A manufacturer was replacing acrylic guards every six months. The guards were crazing around the bends and cracking at the mounting holes, not wearing out from impact. The material was not the problem.
The guards were failing in two distinct ways, and treating them as one problem is why the previous supplier never fixed it.
Crazing at the bends. Fine surface cracks radiating from the bend line. This is stress cracking: the bend put residual stress into the material, and the coolant mist provided the chemical trigger. Stress plus solvent is the classic craze mechanism in acrylic, and neither factor alone would have done it.
Cracking at the mounting holes. Cracks starting at the bolt holes and running into the panel. The holes had been drilled with the bolts bearing directly on the acrylic, so any vibration or over-tightening concentrated load at the hole edge.
Both symptoms pointed at the same underlying issue: the guards were designed as if acrylic were sheet metal.
Annealing after bending. This was the single biggest change. Thick bent guards were annealed to release forming stress before any hole was drilled. It added about one day to the schedule and roughly 12% to the part price, and it removed the crazing mechanism entirely.
Grade selection. The standard cast grade is not the best choice for continuous coolant exposure. We specified a grade with better chemical resistance and confirmed it by coupon test: offcuts immersed in the actual coolant for the expected contact period, then inspected for clouding.
Isolated mounting. Bolts now pass through compliant bushings, so the acrylic is never in direct contact with a metal fastener under load. Holes were routed rather than drilled, and every hole was placed at least 1.5× its diameter from the panel edge. No threads were cut into the acrylic.
Hole tolerance. The mounting pattern was machined to ±0.1 mm, which matters when a guard has eight bolts and the frame is fixed — a 0.3 mm error across the pattern means the last bolt will not go in.
The pattern was consistent: crazing around the bend, then a crack from the nearest mounting hole, within about six months. Neither symptom pointed at the sheet grade. Both pointed at the sequence of operations and at how the panel was held.
The original parts were bent first and drilled afterwards. That puts a hole in the highest-stress zone of the sheet — the bend region — and it puts it there after the material has already been worked. The corrected sequence is machine, drill, anneal, then bend. Annealing sits between the two because it releases both the manufacturing stress in the sheet and the stress introduced by machining, so the bend is formed in material that has somewhere to give.
The second change was to stop over-constraining the panel. A machine guard bolted rigid on four sides to a vibrating frame has nowhere to move, so the vibration concentrates at the holes. Slotted holes with a compressible washer let the panel breathe, and a 12 mm hole-to-edge distance instead of the original 6 mm removes the crack initiation site. Neither change costs anything per part.
| Symptom | Root cause | Change |
|---|---|---|
| Crazing around the bend | Bent without annealing | Anneal at 80 degrees C, 2–4 h, controlled ramp, before bending |
| Crack from a mounting hole | Hole drilled after bending | Drill before bending; 1 mm chamfer on the hole |
| Repeat failure at the same corner | Hole 6 mm from the edge | 12 mm hole-to-edge, external corner radiused 5 mm |
| Panel chattering in service | Rigid four-side fixing | Slotted holes, compressible washer, controlled torque |
The guards also moved from a generic 5 mm to a specified 8 mm on the doors, not because 5 mm was too weak but because an 8 mm door stops chattering against the frame at the machine's running frequency — a noise and wear problem that had been read as a material problem.
The new guards cost roughly 35% more per piece than the ones they replaced. Against that, the replacement cycle went from twice a year to beyond three years without crazing, and the downtime for guard replacement — which stopped a production line each time — disappeared.
The honest summary is that the previous guards were not cheap, they were inexpensive. The purchase price was lower and the cost per year of service was several times higher.
Replacing a failing guard set is usually a small order with a large consequence, which changes how it should be quoted.
| Item | Typical range | What drives it |
|---|---|---|
| Set size | 6–40 panels | Machine count and whether doors are included |
| Sampling | 3–7 business days | Whether a frame drawing exists; a photo with three measurements is faster than a site visit |
| Production | 7–15 business days | Annealing, which runs per batch rather than per panel |
| Cost split | Sheet 20–30 % | Machining and annealing dominate at this size |
| Service life after fix | 3–5 years instead of 6 months | Annealing plus a compliant mounting detail |
The economics are worth stating plainly. A guard set that fails twice a year costs the panels plus two unplanned line stops; the same set corrected costs perhaps 15 percent more per panel once and then runs for years. On a machine that runs three shifts, the line stop is the larger number by an order of magnitude.
For a plant with several machines, the useful first step is not a quote but a survey: photograph each guard, note whether it is bent, and check the hole-to-edge distance. That survey usually shows that two or three of the failures share one cause, which means one process change rather than a series of replacements.
More projects
Coolant, cleaning agents, ambient temperature and vibration all change what we specify. A drawing alone will get you a part that fits; the environment is what makes it last.
