Ideally, discussion of graduation should occur at the first committee meeting (and that will be your prelim if the topic of prelim is your thesis topic), if not earlier. There is a saying in clinical medicine, "discharge planning begins at the time of admission," and it's equally applicable to the PhD candidate.
A good way, I think, to begin this discussion is to be proactive and organized. The way I did it was at the first thesis committee meeting I handed out a sheet to everyone on the committee, outlining the main goals and specific aims in bullets, then a tentative timeline to graduation (appx 2-3 years). In subsequent meetings the same timeline, with some adjustments, was handed out again and again to reinforce this into every faculty's head. Though they had many objections scientifically, no revision of the timeline per se was ever attempted--and I suspect that people on the committee figured that the timeline was not written by me but by my adviser, while my adviser thought it was outlined by someone other official in the MDPhD program. The more formal the better, IMHO.
Committees are highly variable. My committee was very amenable for me to set my own pace, which resulted in a "quick" PhD, but the paper was stuck in limbo for a LONG time, which I think affected my residency application in a negative way. For a variety of reasons, however, I did not want to stay for another year, and in the end everything turned out fine. Sometimes committees can be very unfriendly. You have to be firm but nimble. Know your own bottom line. Remember everything's negotiable...also remember there is no reason for faculty members to hold you back, especially if you are going into a field for primarily clinical reasons, and your PhD is unrelated, and there is little chance that your paths will intersect again. Often the harder your push the more antagonistic people become. You have to spin it such that they feel like they are on your side, even though you are simply manipulating them to do what you want them to do. Often used techniques include evoking competitors, pitting one project against another (i.e. you want me to do this impossible experiment, and I don't want to do it, but perhaps instead of killing myself I can help this other postdoc to do these other experiments, and he can get a paper in. Wouldn't that be a win-win? Of course I'll only do this if you let me out.), writing a manuscript (this one often works, because even if you don't have everything, if the paper is written it FEELS like you've done a lot), presenting your work as much as possible at conferences, etc.
The idea is to make it SEEM as if you've done a lot, even when you haven't or couldn't. You want to start this campaign as early and as comprehensively as possible. When you hit a snag make sure to outline all the possible reasons and ways in which these reasons can be addressed. You don't actually have to do any of it.
This is not to say though that you shouldn't love your science. I think counting an adviser out for the fact that they usually don't graduate in 3.5 yrs can be silly, especially if the topic that you are interested in can't be done in a fast way. If you think it's a breakthrough but will take 5 yrs, you should probably just push through and do it. But given that you like your project and things are going in generally a good way, spin becomes absolutely essential. Plus it'll only become more and more essential if you want to continue this path as you sell your future projects to a plethora of study sections, journal editors, foundation directors, and most importantly your employees and colleagues in the community. A lot of great scientists are buried in obscurity because of the lack of these skills. Might as well start. And these skills are equally applicable in clinical medicine.