I'm measuring the distance between the two cords of a bifilar pendulum and its currently the lowest significant figure data I have. I considered wrapping a string around the two cords and halving the length of that, but the string needs to be taught to get a good measurement which then moves the cords together.
I'm aware that you can get a vernier caliper to measure lengths less that 15 cm, but that is too short for my needs.
EDIT:
It seems to me that I'm coming across as a bit idealistic but here's why getting four significant figures:
From a comment:
You have the treads of finite width
I've used a type of nylon fishing wire that claims to be 0.4mm in diameter which isn't stretchy.
how parallel are the threads
I was planning to measure the distances between the threads at the top and the bottom of the pendulum and then adjust until parallel
how are the threads attached to the rod and the points of suspension, is there a knot at the end of the threads, etc
I've made a rig out of some nuts, a threaded rod and two washers that gives a distinct point where the filament sways from.