This is basically a tap/branch circuit issue.
First let me describe the circuit design.
We are feeding 6 – 400w, 480vac light fixtures spaced over a 500 span. The light fixtures are installed outdoors recessed in metal panels attached between support beams of a bridge.
We have a 480 vac, 20 amp 2 pole breaker feeding two #10 awg conductors, one #12 bare copper ground conductor in a ¾” conduit. (#10 was used for voltage drop)
The ¾” conduit runs the entire length and has a type T conduit body installed in line with the conduit near each light fixture.
Attached to the conduit body is a section of ¾” flexible liquidtight metal conduit (sealtight) that is routed to the fixture’s electrical box.
Inside of the sealtight is one #12 bare copper ground conductor and two #12 conductors that are spliced at the conduit body to the #10 conductors and at the fixture’s electrical box to the #16 or #18 fixture ballast leads.
The argument between the electrical contractor and the code inspector deals with the length of the sealtight and the conductors that lie within.
The electrical contractor considers the ¾” flexible liquidtight metal conduit and the circuits within to be branch circuits. As such there is no limit to the length of the sealtight and conductors between the conduit body and the fixture (within reason).
The code inspector states the ¾” flexible liquidtight metal conduit is limited to a 6’ length based on section 351-8 Exception No. 3 of the (1999 NEC) Code. He also states that the conductors within the ¾” flexible liquidtight metal conduit are to be considered tap conductors and are subject to section 410-67 (c) and are therefore limited to a 6’ length.
My take on the issue is that section 351-8 Exception No. 3 applies to the mounting distance requirements for flexible liquidtight metal conduit, not the length. I also feel that section 410-67 (c) applies to the smaller conductors that connect to the ballast that are not protected by the 20-amp circuit breaker.