Apparently a 0.01 microFarad capacitor between the input and ground helps; I will see if I can find one today and report back.

Correction it is 0.1 uF capacitor, not 0.01. It prevents accidental tripping of the sensor by noise in the circuit, but doesn't help with the grounding issue per se.

Karl, no acknowledgement of Jeff's response? I am glad you posted as I am always confused by these terms. I have tried many techniques to eliminate or reduce noise and false tripping over the years. If I had any luck it was probably accidental. Jeff has a very good understanding of it and is very gracious in sharing. Safety in a home shop should be just as important. Who is going to see you being shocked? Getting cut by a spindle that didn't turn off? Or turned on when you were changing a tool? Self rescue? Just something to think about. We all think it will be someone else.

Oh sorry I have been too busy in the garage and I missed Jeff's post.

Point by point then:

1. I take your point (but who is Will Robinson?). Fools rush in where angels fear to tread. I assume the PDMX will start to melt if the spindle itself goes hot, which should be a good clue that there is a problem. I suppose one of the hot wires to the speed control could somehow contact the end of the green wire where is is screwed to the bearing; this is indeed a safeguard that has been undone, because that ground wire attachment is up above the plastic housing for the bearing in the same vicinity as the powered wires.

2. I understand this - I think. The green conductor is no longer connected to Earth. It is connected to the PDMX common pin, and to the armature. Therefore I am not using earth ground as a return path; I am only using a wire that used to be connected to Earth ground to carry a low voltage signal back to the breakout board.

3. I am not using Earth ground on the wire from the spindle (see above). Yes the wire is in the same bundle as the power wires to the router. Fortunately the router is off when I am touching tools to plates, so I don't seem to have the noise problem that other people refer to, where the LED trips prematurely and the Z axis goes up before touching the plate.

All in all it sounds like running a new wire is a good idea, or perhaps just disconnecting the green one from the bearing and putting a clip on it to hook to the bit. I'll probably do that when the current eight hour machine run is done, and in the meanwhile I'll unplug the spindle when changing tools (thankfully there is a plug on the Z axis).

Thanks for the input and your concern,
Karl

Looks like your "lost in Space"
 

OK the upshot of all this is that the porter Cable 892 basically stopped working. I had the bearings replaced for $50 and it died again, so I replaced it with a 690 and another 890 from eBay.

Overall I think the 690 is a much better built machine with a cast metal top bearing mount, rather than plastic, and a ground wire to the earth.

I am running another 890 with the alligator clip this time for the moment, until I can rig the SPID at which point I will switch back to the 690 for reliability. But the twin lead wire is a keeper; pretty easy to use.

 What encoder count did you use? Or did you simply use the single one?
Dan