Interlacing progressive footage...why?

Please help me with my workflow. I am starting with 960x720p DVCPROHD and I want to end up with letterboxed 720x480p MPEG2 that is re4asonably sharp.

As an intermediuate step, I create a sequence in FCP7 which is NTSC DV but with the field order as "none".

I export this and import into Compressor. I use MPEG2 program stream present set for SD DVD. I also export 16 bit aiff audio. I then open both in MPEG Streamclip (renaming the clips so they have the same name, just different extentions for video and audio). From MPEG Sytreamclip I export a (muxed) mpeg with mp2 audio.

Probkem is, unless a add a de-interlace filter in FCP, I get comb and poor resolution. I'd rather not have to do this render every time. Is there a better way?

BTW, the reason for exporting the muxed mpeg is two fold. I need to author a simple no menu disk in Toast for public access t.v. where playback is via a controlled DVD deck for some facilities, and I need a mused mpeg2 with mp2 audio for directly adding a file (without having to ingest) to a Nexux digital video server .