Troubleshooting

Pultrusion Troubleshooting: 8 Costly Mistakes That Wreck Profiles and Margins

The eight process and estimating errors that quietly ruin pultruded profiles and margins, each with a symptom, root cause, and a fix tied to a real number.

The most expensive pultrusion mistake is running pull speed faster than the resin can gel. Symptom: transverse cracks, blistering, or an exotherm peak that migrates toward the die exit. Root cause is dwell time in the heated die dropping below the cure window, usually when someone raises line speed from 24 to 36 in/min to hit throughput without recomputing residence time. Fix: for a 36 in die and a resin needing 90 seconds at peak, max speed is 24 in/min, not 36. Check the Cure Die Dwell calculator before every speed change and treat 90 seconds as a floor, not a target.

Fiber volume fraction errors come from weighing glass by count of rovings and ignoring the mat and veil. Symptom: a profile that hits dimensional spec but fails flexural modulus by 10 to 15 percent. Root cause: operators track only the 62 continuous rovings and forget two layers of 0.75 oz/ft2 continuous filament mat, so the calculated Vf reads 55 percent while the real reinforcing roving fraction is closer to 42 percent. Fix: enter every reinforcement layer into the Fiber Volume Fraction calculator by yield (yards/lb) and areal weight, then reconcile against measured burnout at 565 C.

Unit slips between yield and linear density wreck resin and fiber cost quotes. Symptom: a quote that is off by a factor of 1.5 to 2 with no obvious arithmetic error. Root cause: glass yield is quoted as yards per pound (a 113 yield roving) while resin dosing is tracked in grams per foot, and someone mixes the two into one spreadsheet cell without converting. Fix: standardize on lb/ft for both, run Fiber Cost Per Foot and Resin Cost Per Foot in the same units, and sanity-check that fiber plus resin mass equals the Profile Weight Per Foot within 3 percent.

Ignoring resin pickup variability inflates the bath consumption estimate and hides real waste. Symptom: monthly resin usage running 8 to 12 percent over the calculated figure. Root cause: the estimate assumes 100 percent of bath resin ends up in the part, but drag-out, wet-out squeeze-off, and bath skinning add loss. Fix: apply a realistic 1.08 to 1.12 waste multiplier in the Resin Bath Consumption calculator, and if pot life is under 40 minutes at bath temperature, expect skinning losses at the high end of that band.

Quoting from nameplate line speed instead of effective throughput is the classic margin killer. Symptom: a job that looked profitable at 30 in/min but lost money on the floor. Root cause: die changeovers, roving creel splices, and start-up scrap drop real uptime to 65 to 75 percent, so 30 in/min nameplate becomes 20 to 22 in/min effective. Fix: feed measured availability into the Line Throughput calculator and quote from the effective figure. A 10 percent uptime error on a 5,000 ft run is 500 ft of unbilled capacity.

Treating startup and threading scrap as negligible burns fixed length off every run. Symptom: yield that falls short of the theoretical figure by a consistent 20 to 60 ft per job. Root cause: threading the die, reaching cure equilibrium, and trimming the lead end consume 15 to 40 ft before saleable profile emerges, and that length is nearly constant regardless of run size. Fix: model it with the Scrap Length Cost calculator. On short 300 ft runs, 40 ft of scrap is 13 percent of the job, so batch small orders to amortize the fixed loss.

Undercounting heater energy leads to phantom margin on thick sections. Symptom: energy cost per foot that is double the estimate on 0.5 in thick profiles. Root cause: die heater load scales with cross-sectional mass and cure exotherm management, not line length, so a thick section at slow speed draws far more kWh per foot than a thin one at high speed. Fix: run the Heater Energy Cost calculator per profile geometry rather than using a plant-wide average, and expect a 3x spread between a 0.125 in wall tube and a solid 1 in rod.

Building profile weight from nominal density rather than as-cured density skews every downstream number. Symptom: shipping weights and cost-per-foot that drift 4 to 7 percent from actuals. Root cause: using 1.9 g/cm3 handbook density when voids, filler loading, and real Vf put the cured composite at 1.75 to 2.05 g/cm3. Fix: back-calculate density from a weighed 12 in sample, feed it into Profile Weight Per Foot, and reconcile against the summed fiber and resin masses. A 5 percent weight error on a 40,000 lb annual volume is 2,000 lb of mischarged freight and resin.

Published 2026-07-01.