We use the term "ground" a lot, as a noun, as a verb, as a concept, and an answer to many design questions. And we should, since ground and grounding are important factors of almost every design.

But lately, I have stopped using that term where it does not apply—and that' a lot of places. In fact, you might say that I have become "ungrounded." For battery-powered equipment, as well as automotive systems and airborne or space systems, there is a "common" but there is no ground, unless we are using the term as another word for common.

But w shouldn't. A real ground refers to a circuit point which is Earth-connected, and thus can take advantage of, or needs, the huge reservoir of electrons that the Earth offers. This ground not only acts as a signal common (not very good one, admittedly), it also provides a relief path for ESD, overvoltage, and system faults. An ungrounded appliance, for example, can be lethal if its metal enclose accidently connects to the AC mains; a grounded one will pop the local circuit breaker.

Inherently, battery-powered devices are ungrounded. Sometimes this is a virtue, as when you need to make floating measurements across individual cells in a series string of batteries, and the top cell may be at a potential of hundreds of volts above Earth ground. [An aside here: remember that "voltage" is a shorthand tem for "potential difference" and you should always keep in mind the unspoken "with respect to what" part when you say "the voltage is 'x' at this point."]

The interesting thing is that there are times when you want your system grounded and there are times when you deliberately don't. In the case of the battery-cell string, you want to make sure the instrument is ungrounded (floating) so that it can be protected against high potential in the string and also make a differential measurement across a given cell. But in a reverse example, many AC-powered medical instruments carefully isolate the patient's I/O so that any failures in the instrument itself do not ripple to the patient.