I believe I'm with Don on this one. I would say a literal interpretation would require securement within 3' of the transition between raceways.
It would appear that the NEC makes a clear distinction between "conduit" and "tubing". I'm just not sure why. For the installation in the OP there are two different types of raceway and each could be considered as "terminating" at the joining of the two.
344.30 Securing and Supporting.
(A) Securely Fastened. RMC shall be securely fastened within 900 mm (3 ft) of each outlet box, junction box, device box, cabinet, conduit body, or other conduit termination.
358.30 Securing and Supporting.
(A) Securely Fastened. EMT shall be securely fastened in place at least every 3 m (10 ft). In addition, each EMT run between termination points shall be securely fastened within 900 mm (3 ft) of each outlet box, junction box, device box, cabinet, conduit body, or other tubing termination.
FWIW, I wouldn't have a problem if there was not securement within 3 feet in the installation that Gus described in the OP. I just can't put my finger on the safety issue.
Pete
It would appear that the NEC makes a clear distinction between "conduit" and "tubing". I'm just not sure why. For the installation in the OP there are two different types of raceway and each could be considered as "terminating" at the joining of the two.
344.30 Securing and Supporting.
(A) Securely Fastened. RMC shall be securely fastened within 900 mm (3 ft) of each outlet box, junction box, device box, cabinet, conduit body, or other conduit termination.
358.30 Securing and Supporting.
(A) Securely Fastened. EMT shall be securely fastened in place at least every 3 m (10 ft). In addition, each EMT run between termination points shall be securely fastened within 900 mm (3 ft) of each outlet box, junction box, device box, cabinet, conduit body, or other tubing termination.
FWIW, I wouldn't have a problem if there was not securement within 3 feet in the installation that Gus described in the OP. I just can't put my finger on the safety issue.
Pete