Big static discharges can screw up all *sorts* of electronic equipment, and metal detectors are just sensitive inductive sensors.
Storytime: The local paper mill had a packaged product distribution conveyor that was chewing up bearings because it was discharging static through them.
The solution was to switch to an insulated bearing. No more discharge through the bearings, right ? Problem solved !
A new problem appeared; the VFD that ran that conveyor would stop and recover a couple of minutes later. The only evidence ? A diagnostic code on the DeviceNet scanner. This became known as a "DeviceNet problem" and then a "DeviceNet bug" and then eventually led to a Company-wide e-mail blast that there would be no DeviceNet used anywhere because it was junk and unreliable and nothing Rockwell was able to do would help and Ken Roach was an idiot because he said the problem isn't the network.
Bearings are replaced by millwrights. Different union than the electricians who replace VFDs.
But both of them came to see when management hauled me up there to answer for the "DeviceNet bug". I brought my oscilloscope, and my protocol analyzer, and started asking questions.
At one point I had my hand on the conveyor frame and on a handrail when the steel laces came around to the head pulley and discharged static.
"OW ! Mother****er ! What was that !"
The millwright laughed. "Oh, yeah, it'll bite ya. I used to have to replace the bearings on the head pulley all the time. I put in a plastic hub bearing."
The electrician took notice. "You did ? Where's the grounding wire for this section ?" We found it hanging, probably left that way by a welder who repaired a walkway handrail.
With no route to ground through the bearings, and no route to ground through the frame, the static built up to voltages high enough to jump around the pulley bearing race and went to ground through the only path... right through the VFD's chassis ground wire.
The reason the VFD showed a DeviceNet error code is because it was rebooting after having its memory corrupted by the voltage spike.
Years later, I hired a new engineer who had been a junior guy at the paper mill around that time. I was showing him how to configure POINT I/O on DeviceNet, and asked if he had any experience with it.
"Oh, we had a little of it at the mill. But it was too unreliable. There was this one conveyor that would fault all the time because of the network and Rockwell never could figure it out."