Why the Reed Switch Fell Off: A Close Look at the Mount
A reed switch (magnetic cylinder position sensor) on a pneumatic cylinder. It clipped onto the cylinder’s T-slot with a screw and a plastic clamp. After three months, it fell off. The cylinder kept cycling — the sensor was on the floor. The PLC never saw the home position. The machine faulted out. We pulled the failed switch apart. Here’s what we found.
What the switch looks like inside
A reed switch is a glass capsule with two ferromagnetic reeds inside. A magnet (in the cylinder piston) brings the reeds together, closing the contact. The capsule is potted in plastic. The plastic housing has a dovetail that slides into the cylinder’s T-slot. A screw clamps the dovetail against the slot wall.
The failure wasn’t the reed capsule. It was the plastic clamp. After three months of vibration (the cylinder cycles every 2 seconds), the screw backed out. The clamp loosened. The dovetail slid out of the slot. The switch dropped.
Why the screw backed out
Three reasons:
The screw was a standard M4 machine screw, thread-formed into the plastic clamp. Plastic threads strip under vibration. A few dozen cycles and the clamp’s internal threads are worn. The screw holds nothing. It should have been a screw into a metal insert (brass or steel) molded into the clamp, or a separate clamping plate.
No thread locker. The screw was dry. Vibration + no locker = backed out. A single drop of Loctite 243 would have held for years.
The cylinder cycles every 2 seconds. That’s 43,200 cycles per day. The vibration from the cylinder’s end-stop impact shakes the switch every cycle. A slow cylinder (10 seconds per cycle) wouldn’t shake it loose. A fast cylinder does.
What we replaced it with
We swapped the clip-on reed switch for a magnetic sensor with a bolt-on mounting bracket. The bracket bolts to the cylinder’s end cap (not the T-slot). The sensor slides in the bracket and locks with a separate screw into a metal thread. No plastic threads. The bracket doesn’t vibrate loose because it’s bolted to the end cap, not clamped to the T-slot.
We also added a strain relief on the cable. The cable wasn’t pulling the switch off, but the cable’s weight on the sensor added a moment that helped the screw back out. A cable tie holding the wire to the frame removed that moment.
Mounting styles compared
| Mount style | Hold method | Vibration resistance | Best for |
|---|---|---|---|
| T-slot clip (plastic) | Screw into plastic clamp | Poor (fast cycles) | Slow cylinders, infrequent motion |
| T-slot clip (metal) | Screw into metal insert | Good | Most automation |
| Bolt-on end cap bracket | Bolt to cylinder cap | Best | Fast cycles, high vibration |
| Adhesive pad | Sticky pad | Only for static | Never on a moving cylinder |
The installation detail that matters
When you install a reed switch, slide it along the T-slot until the LED turns on (the piston magnet is close). Then slide it back until the LED turns off, then forward again until it just turns on. That’s the trigger point. Lock the screw. The switch should trigger just before the piston reaches end-of-stroke — not at the end, where the impact could damage the reed.
For a cylinder that needs to signal “home” at the retracted end, position the switch 2–5 mm before the piston fully retracts. The signal comes before the impact, not after.
The sensor that didn’t fall off
We now spec bolt-on brackets for any cylinder cycling faster than 1 cycle per 5 seconds. The clip-on T-slot sensor is fine for a clamp that moves once per minute. It’s not fine for a 2-second cylinder on a pick-and-place. The replacement sensor has been on for 14 months without touching. The screw is still tight.
The mounting rule: Reed switches don’t fail electronically. They fail mechanically — the clamp loosens, the cable pulls, the screw backs out. For fast cycles, bolt the sensor to the end cap. Add thread locker. Strain-relieve the cable. The sensor that stayed put wasn’t better — it was mounted better.
