I've seen it before and I ran across it again tonight.
When doing a manual capture, with Ctrl-C, It fills in the blank with a default file name and a drive to do a cpture on (I don't remember if it uses a default drive, or the last one a capture was done on).
The problem is that is you know the capture isn't going to fit on the drive listed, changing the drive letter to somewhere else has zero affect on the available time. For example, I knew that the F: location didn't have enough space, so I changed it to E:, as well as changing the file name. I didn't give it a length, as I was going to manually stop it. I told it to start and it filled in the available time based upon the F: drive, not the E: drive that I just entered. I could have been in a world of hurt, but because of what I was doing, I was able to stop the capture within the 37 minutes, during a commerial break, I go from there.
The software needs to look at the drive letter when the user starts the capture, not the drive letter it originally put in the GUI.