Δ√QUOTE="Andy Resnick, post: 5951029, member: 20368"]The uncertainty/statistical spread of a discharged battery's mass is likely many times larger than a pg[/QUOTE]
Looking at a harmonic oscillator, torsional:
Originally, I had thought that since the period is proportional to the square root of the mass, ... could be faulty thinking.
Background:
Time period of oscillation of twisting wire,
https://www.colorado.edu/physics/phys1140/phys1140_sp05/Experiments/M4Fall04.pdf
or frequency,
ω = √[ (
k/
I )
I - moment of inertia of the mass(s)
https://en.wikipedia.org/wiki/Cavendish_experiment
For the Cavendish balance,
This is the regular formula, for two masses separated horizontal distance 2L, and hanging from a wire.
T -period
m - mass
L - distance to wire
k - torsion coefficient of the wire
The mass hanging from the wire could be a single battery, or two on extended arms similar to the Cavendish setup.
We should have, with the battery uncharged
T
1 = C√( m )
where C is a constant composed of 2, π, √L
2/
2k
Similarly, with the battery charged,
T
2 = C√( m + Δm )
Squaring each equation, and subtracting
T
12 - T
22 = C
2 Δm
or
Δm = T
12 - T
22 / C
2
Like you said could be impractical due to errors in measurement for C.