In direct warping and sectional warping installations, beam edge density is governed by the cumulative radial stress applied to the yarn layers at the flange-adjacent zones. The creel tension map is the spatially indexed set of per-end tension values collected across the creel matrix, with row, column, side, and package-path data attached to each record. On a high-speed filament line running at 600–1,200 m/min, the edge positions of the beam, commonly the first and last 60–100 mm of winding width, are supplied by the outermost creel columns. These columns are more exposed than centre columns to path-length differences, guide-angle accumulation, and brake drift. When the tension map shows a systematic offset between outer and centre columns, the beam develops edge density deviation. The defect may appear as soft edges, hard edges, or asymmetric flange build-up. A well-constructed tension map does more than display absolute tension; it separates the geometric tension component from the brake-offset component, converts the tension history into a radial density signature, and drives either closed-loop electronic tensioners or manual creel interventions. The industrial requirement is not uniform tension for every end, but a tension profile that produces an acceptable Shore A hardness profile across the beam after doffing. Commissioning data on V-creel and linear creel machines indicates that tension mapping is most effective when the measurement is synchronised with beam length rather than elapsed time, because edge density is a function of accumulated stress per radial millimetre, not instantaneous speed.

What Happens at the Flange When the Last Eight Creel Columns Carry a Residual Tension Offset?

For a beam width of 1,800 mm and 1,200 ends, the outermost eight columns contribute 8/1,200 of the ends but their yarn paths are the most angular. A residual tension offset of 0.03–0.08 cN/dtex on those columns is sufficient to generate a measurable edge hardness change. The mechanism is not simply a change in yarn lay tension; it is a change in the local packing density as each subsequent layer compresses the previous one. The edge zone receives yarn under elevated stress, so the yarns are elongated more during wrapping and then relax after laydown, generating a local squeeze on the underlying layers. Over thousands of layers, the edge radial density increases, and the beam surface hardness climbs. If the offset is negative relative to centre columns, the edges remain softer, allowing the pressure roller to act more deeply at the flange and sometimes producing barrel sag at the edges. If the offset is positive, the edges harden first, which shifts the pressure roll reaction away from the centre and can create a cupped diameter profile. The map records this offset at each column; the analysis then identifies whether the last eight columns share the same offset or whether only the final two columns are drifted. Shared offset across eight columns usually indicates a machine-level geometry problem or a side-specific creel tensioner calibration error. Drift confined to the final one or two columns usually indicates package diameter differences, worn guides, or brake hysteresis in the outer positions. Field records on sectional warpers show that edge density defects are often preceded by a gradual tension splitting between the left and right outer columns, developing over 2–4 doffs. When the map compares left and right side residual offset for the edge band, the onset of asymmetry is usually visible before the durometer reading moves.

Node addressing and the conversion of creel position to beam edge coordinate

Electronic creel tension mapping systems assign each tensioner a physical address that encodes side, column, and row. On a 1,200-end V-creel, the node address is translated into a beam lateral coordinate by applying the reed geometry, the v-reed convergence factor, and the warper's lateral spreading profile. The outermost columns on the left side map to the beam edge from 0 mm to 65 mm; the outermost columns on the right side map to the beam edge from 1,735 mm to 1,800 mm. The exact mapping is not linear for all creel designs, because the angle at the v-reed changes as a function of column. Tension samples are taken from each position at update rates between 100 Hz and 500 Hz depending on the fieldbus cycle. Each sample is time-stamped with the machine's beam length encoder pulse, permitting the conversion from a speed-dependent time series to a radial build history. Three-roller strain gauge tension sensors are typically placed at the output of each tensioner; these sensors measure the normal force on the centre roller and convert it to yarn tension through a calibration constant. The calibration is performed with a traceable force reference per ISO 7500-1:2018 class 0.5, and thermal compensation is applied for the 20–45 °C machine envelope. The signal is filtered with a low-pass cutoff of 10–20 Hz to remove package ballooning noise while retaining the lower-frequency tension drift that maps into beam edge density. The recorded tension matrix is then reduced to an edge-band statistic: the median tension of the first and last 8–12 columns per side is subtracted from the median tension of the centre 40 % of columns. This side-to-centre differential is mapped against beam radius and compared with a control band. When the differential exceeds the control band for a defined radial segment, a flag is generated.

Closed-loop electronic creel tensioners use the mapped differential as setpoint trim. The edge-band differential is passed through a deadband of ±0.02 cN/dtex; if the differential remains outside this band for more than 3–5 s of winding, the tensioner output for the affected columns is adjusted by a proportional-integral control loop. The tuning constants are machine-specific, but a starting integration time of 0.5–1.0 s and a proportional gain of 0.05–0.15 cN/dtex per 0.01 cN/dtex error are used on commissioned lines. The control action must be gentle because an aggressive edge-tension correction can produce a new density defect: a local overcorrection creates a ridge at the shoulder of the beam edge. The maximum permitted rate of tension change per position is typically 0.01 cN/dtex per second; this prevents tension shocks that could break filaments or cause slubs. The control system also serialises its actions by creel side, correcting left edge and right edge separately to avoid introducing a side-to-side asymmetry. The winding speed is ramped to full speed only after the mapped differential is within the control band; otherwise the first 20–30 mm of beam build inherits an uncorrected edge density. Each end's tensioner is individually addressable over a fieldbus such as an IEC 61158-compliant EtherCAT or CANopen network, allowing per-position loop times that are shorter than the dynamic response time of the yarn path. Mapping also allows cascade control: a slow outer loop trims side-to-side tension difference based on post-doff durometer data; the fast inner loop maintains the distributed per-end tension profile. This cascade structure is important because beam edge hardness cannot be measured in real time on most machines; the durable edge density error is detected only after doffing. Therefore the tension map is a feedforward and feedback surrogate for the unmeasurable density during winding.

