Why can't a lidar and a drive wheel share one compartment?

A: The requirements are opposite. The lidar needs clearance; the 12 kg wheel needs preload and a hard stop, so foam soft enough for the wheel cannot hold the unit.

Is a 5 mm window gap enough in transport?

A: Yes, if the other five faces are restrained. The 5 mm covers foam recovery plus wall deflection: about 1.2 mm at 3 g, plus 0.5 mm of compression set.

Where does the 2 mm axial limit come from?

A: From the gearbox input shaft tolerance. Most AGV gearboxes accept 0.05 mm of radial eccentricity, and at 150 mm diameter a 2 mm axial offset is about 0.03 mm at the rim.

Is a 4 mm wall too thin for this duty?

A: The case sees point loads, not distributed pressure. Four millimetres of PP plus six ribs 3 mm high gives a section modulus close to a 7 mm flat panel while saving 1.8 kg.

Can one enclosure hold spares for four vehicles?

A: It can, but swap time goes from 90 s to 4 minutes, and four lidar bays push the weight past 45 kg, needing two people.

AGV Spares Case

Quick answer: An agv spares case fails at two contact faces: the lidar window that must never touch foam, and a 12 kg wheel needing 2 mm axial play.

Where an AGV spares case actually fails
Why the optical window must not touch foam
Six faces of the lidar bay
Axial end stops and the 8 mm damper
Reaching a 90 second swap
FAQ

1. Where an AGV spares case actually fails

A welding line at a vehicle plant runs 26 automated guided vehicles and keeps four spare-part enclosures. Twice the fitters reported the same pattern: a freshly fitted lidar lost point density within three days, and a new drive wheel pulled the vehicle 15 mm off line over a 40 m straight. Neither fault traced back to electronics.

The lidar died at its window. The coating is a few hundred nanometres thick, and one slide against foam leaves a scratch 2 μm to 5 μm deep. On a shop floor at 45 ℃ the foam also releases low-molecular compounds that settle on the glass as a faint haze. Transmission falls 3% to 5% before anyone notices.

The wheel died along its axis. A 150 mm, 12 kg wheel hangs on a φ20 mm shaft whose only axial stop is one circlip; a forklift knock shifts it 2 mm to 3 mm, and the wheel no longer runs concentric with the gearbox.

Adding 2 mm of wall thickness fixes neither problem.

Spare	Symptom	Deviation	Case-side fix
Lidar φ70×80 mm	Sparse cloud	3% loss	Window left open
Wheel φ150 mm	Straight-run drift	3 mm float	Stops at 2 mm
Hub bearing	Grinding	0.1 mm	8 mm pad
Lidar plug	Dropouts	4 mm	8 mm slack

2. Why the optical window must not touch foam

Foam is soft, but what it does to a window is not. Expanded polyethylene and EVA outgas slowly under sustained load, and the residue lands on the coating as an invisible film. Abrasive dust in the cell structure also rubs across the glass through every vibration cycle.

AGV spares case open to show die-cut foam pockets, one square bay and several long slots, egg-crate lid foam, twin latches

The fix is an air gap in front of the window. The cavity wall carries a 5 mm step, and the glass touches only a 1.5 mm ring at the edge of that step. The ring is faced with 2 mm black EVA at 30 Shore A, soft enough to stop rattle and too soft to carry load.

The bay also has to be sealed off from the wheel bay. A 12 mm EPE partition separates them, its top edge standing 8 mm proud of the foam face. A 3 mm notch at the bottom stops the closing lid from pressurising the cavity and lifting the foam.

A window can touch something soft. It just cannot be pressed against it.

3. Six faces of the lidar bay

Only one of the six faces is meant to be empty. The other five have to work, and getting any one wrong lets the unit rotate a few degrees under 3 g of vibration.

Bay face	Contact	Gap
Window side	Open air	5 mm
Window rim	2 mm EVA ring	1.5 mm preload
Floor	20 mm EPE	0 mm
Ceiling	10 mm EPE, webbing	2 mm
Sides	12 mm EPE	3 mm
Cable side	φ24 mm channel	8 mm

The floor takes the whole weight. EPE at 45 kg/m³ compresses about 1 mm under a 0.6 kg load and recovers in 2 s.

The webbing across the top is a retention device, not a clamp. Pulled snug it still leaves 2 mm to the lid, and a finger press shows movement.

The cable channel needs 8 mm of slack. Tie a harness tight and the load lands on the connector, where a 4 mm shift breaks contact intermittently.

4. Axial end stops and the 8 mm damper

A wheel wants to sink when the case lands and to slide along its shaft when the vehicle moves. The pad handles the first, the stop handles the second.

AGV spares case factory floor, red crane girders, blue racking, steel moulds on concrete, one green machine at right

Stop element	Build	Value
Axial blocks	15 mm PP	1 mm per side
Shaft bushing	3 mm steel, H8/f7	0.02 mm
Floor support	8 mm nitrile	1.5 mm compression
Tread	5 mm EPE wrap	Grooves clear

One millimetre per side gives the 2 mm float the specification asks for. The blocks carry no weight, so 15 mm of PP is enough, provided the block is moulded into the same rib network as the case floor. Bolted on afterwards, it will lean under the 12 kg inertia.

The 8 mm nitrile pad works on a different principle: it stretches the impact pulse instead of blocking it. On a 3 g sweep from 5 Hz to 200 Hz for 30 minutes, acceleration at the shaft drops from 3 g to 1.4 g, and hub clearance is still 0.1 mm after 500 km of road transport.

The tread stays unloaded: a 2 mm permanent set takes the effective diameter from 150 mm to 149.6 mm.

5. Reaching a 90 second swap

Ninety seconds is the number the maintenance room is measured against; a minute of lost output costs far more than the enclosure does.

Swap step	Time	Where it stalls
Release two latches	8 s	320 mm spacing
Lift lid to 100°	5 s	2 N·m hinge
Slide out tray	25 s	180 mm travel
Open webbing	20 s	One-hand buckle
Refit and reseat	32 s	Two bosses

The tray is the only moving part. It carries the unit out of a bay 180 mm deep so the operator never reaches past the window. Two cold-rolled slides rated 15 kg give the travel, a 1 mm stop at the end.

Two φ8 mm bosses on the bay floor solve the refit. The tray drops in one way only, more reliable than a label inside the lid.

Hitting 90 seconds comes from removing steps, not from moving faster.

Conclusion

An AGV spares case is a constraint problem on two faces, not a strength problem: 5 mm of air in front of the lidar window and 1 mm of clearance at each side of the drive wheel. Fix those two numbers and the remaining bays can follow a conventional insert layout.

This article only covers the enclosure side: its dimensions and what the design values mean. It does not address handling, loading or use of anything carried inside.