How would you make it so that it can be loaded by the GPS like a real one? Is there a special programming involved in this as in a language, or could you yourself construct in the GPS by entering coordinates etc?
As a practical answer: you could of course do it, by
- reverse engineering the copy protection scheme on the flash cartridge
- reverse engineering the database encoding
- inserting your own procedure into the database
- re-encoding the database
- programming a flash cartridge with the new database
and repeat this every 28 days, because the Jepp database changes will wipe out your edits.
I honestly have no idea if anybody has done this, and very much doubt anybody would bother.
It is widely believed a number of people have cracked Step 1, because that allows database subscriptions to be shared among GPSs. The copy protection scheme varies between Honeywell and Garmin but cannot be that hard, and I know some MFDs are trivial and are cracked simply by purchasing an unusual flash card writer from a German company.
If you create a set of user waypoints, perhaps also using some existing database waypoints, and load them into a flight plan, you get the same effect anyway - except you don't get the automatic lateral sensitivity increase at/past the FAF (you have to do it manually).
Commercially, nobody is going to touch this, for liability reasons. IAP design is a well documented procedure (TERPS in the USA, something else over here) and anybody can do it, but if you want to go low down you need a physical obstacle survey done. If you are happy with an MDH of say 600-800ft, you can just do it off a 1:25k contour map, and any remaining uncertainty is cleared by a test flight in VMC.
Very few people have been killed by flying DIY approaches. Those that have probably did not do the procedure design properly. I know one bloke who did a CFIT on a DIY IAP and amazingly survived, but he forgot to design the missed approach segment, and ...... crashed when he went missed
