Hi,
Oh, about the length of one electron
In any wire of any length that has current flow we might see some radiation. The radiation occurs when electrons get accelerated, and electrons get accelerated when the current changes. It doesnt matter if they are in a wire that is 1 inch long or 1 mile long or 1mm long.
You could look up some antenna gains though if you want to see what is considered 'better' for radiating by considering what is better for receiving There's a formula somewhere that will relate the length of the antenna to the wave length.
Also note that if you have electron acceleration in one direction in one wire if you have another wire close to it with acceleration in the opposite direction the radiation will partly cancel. The two vertical sections of the upside down square loop have such a configuration so we dont see as much radiation along the entire 3*1.6cm length.
If you want to know more details you should find a site that is dedicated to stuff like this. What i used was the Biot Savart Law and solved for a square loop that has one side missing, using a direct multi dimensional inegration method that makes no approximations. That means the result is exact, but unfortunately we get these other side issues coming into play such as the mutual inductance of the rest of the circuit and so we are always working with an approximation once we leave the drawing board, or in today's world, once we leave the symbolic math software environment
This makes SOME real world experimentation always necessary. I site dedicated to this stuff may have people who do this on an every day basis, and so can provide some ideas that apply once you have something up and running. But i dont think there will ever be a circuit that like this that does not need at least some adjustment. The usual way is to build a coil based on a textbook formula and make it have a slightly larger inductance than needed, then spread the turns slightly in order to adjust it once it is in the actual circuit.
