The Guard That Cracked When a Part Flew Out: Impact Resistance, Not Decoration
A machining cell with a clear polycarbonate guard. The door was 5 mm thick, bolted in an aluminum frame. A broken tool sent a 50 g steel chip across the cell at 30 m/s. The chip hit the guard at chest height. The guard cracked — a 300 mm spiderweb from the impact point. The door held (no breach), but the crack meant the guard needed replacement. We changed the spec. This is how guards should be sized for impact.
The energy of a flying part
A 50 g chip at 30 m/s has kinetic energy:
E = ½ · m · v² = ½ × 0.05 × 30² = 22.5 J
A 5 mm polycarbonate sheet stops about 10–15 J before cracking. We were at 22.5 J. The guard cracked — it absorbed the energy but fractured. It didn’t shatter (polycarbonate is ductile), but the crack compromises future impacts. The next chip might breach.
Guard material and thickness
| Material | Thickness | Impact energy (J) | Use when |
|---|---|---|---|
| Acrylic (PMMA) | 5 mm | 5–8 J | Low risk, no flying parts |
| Polycarbonate (PC) | 3 mm | 8–12 J | Light machining |
| Polycarbonate (PC) | 5 mm | 15–25 J | General machining |
| Polycarbonate (PC) | 8 mm | 30–50 J | Heavy machining, casting |
| Wire mesh (steel) | 3 mm wire | 50+ J | Grinding, high-mass parts |
| Safety glass (laminated) | 6 mm | 20–30 J | Visual access, low speed |
What we changed
We upgraded the guard from 5 mm to 8 mm polycarbonate. The 8 mm PC stops 30–50 J — more than the 22.5 J chip energy. The guard didn’t crack when we tested it with a controlled impact. The frame was reinforced with gussets (the original frame bent slightly at the impact point).
For the heavier machining cell (1 kg parts at 20 m/s = 200 J), we switched to wire mesh with a 3 mm polycarbonate inner layer. The mesh stops the heavy projectiles; the PC stops the small chips and provides visibility.
The guard door interlock
A cracked guard is a breach even if it doesn’t open. The safety circuit should monitor the guard. We added a safety interlock switch on the door — if the door is opened (or the guard panel is removed), the machine stops. The cracked guard triggered the interlock when the operator opened the door to inspect. The machine wouldn’t restart until the guard was replaced.
For a production machine, the guard is a safety component, not a cosmetic cover. It needs to be rated for the worst-case projectile in the cell. Don’t spec 5 mm PC for a machining cell that throws 50 g chips at 30 m/s. The math says 15 J; the reality is 22.5 J. Oversize by 2×.
The mounting matters
A guard that’s clamped in a frame is stronger than one that’s screwed to a thin extrusion. The original guard was held by 4 screws in a 30×30 aluminum extrusion. The impact bent the extrusion. We upgraded to a 40×40 extrusion frame with a clamp bar around the PC panel. The frame didn’t flex. The panel took the full impact without transmitting energy to the frame.
The cost of a cracked guard
A cracked polycarbonate panel is about $50. A breach (chip through the guard) could injure someone. The 8 mm panel is $80. The 30% extra cost for the thicker panel is trivial compared to the downtime and risk. We now spec 8 mm PC as the default for any cell with rotating tools. 5 mm is for light assembly with no projectiles.
The guard rule: Calculate the worst-case projectile energy (½mv²). Size the guard for 2× that energy. Use 8 mm polycarbonate for machining cells. Wire mesh for high-mass projectiles. A cracked guard is a failed guard — replace it, don’t tape it. The guard that cracked was doing its job; it was just undersized for the impact.
