I used to work at a joint test facility testing military communications systems at Fort Huachuca. When I first got there back in 1999 they had this really cool static display:
This was an experimental CSP run by the NREL. It originally had a stirling engine in it, but they had so many problems with it that the replaced it with something else, might have been a stirling engine from another manufacturer, the guy that did the maintenance on it wasn't sure. Anyway, that engine didn't work any better.
I call it a static display because the first two years I worked there is was broken and they had been waiting for parts to repair it since before I got there. They finally got the parts in and repaired it, it ran for about three weeks and then broke again. In the five years I worked there, I don't think it was up and running for more than 3 or 4 months total.
Sometime around 2005 they finally gave up on it and removed it, I imagine it was latter sold for scrap.
Stirling engines can work fairly reliably if they are CONSTANTLY running, however they don't respond well to operating intermittently, like running during the day and being off at night. Part of the problem is that in order to work they require very tight tolerances. Since they tend to be made out of a variety of dissimilar materials that expand and contract at different rates when hot vs cold, keeping tight tolerances under both conditions is challenging.
Another part of the problem is that to get really high efficiency they tend to use helium, or some other gas for a working fluid. With the movement for tracking and the high temperature differentials they would frequently end up with leaks which would shut it down again.
There are probably other problems as well, those were the two main ones that stuck out in my mind.