Assuming this "force" represents an actual "part" of the system, and is therefore a "primary (weight)" load, then there are a number of reasons to opt for the "rigid element" technique rather than the "Fn" technique:
1) By using a "rigid element" of the appropriate weight, this weight will always be included in the system analysis when the load case contains either "W" or "WNC". There is no chance that you'll forget to include "Fn" when you setup a new load case.
2) Should you ever need to perform dynamic analysis on the system (maybe not you, but someone else a few years from now), the mass of the "rigid element" would be automatically included in the system mass matrix. The "Fn" on the other hand would not be included, and to "correct" the dynamic model you would have to include a "lumped mass" in the dynamic input. This is because we don't know what the "Fn" vector means - is it a mass (i.e. part of the system), or some type of load?
Set things up to be "simple", "obvious", and in a manner that hopefully prevents you from "shooting yourself in the foot" sometime in the future. I'd use the "rigid element". Just make it 1" long with the appropriate weight, and attach it to the node where you would have applied the force.
_________________________
Regards,
Richard Ay - Consultant