Would spinning salt emit radiation? It wouldn't need to be salt. 
Basically I was initially thinking about a mechanical transmitter, essentially just taking two equal opposite charges and fixing them to the opposite ends of a pole. Then you spin the pole around it's center (like a baton twirler) and it will emit some (mostly dipole) radiation.
Take an identical setup and fix the centers of the two poles together so we have a cross and spin that in the plane of the cross (picture a tire iron), and we should have a decent quadrapole moment and thus some quadrapole radiation (supposing you could spin it fast enough).
This made me consider iterations of this until you've a ring of alternating charges side by side spinning in a circle. Naturally I thought of table salt (some ionic crystal or another). So classically I get that it should radiate, but since it's a bound quantum state I'm not sure? 
Also, the strongest multipole moment I imagine would be some ridiculously large number for such a setup (probably making the radiation exceedingly weak and hard to detect?).
Sorry no pictures or equations, I'm tired
Anyway, for those interested I was initially considering how the moment of inertia of our initial pole would end up being dependent upon rotation speed.
 A: In order to radiate, a system must have an oscillating dipole (or higher multipole) moment.  Salt does not consist of discrete molecules, but rather of positive and negative ions arranged so that the moments cancel out very accurately in macroscopic crystals.  Some other substances (e.g., quartz) do consist of oriented dipoles, and you might think that crystals would present bound surface charges and dipole moments analogous to permanent magnetic moments.  (You could call them electrets.)  However, the surface charges attract free charges that neutralize them, at least when in equilibrium.  Such substances turn out to be piezoelectric instead.  
A: First, note that you are using a classical approximation for both the composition of a salt and the nature of electromagnetic radiation. This approximation breaks down both at very small length scales (where the distribution of the electron clouds within the salt becomes important) and at very low intensities of radiation (where the discrete nature of electromagnetic radiation becomes important).
Also, note that even in the classical approximation, you cannot consider the radiation of an individual charge in isolation from its environment. In the radiation zone (namely, at distances much larger than both the spacing between charges and the wavelength derived from the frequency of oscillations), the contribution of every charge in the sample is important (since they're all at nearly the same distance from a point in the radiation zone), and the total radiation is the sum of the oscillations in the field from all of the charges.
If you pick a charge in the middle of a salt, you can find an opposite neighboring charge, which means you have an electric dipole (and, in particular, the monopole moment for this distribution is zero). For this dipole, you can find a neighboring dipole pointing in the opposite direction, which means you have an electric quadrupole (and, in particular, the monopole and dipole moments for this distribution are zero). For this quadrupole, you can find a neighboring oppositely-oriented quadrupole, which means you have an electric octopole (and in particular, the monopole, dipole, and quadrupole moments for this distribution are zero). You can continue this process until you reach the end of the salt crystal. For a salt crystal consisting of $N$ ions, you may have a nonzero electric $2^{\lfloor \log_2 N\rfloor}$-pole moment, while all lower moments are zero.
It turns out that in the radiation zone, the radiation from an electric $2^\ell$-pole is suppressed by a factor of $1/(1+2\ell)!!$ relative to an electric monopole (source: https://en.wikipedia.org/wiki/Multipole_radiation). For a macroscopic crystal with $N=10^{24}$ ions, this corresponds to $\ell=79$, which means that the salt may have a nonzero $2^{79}$-pole moment, whose radiation is suppressed by a factor of $3/159!!\approx10^{-141}$ relative to your dipole. This means the radiation is certainly undetectable and would break the classical approximation even if it was detectable.
