What "tensioning" is not
The word invites a wrong picture. Pull hard on a string, the string gets longer, done. That description would be correct for a rubber band and roughly correct for a steel cable. It is almost entirely wrong for a bundle of UHMWPE and LCP filaments held together by a helical twist.
A bowstring under a rising load does something more interesting than stretch. Individual filaments shift laterally within their strand. Individual strands shift laterally within the bundle. Air is displaced from between fibers and replaced by wax as the internal temperature rises. Filaments that were sitting a little too high or a little too low in the bundle slide until they find the layer that matches their unloaded length. The bundle contracts radially as the twist self-clamps. And underneath all of that motion, the total length of the string changes by an amount so small that a shop tape can barely detect it.
Tensioning is a re-arrangement operation, not a lengthening operation. The load is the environment inside which the re-arrangement is allowed to happen.
The load ramp
Every bench that produces stable strings uses some version of the same load sequence: a low equalization pass, a moderate working load, and a high burnish-and-hold. The specific numbers below are Axial's — the loads and dwell criteria this bench has converged on. They are not published material-science thresholds; they are a working recipe.
On the Axial bench: 40 lb for equalization, 50 lb for twisting, 100 lb for measurement, 200 lb for the tensioning hold, and 300 lb for the final stretch after serving. Each of those loads is doing a different job on the bundle, and no two of them are interchangeable.
The 200 lb hold is the load this bench uses for the main dwell. It sits at a useful point in the material's response: high enough that every filament in the bundle is participating in the load, low enough to stay clear of the region where local filament stress starts leaving marks.
Dwell — the load-time product
Load alone does not settle a bundle. Load applied for a specific length of time settles a bundle. The name for this in polymer mechanics is dwell: the interval during which a viscoelastic material is held under load so that time-dependent deformations can occur.
UHMWPE — Dyneema, the primary fiber in 452X — is a strongly viscoelastic material. Its stress-strain curve is not a line. It is a family of curves, and which curve applies at any given moment depends on how long the load has been applied. Load the fiber briefly and it behaves stiffly. Hold the same load for minutes and the fiber slowly elongates as its polymer chains re-orient along the tension axis. Hold it for hours and the elongation continues, more slowly, but without an obvious upper bound. That continuing elongation is creep, and it is the enemy of a stable bowstring in service.
The purpose of the tensioning-and-burnishing dwell is to eat that creep at the bench, so it doesn't happen on the bow. A string that has been held at 200 pounds for six or seven minutes at 115°F (burnishing) has already experienced the polymer chain re-orientation that would otherwise happen over the customer's first several hundred shots. When that string comes off the jig, its creep budget for the next thousand cycles has been substantially spent. What remains is a small, predictable equilibrium settling, not the large, unpredictable settling of a bundle that was never given time under load at the bench.
The drop-rate signal
A load cell attached to a string under a fixed displacement shows something instructive. The moment the string reaches 200 pounds, the load cell reads 200 pounds. Hold the string at that displacement without adding any more, and within seconds the reading begins to fall. Not because the string is failing. Because the string is elongating microscopically as its filaments re-arrange, and each micron of elongation relieves a small fraction of the total load.
The drop rate is fastest in the first thirty seconds. It slows through the first minute. By minute three or four, it has fallen to a small fraction of what it was at the start. On the Axial bench, the tensioning hold ends when the drop rate falls below one pound per minute — the curve-slope threshold where the remaining creep is small enough to be finished by the final 300 lb stretch without leaving anything measurable on the customer's bow.
The threshold is not a stopwatch value. It is a curve-slope value. A string that has been under load for six minutes and still shows more than 1 lb/min of drop is not ready. A string that reaches the threshold in four minutes is ready in four. The bundle decides, not the clock.
The temperature dependence
The creep rate of UHMWPE is strongly temperature-sensitive. That is well established in the polymer literature: higher temperature (within safe limits) lowers the activation energy for polymer chain motion, so the same amount of settling completes faster. Heat is a bench multiplier, not a cheat.
Axial's burnishing window is 110–120°F. That range does three useful things at once:
- Chain mobility is high enough that the tensioning dwell completes in minutes rather than hours.
- Traditional bowstring waxes reach a soft-flow state that migrates cleanly into the bundle instead of pooling on the surface.
- Vectran, where present, is unaffected — LCP fibers are stable well above 200°F.
- UHMWPE stays comfortably below its continuous-use temperature ceiling. (DSM cites Dyneema's continuous-use maximum around 70–80°C / 158–176°F; 110–120°F leaves comfortable margin.)
Below 110°F, the dwell takes too long to be practical. Above 120°F, wax starts to migrate out of the bundle instead of into it, and the fibers move closer to a temperature ceiling that is worth respecting. The window is narrow, and the narrowness is a feature — it means the process is well-defined.
The specific-load, specific-time hypothesis
A useful way to think about the tensioning stage is as an integration: creep-experienced equals load times time, adjusted for temperature. Any polymer scientist would recognize this as the loose form of a time-temperature superposition principle — the observation that a viscoelastic material's response at low temperature over long times can be predicted from its response at high temperature over short times. The bench recipe encodes this principle in tools:
| Bench variable | What it controls | Effect on the bundle |
|---|---|---|
| Load magnitude | Depth of the settling | Higher load reaches into filaments that lighter loads never engage. |
| Dwell time | Completeness of the settling | Longer time allows more creep to occur at the current load and temperature. |
| Temperature | Rate of the settling | Higher temperature (within window) speeds the same total settling. |
| Radial pressure | Contact between filaments | Set by twist rate and axial load; determines whether wax and fiber-to-fiber contact are close enough to interact. |
Every stage of the tensioning and burnishing process is tuning one of these variables. A string that fails to stabilize was almost always shortchanged on one of them: pulled to a lower load than 200, held for less time than the drop-rate threshold required, burnished at a temperature outside the 110–120°F window, or twisted at a rate that left the bundle too loose for radial pressure to develop.
What settling looks like from the outside
An archer will never see any of this happen. What they will see is the difference between a string that stabilizes in the first hundred shots and a string that never quite stabilizes at all. That difference — the difference between a peep that stays put after the first tuning session and a peep that keeps rotating for a month — comes down entirely to how much creep was consumed at the bench versus how much was left for the archer to consume in service.
Connections
This is why strand count matters beyond raw strength: more strands lower the load per filament, which lowers the creep rate at any given total load, which shifts the dwell curve. It is why the center serving must be applied at the same 100-pound load the bundle was measured at — and is expected to operate at: any other load leaves the served geometry mismatched to the settled bundle. It is why the Part 5 tensioning stage is not a checkbox but a decision, made against a measured curve.
← The helix · Science & Mechanics
Published 2026-07-04 · Axial Bowstrings