Process conflicts intensify when the warping speed is increased to reduce cycle time. At speeds above 900 m/min on textured polyester, the balloon geometry at the package changes, and the effective tension measurement at the tensioner output includes a larger aerodynamic component. The tension map then overstates the yarn stress at the beam for outer columns because the balloon length is longer. This aerodynamic offset is speed-dependent and can be estimated by running the same creel at three speeds: 300 m/min, 600 m/min, and 900 m/min. The edge differential is then plotted against speed; if the slope exceeds 0.04 cN/dtex per 100 m/min, the map cannot be corrected by brake trim alone and must be supplemented by package positioning changes. Some mills insert balloon breakers or swivel guides at the outer columns to neutralise the speed-dependent component. The beam edge processing window is therefore constrained by both tensioner response and yarn balloon behaviour.

If the Creel Uses Passive Ball-Bearing Tensioners Instead of Closed-Loop Electronic Brakes

Passive tensioners cannot execute automatic per-end trim. The tension map is reduced to a historical record and an operator adjustment guide. The first step is a static map at doff or thread-in speed; each tensioner is set to a target using a hand-held tensiometer. The second step is a dynamic map at production speed because passive tensioners exhibit speed-dependent hysteresis. The difference between the static and dynamic maps is used to identify which positions require cleaning, ball-bearing replacement, or package rotation. On a V-creel of 1,200 ends, edge columns with a dynamic-to-static tension difference of more than 0.06 cN/dtex are flagged for mechanical service. Passive edge control is obtained by assigning packages with controlled diameter and consistent unwinding characteristics to the outer columns, and by adjusting the creel column alignment relative to the v-reed. Some installations use side-specific passive pre-tensioners to compensate for path geometry, but the compensation is fixed and cannot follow package diameter decay. The beam edge density response is therefore more variable across the batch. If the package lot has a diameter coefficient of variation above ±3 %, the passive creel cannot maintain edge density within a narrow Shore A band without physical sorting. Mapping in passive creels is still necessary because it identifies the edge columns that drift first; otherwise the warp beam is doffed before the edge density error becomes visible.

After doffing, the beam is placed in a controlled environment for a stabilisation period before surface hardness mapping. The hardness probe is applied along the winding surface at 100 mm intervals, with additional measurements at 0, 25, 50, 75, 100 mm from each flange. The recorded Shore A values are compared against a control band established for the beam build. Density is inferred from the surface hardness because direct density measurement on a wound beam requires either calculated volumetric mass or destructive sampling. The standard operation calibrates the durometer per ISO 7619-1:2010 but notes that the Shore A reading on a wound yarn package is a comparative process metric, not a material hardness value. Edge density acceptance is usually defined as a maximum side-to-centre hardness difference of 4–6 Shore A points. A larger difference triggers immediate review of the tension map from the same beam, starting with the outermost columns. If the creel mapping data shows a corresponding edge tension offset, the next beam is corrected by derating the flagged tensioners. If the durometer profile is asymmetric, the tension map is inspected for side-specific path or brake asymmetries. The correlation between durometer edge reading and tension map edge differential is strongest for beams doffed at stable ambient conditions; variations in relative humidity above 65 ± 5 % RH alter the friction coefficient of the yarn and weaken the correlation. Published data for the specific transfer function from edge tension to Shore A hardness on all yarn classes is limited; the numerical control limits are therefore developed from machine commissioning records and should be revalidated after any change in yarn supplier or package geometry.

A Single Beam Sometimes Transitions from Soft Edge to Hard Edge Within Two Doffs

The tension map from a single beam is not a static entity. Package diameter decay changes the unwinding path length and the moment arm at the tensioner. The outermost creel positions are more sensitive to package diameter because the yarn balloon diameter is larger and the package radius decreases faster relative to the fixed guide. As the beam build progresses, the pressure roller penetration also changes the force distribution. A beam that doffs with acceptable edge density after one lot change can doff with hard edges on the next lot if the incoming packages have a different diameter or if the ambient humidity changes by more than 5 % RH without reconditioning. To handle this transition, the tension map is collected for every doff and saved by lot number, package lot, beam build length, and warp speed. This historical map is used as a roll-forward diagnostic: if the second doff of a new package lot shows an edge-band differential above 0.05 cN/dtex while the first doff was inside 0.02 cN/dtex, the beam edge density is likely to shift before the third doff. The correction window is narrow because the next beam is already in process. For operations without electronic tensioners, the safe response is to reduce warping speed by 15–25 % and increase the pressure roll load by a small amount to compensate for reduced packing efficiency. Published data for the exact speed-to-density transfer function in warping beams is limited; the available commissioning evidence is machine-specific and cannot be generalised without a defined beam width, yarn type, and filament or spun yarn composition. The mapping system should therefore be treated as a process control tool, not as a guarantee of edge density uniformity.