Troubleshooting

Why Your Thermal Spray Numbers Are Wrong: Common Coating Mistakes and Fixes

The costly errors that throw off thermal spray and hardfacing estimates, each with a symptom, root cause, and a numeric fix.

Symptom: you quote 2.4 kg of powder per part and the booth burns through 4.1 kg. Root cause is almost always a deposit efficiency (DE) assumption pulled from a spec sheet instead of your gun. Vendors quote DE at 65 to 75 percent for HVOF WC-Co, but a worn nozzle, a 15 degree spray angle off normal, or a 6 inch standoff drift can pull real DE to 45 percent. Fix: mask a coupon, weigh before and after, and back-calculate DE from the Powder Consumption calculator. If measured DE is 20 points below the sheet, every powder order is short by a third.

Symptom: coating comes in at 0.008 in when the print calls for 0.012 in, and the sprayer just adds passes. Root cause is treating thickness-per-pass as fixed when traverse speed drifts. A pass laying 0.0008 in at 500 mm/s drops to 0.0005 in at 800 mm/s. Fix: lock traverse in the program, not by hand, and verify pass buildup with the Coating Thickness Buildup calculator before committing to a pass count. Adding blind passes to hit thickness inflates powder and cycle time 30 to 50 percent and can overheat the substrate past 350 F.

Symptom: your grams-per-part looks right but the monthly powder invoice is double the forecast. Root cause is unaccounted overspray, which the per-part math quietly ignores. On small parts sprayed on a rotating fixture, 40 to 60 percent of powder never lands on the part. A 55 percent DE quote already bakes some of this in, but edge and fixture losses stack on top. Fix: run the Overspray Loss calculator against actual booth throughput. If part-facing area is under 30 percent of the spray cone, expect real consumption 2x to 3x the coated-surface estimate.

Symptom: wire-fed arc spray jobs run long and the wire spool empties faster than the deposition math predicts. Root cause is confusing wire feed rate with deposited mass. Two 1.6 mm wires at 150 g/min feed 300 g/min, but at 70 percent DE only 210 g/min bonds. Fix: separate feed rate from deposited rate in the Wire Consumption and Deposition Rate calculators, and never size a spool order off feed rate alone. Mixing metric wire diameter with imperial feed settings is a classic unit trap that silently doubles or halves the estimate.

Symptom: bond coat usage is consistently under budget on paper but the NiCrAlY drum runs dry mid-run. Root cause is estimating bond coat as a flat percentage of top coat instead of a real thickness. A 0.004 in bond coat under a 0.015 in top coat is not 5 percent of powder, it can be 20 to 25 percent once its lower DE and full-surface coverage are counted. Fix: size bond coat as its own layer in the Bond Coat Usage calculator with its own DE, typically 5 to 10 points below the ceramic top coat.

Symptom: the quote looks profitable until masking eats the margin. Root cause is masking treated as a fixed 10 minutes when geometry says otherwise. A part with eight threaded holes, two bores, and a sealing face can take 25 to 40 minutes to mask and demask, at a loaded rate of $45 to $65 per hour. Fix: estimate masking off feature count, not part count, using the Masking Labor Cost calculator. On low-volume aerospace work, masking labor often exceeds spray time by 2x to 3x and is the single most under-quoted line item.

Symptom: spray time per part looks fine but the booth still misses its daily part count. Root cause is ignoring non-spray time inside the cycle. The arc may run 90 seconds, but fixture load, preheat to 200 to 300 F, cool, and unload push real cycle time to 6 to 9 minutes. Fix: model the full door-to-door cycle in the Spray Time Per Part calculator and cross-check against measured Booth Utilization. A booth showing 35 percent arc-on time is normal, so sizing capacity off arc time alone overstates throughput by 2x.

Symptom: first-pass yield reads 92 percent but rework quietly consumes a fifth of the schedule. Root cause is counting only scrapped parts and ignoring strip-and-recoat cycles. A part reblasted and resprayed costs the original powder plus abrasive, labor, and booth time, often 1.5x to 2x the first attempt. Fix: log every recoat in the Rework Cost calculator, not just rejects. If measured DE, masking, and overspray inputs are guessed rather than sampled, rework rates of 8 to 15 percent are common and every wrong assumption above compounds into it.

Published 2026-07-01.