According to Robert11 :

Have been trying to get some smoke detectors interconnected, and
during the trouble shooting of the problem, I measured (with an analog
voltmeter) the
voltage between the white neutral, and the bare copper ground wire in the
box.

Was very surprised to see that it was about 2 V AC.

....

BTW: How "common" is it to see voltages of this magnitude between the white
neutral and ground ?

[The other postings touched upon this, but sufficiently buried to perhaps
be worth repeating.]

You measured with an analog voltmeter, that will _normally_ have an internal
resistance in the 5K-50K "ohms per volt" range. Your voltmeter _should_ have
that info prominently displayed on the face of the meter. If your meter
has that information (and it's not something extreme like 1M ohms per volt),
then the voltage difference between neutral and ground is real.

You have to be really careful interpreting results with a DVM (or other
devices in the 1M+ ohm/volt range) due to inductive pickup. While a DVM
reading would be real if the neutral and ground were properly connected
back to the panel, if there was a break in the wires, the DVM could read
virtually anything (in fact 120V in some cases).

[If you're using a DVM, you can try to "load" the voltage you're seeing.
A 50K ohm resistor would work. So would a low wattage 120V lightbulb.[+]
If you still read 2V, the voltage is real.]

A 2V difference between ground and neutral would be "normal" if there's
a highish load in operation downstream of the smoke detectors due to
voltage drop in the neutral (IxR), and there was a fair bit of wire
between the smoke detectors and the ground/neutral interconnect in the
panel.

But it does seem worth checking a little further. Try killing that
breaker. Do you still see 2V between neutral and ground? If you do,
you probably have excessive resistance in the neutral-ground
interconnect in the panel[*]. Check for voltage between
neutral and ground on other circuits.

[+] Values chosen so you don't fry something. It'd be best to
pick a resistor in the 10K range, say, but you may not have one
handy, and it really should be at least 1W - just in case something
goes wrong and you get 120V.... Best not to have "test equipment" that
will melt - especially if you tried this trick again to test whether
an 80V reading was real or not.
[*] A reference was made to a major load imbalance in the two hot
legs of your service possibly causing this. No, it wouldn't, because
no matter how hard they "pull", resulting in different neutral-hot voltages,
they shouldn't pull a solid neutral/ground connection in the panel
apart voltage-wise on an unloaded circuit at _all_.
--
Chris Lewis, Una confibula non set est
It's not just anyone who gets a Starship Cruiser class named after them.