Does sound like a laptop issue, I have read many issues with laptops and CNC. I would get a cheap XP pc off eBay myself and dedicate it to cnc.

First of all, thank you to Everyone  who shared ideas and gave input. Here some feed-back.

- The Power management of a laptop may indeed play a role, but is not ( ), in this case, the main reason for erratic USB connection losses.
- The only work around ( not a sustainable solution ) that was identified is to run the rig with the laptop ON BATTERY only.
- For those who are interested, I have attached a scheme of the set-up.

This set-up deserves a few comments:

1) The 5V power supply for the UC300 and BoB can be a cheap "plug in the wall" device or a "lab grade low ripple" unit, the behavior is similar: erratic USB disconnection are observed unless, the laptop is disconnected from its 18 V PS.

2) Between TP1 and TP2 (see scheme), using the "plug in the wall" device, there is 68 V ac measured. The same amplitude is measured between the shield of the "plug in the wall" and any of the "0" and "5V" wires, when everything is connected. Main contribution is AC 50 Hz, shape is distorted from a pure sine, verified with a scope. If the 5 V PS is checked alone, this AC is "only" about 10V.

3) Between TP1 and TP2 (see scheme), the nasty 68V ac drops to 3 V ac (shape identical) when the "cheap" 5V PS is replaced by the "clean" 5V dc.

4) By design, the BoB has its PCB plane ground on the LPT connector.

5) The shield of the USB B connector on the UC300 unit IS NOT connected to the PCB plane by design. This link takes place in the laptop, when the shield leaving the computer is bonded to the "0V" of the USB connetor.