Eric
You are correct, as long as the spindle impedance is still relatively low compared to the input impedance of the ESS input that is sensing that connection it will still work fine.
So of course you can’t really know for sure unless you make a spreadsheet…
I chucked up a 1/2" downcut endmill (unfortunately I don’t have any precision dowels on hand) and did 10 touchplate runs with the magnet on the bit, 10 with the magnet on one of the bolts that holds the tramming plate on, and 5 more with the magnet stuck to my rotary unit’s chuck. I gave the spindle a spin with my fingers between each one. I had to rehome on the 8th run because the magnet was on a dirty part of the bit and I had to EMO the machine.
I recorded the machine coordinates after each run, normalized it to graph with better resolution, and plotted it.
There is more variation in the X and Y than I’ve seen on past tests, but this bit was a little harder to align the edge to X and Y directions, so I think this was my ability to align the tip perpendicular to the Axis. Wish I would have had a dowel. Z is very tight. Because Z was so repeatable, you can see that the Z re-home did shift a couple thousandths.
From the data, you can’t see any variation between where the ground wire is connected, even on my spindle that has 2k ohm worth of variation in the bearing resistance.
At this point I would call grounding through the chassis more reliable than the magnet because the magnet wire can break, you can forget to put it on, or you can set it on some tar on the bit (that has happend a couple times to me).
I have a couple lasers with the touchplate rod that aren’t grounded so I may keep my magnet and wire on the touchplate, but its nice to have the option and backup through the chassis as well.