I have been working on a PhD for over 6 years, and in that time have published 5 (conference) papers, each of which I intend to convert to journal articles after my PhD. My supervisor has been insisting that I do some experiments which have turned out to be considerably more challenging than either of us expected. Our funding has run out, so I recently asked my supervisor if it would be okay to scrap the experiments and defend what I have done so far at the end of the year. Based on the volume and quality of my work, I think any objective person would feel that I have done more than enough for a PhD. (Indeed, I've had a professor who clearly has read my work ask me out of the blue at a conference why I haven't graduated yet, as he thinks I've done way more than is typical for a PhD.)
My supervisor's reaction was particularly bad: He accused me of reneging on a promise, suggested that I've been conning him for years, etc. I agree completely with him that the experiments would be nice to have, but I disagree that they are essential, and I particularly disagree that they are worth going into debt over. I think we could get a good paper detailing what we've done so far and what the challenges have been. Unfortunately that option was unacceptable to him too.
Any attempt to discuss this further with my supervisor goes nowhere. What are my options at this point? I told him that in the absence of funding I will be dropping out at the end of the semester. He said that is a waste, which I agree with, but ultimately I don't want to continue what I'm now viewing as abuse.
I appreciate everyone's responses. Taking the advice I received here and elsewhere, I decided to complete one additional set of experiments and then speak with my supervisor again.
These experiments had a new feature that we've wanted for a while that my supervisor hoped would solve some of the issues we've been having. The new feature ultimately it did not solve the problems we've been having, though it helped narrow down the cause of the problems.
I spoke with my supervisor about the experiments, emphasizing that we need to do many many more experiments to get satisfactory results, far more than we have time for. He agreed when I described the results of the experiment. We then had a discussion about what he expects me to do to graduate, and again, he made clear that he doesn't care about what I did previously and only wants the new experiments and some computer simulations of the experiments. I tried to justify what I did previously as necessary to do the new experiments correctly, but he wasn't convinced.
Some folks here and elsewhere have suggested that my situation would be different if I had journal publications. While my supervisor's issues didn't sound at all like "you need journal publications" to me, I asked if submitting my previous work to journals soon would make a difference. He thought for about a minute and the answer was yes, it would.
Consequently, we came to an agreement: I submit my previous work to journals before end of the semester and write up a chapter in my dissertation on the already completed experiments (likely will remain unpublished aside from the dissertation). I will switch to a part-time position at the end of the year, move to a cheaper location, get a job, and return late in the spring semester to defend. In that time I should at the very least have reviews for some of the journal articles I'll submit, which is sufficient for my supervisor. My supervisor and I have agreed on which journals to submit the articles to as well.
I may update this answer later after my defense.