Apologies for the delay responding while we investigated this issue.
It seems that the problem lies with the way the frame size is reported from MXF files, as 1920×1080, instead of the native DV size of 1280×1080 which is what FCP7 expects. If you send a single P2 clip from CatDV (as a merged clip) FCP seems to read the frame size information from the file and copes, whereas if you send a spanned clip it has to go over as a sequence and FCP seems to get confused by the difference between the frame size of the sequence and of the media.
We are still investigating how best to resolve this in CatDV but there seem to be two possible workarounds:
a) for an existing sequence in FCP which appears incorrectly zoomed in, bring up FCP Sequence Settings dialog and enter a custom frame size of 1920×1080.
b) when exporting the merged clip from CatDV, rather than dragging it over using the “magic door”, save it as a FCP7 XML file then edit that file with a text. On line 60 or thereabouts change 1920 so it says 1280 instead, then import the XML file into FCP.