Are there any good jumper wiring guidelines to reduce the chances of a breadboard (or set of breadboards) radiating EM outside various limits (FCC Part 15 in the U.S., etc.)?
No. Breadboard is a nightmare, RF-wise; jumper wires simply look like antennas for sufficiently high frequencies, and the connecting "bars" are long enough to become monopoles themselves...
e.g. maximum length of jumper wires,
Rule of thumb: a wire smaller than 1/10 of the wavelength is not going to be a great antenna. So keep things short!
maximum loop area with respect to the nearest ground return,
That would assume you can be sure any of the ground returns would have low impedance compared to a loop. That's usually not the case for breadboard!
minimum termination resistance of a circuit loop, etc.?
Obviously, you're absolutely right to think about this: when you know where the energy is sunk, there's less chance it can radiate! So:
- Terminate, if possible
- keep wires short relative to wavelength
- don't use maximum drive strength at high switching frequencies
- Add local decoupling capacitors – that might both improve performance and supply stability (and that has the potential of reducing production of harmonics to begin with!) as well as it will shorten out RF to supply/ground.
- adding e.g. a 10 nF capacitor can trade slew rate (which often is desirable) for EMI immunity and emission
- Especially with microcontrollers, it might make sense if an experienced electronics expert or enthusiast (cough that'd be you! :D ) simply designs (or uses an existing design, see e.g. "sharing" on oshpark.com) PCB that carries the microcontroller (hint: sometimes, especially when ordering really cheap PCBs, e.g. from JLCPCB, the difference in cost between the DIP and an SMD variant of an MCU can already amortize the board), decoupling capacitors and the oscillator; keeping the capacitors close reduces supply/ground loops. Having a PCB also makes the project more robust, gives a bit of space to add a label etc.
- Don't fret too much – typical school experiments will not have multiple amperes switched at 10s of MHz es; if you take care to not intentionally build antennas ("I added a 1,5m cable so I can measure this 100 MHz clock on that scope over there"), you'll probably be fine, even with the low-power modes