Hood,
  My Geckos, BOB, power supplies, etc. are mounted in an enclosure behind the mill - basically inaccessible, and a fair distance away. The PC is right next to the mill, a few feet from where I sit. So, the pendant is more logically connected directly to the PC, rather than to the BOB. So, a PC PP connection makes more sense.Â
Still dont understand, if your bob etc is in the enclosure then surely the cable comes from it to the PC? If so what is stopping you putting the SS inside the computer case, connecting the cable from the BOB to it and also connecting the cable from the pendant to the SS. As far as I can see it would be exactly the same cabling as you have now.
As for the performance aspects, as I reported in the original post, my pendant, connected through SS, is virtually useless for other than very coarse jogging. The shortest button press I can manage moves on the order of 0.100" or more. Sometimes much more. With the pendant on the PP, I could easily jog about 0.010". The length of the button presses on the pendant are obviously being affected by the latency of communications with the SS.
I would say that is not the issue, there must be an issue with the way the Brain is working with the plugin, the USB2 is much fatser than the PP as far as I know.
I'm using the latest SS plug-in, just downloaded last night.
Try downloading an earlier one and see if the problem still happens, as I said I seem to remember a problem with a simple Brain I was using but its sorted now. However I am not sure when it got sorted as I have been using Beta plugins for quite a while now doing threading testing so it could well have been the last one on the site that gave me problems. I forgot to bring a SS home with me tonight so cant test here
  Here's another "anomoly" I discovered in my first attempt to do some real work with SS today - I normally do 2.5D milling, with manual Z moves, as I haven't yet had time to mount the servo to the Z axis. Each time a Z move is required, an M00 is inserted, telling me where to move it to. Trouble is, the G-code display in Mach is no longer synchronized correctly, so the M00 lines are a few lines above or below the highlight in the G-code display. This sometimes leaves me having to flgure out which of the potential Z moves is the one I should be doing right now. Not good....
Cant say I have noticed that, will check it out tomorrow when I get to the workshop, well thats if I remember, supposed to do a lot of things today and forgot most of them LOL, must be getting old
Hood