft modulation ranges stay bounded to protect yarn count uniformity. The servo drive limits dynamic draft adjustments to a range of positive or negative twenty percent around nominal draft. At nominal draft 7.33, the operational band spans from 5.86 to 8.80.

If an incoming web section drops from 132 ktex to 115 ktex due to an empty can runout, the autoleveler increases back roll speed, lowering draft to 6.39 to hold delivery count at 18.0 ktex.

Speed exaggerates drafting error. When delivery speed accelerates from 150 to 250 metres per minute, friction between the flax bundles and the faller pins drops by twelve percent. This loss of inter-fibre control broadens the natural drafting wave and shifts the peak wavelength twenty percent longer.

The controller applies a speed-dependent offset matrix to the spectrogram thresholds.

• Feed Splicing Overlaps demand rapid transient clamping when sliver ends double over a thirty-centimetre span.
• Retting Band Variations generate sustained low-frequency mass steps extending across fifty to two hundred metres of sliver.
• Pin Bed Deflections manifest as wandering intermediate chimneys when bent fallers skip over coarse bundle knots.
• Tension Draft Drifts occur downstream when calender flutes pull sliver faster than delivery cylinder release speeds.

Spectrogram Peak Suppression under Varied Moisture

Atmospheric humidity alters the electrical capacitance and mechanical bulk compression of bast strands. Water molecules penetrate the secondary cell walls, swelling flax bundles by up to fifteen percent in diameter while increasing inter-fibre friction. At 75 percent relative humidity, the drafting wave amplitude rises to 1.72 squared millimetres on the delivery spectrogram.

Pectins soften under heat.

Dry conditions below 55 percent relative humidity trigger static cling and fibre fly, generating false mass spikes at the infeed scanning unit. When ambient moisture drops, the threshold algorithm lowers the trigger limit for long-wavelength drift while widening the intermediate filter band. Process stability improves when mechanical settings honor the natural behaviour of the fibre.

Defect

Mass irregularities allowed through the draw frame propagate directly into spinning spindle stoppages and uneven cloth shades. An undetected long-wavelength draft error produces count variation that escapes notice during roving and enters the wet-spinning room. Wet troughs expose defects.

During wet spinning, roving strands pass through a water bath heated to sixty degrees Celsius to soften remaining pectins before final attenuation.

Hot water dissolves inter-bundle bonds, reducing the strand to elementary ultimates measuring twenty to thirty millimetres. If the roving contains a hidden periodic drafting defect introduced by an improperly configured autoleveler, the wet-spinning frame amplifies this defect. Thin spots in the roving possess fewer ultimates to resist drafting tension, leading directly to end breaks at the spinning guide.

Roving tension breaks down.

Greige linen reveals periodic thickness variations as crosswise stripes across plain woven sheeting.

Late correction ruins cloth. Periodic irregularities passing into finished linen yarn create visible optical defects in woven goods. When woven into high-density plain weaves, a repeating mass anomaly with a wavelength matching the fabric reed width creates harsh horizontal bands known as barre marks.

In wet-spun yarns destined for apparel or fine tableware, barre defects cause rejection of entire dyed dye lots because yarn mass variations absorb dyestuffs unevenly.

Downstream Breakage in Wet Ring Spinning

End break rates correlate exponentially with yarn count variation. A mill spinning 26 lea yarn (63.5 tex) from hackled line sliver experiences six to eight end breaks per hundred spindle hours under normal processing conditions. When sliver autoleveler thresholds drift, allowing a three-percent rise in short-term unevenness, spinning break rates jump to eighteen breaks per hundred spindle hours.

Operator efficiency collapses as ring spinners spend their shifts piecing broken ends.

Pieced yarn joints introduce knots or mechanical splices that cause warp breakages during high-speed weaving. An automatic loom inserting picks at six hundred metres per minute stops instantly upon yarn break detection. Every weaving stoppage leaves a starting mark in the linen web, downgrading premium piece goods to second quality.

False draft destroys yield.

Economic Losses from Periodic Mass Irregularity

Financial margins in linen manufacturing rest on yarn count control and processing waste reduction. Hackled line flax trades at significant price premiums over short tow fibres. Converting raw flax sliver into delivery cones generates substantial financial losses when autolevelers wander out of calibration.

A spinner producing fifty tonnes of 36 lea yarn per month absorbs thousands of dollars in off-grade yarn discounts when sliver count drifts beyond commercial tolerance.

Incorrect autoleveler threshold determination causes either destructive over-leveling or unchecked mass drift, driving yarn count dispersion outside trade limits, elevating wet-spinning end breaks, and forcing weavers to downgrade finished linen fabrics to discounted commercial grades.