On March 26, 2002, about 1715 eastern standard time, a Robinson R-22, N7187J, registered to Pompano Helicopters, Inc., operating as a Title 14 CFR Part 91 instructional flight, crashed in the vicinity of Coral Springs, Florida. Visual meteorological conditions prevailed and no flight plan was filed. The rotorcraft received substantial damage, and the flight instructor and private-rated student were not injured. The rotorcraft departed Pompano Beach Airpark about 1 hour before the accident. According to the flight instructor, he wanted to end the training flight by having the student pilot perform a forced landing maneuver with a power recovery at 500 feet agl. He stated he had to say, "engine failure" three times before the student pilot entered an autorotation. He does not remember if he turned the governor off or not. He remained lightly on the controls with the student, and at about 400 to 500 feet agl, he, (the instructor) tried to ease in throttle, but could not get throttle response, and interpreted the condition as a real engine failure. He told the student, "my controls" and assumed rotorcraft control. In the flare, he realized the height of the grass was about 6 feet high instead of the usual 2 to 3 inches, and the touchdown area was very soft and uneven. Additionally, he realized the student was still on the collective controls, and it was being raised prematurely. The result was a hard touchdown, a bounce, a slight forward slide, and the rotorcraft rolled over onto its nose. The two occupants exited the rotorcraft under their own power. According to both the owner/operator and the chief pilot of the flight school after inspection of the rotorcraft and debriefing the instructor, the training flight was the first occasion that the instructor and the commercial-rating trainee were paired up. The proper techniques and procedures for a simulated engine failure were not briefed prior to commencement of the training flight. The engine governor was not turned off prior to announcing, "engine failure" to the trainee. The trainee rolled the throttle into the detent to keep the rpm needles split, and held it there without the instructor detecting that he had done so. Had the instructor detected that the trainee had positioned the throttle in the detent and was holding it there, he would have realized what was causing the lack of throttle response, and that there was no engine failure. According to an FAA inspector, because of the throttle resistance to movement, both instructor and student were convinced they had an engine failure. The instructor took control, performed a real autorotation, but failed to maintain main rotor rpm and selected a tall grass area for landing. Consequently, the rotorcraft made ground contact and rolled over causing damage to the main rotor, tail boom, and cockpit glass.