Thanks for the tips all, it's clear I don't know what I'm doing so I appreciate the help, haha. I'll do my best to answer most of these questions and see what I can make of this.
I want to test this mirror because I'm considering using it for a compact travel dob -- it would be nice to see whether the mirror is up to the task first. It seems to be a DIY mirror that was a leftover from a friend who bought it on ebay from an anonymous seller a long while ago, so I don't have strong reasons to trust that it will work well before I go and build a scope around it.
I've since taken the test rig down, but I can describe it. In this setup the knife edge is placed horizontally from below. The light source was placed at 2x the focal length of the telescope such that the reflected point converged at the same distance as the source. The light source and camera were kept stationary, and I varied the position of the knife edge. I used a canon T7i for imaging, and I found I had to take the lens off in order to get an image -- which means I wasn't able to focus on the mirror, since I had no lens to focus with. I just moved the camera sensor back until the image filled most of the frame. I tried 35 mm and 100 mm lenses but those didn't seem to work for whatever reason -- it seemed like the where the point source re-converged was 'inside' the camera lens, and so I couldn't place the knife edge on it.
My images also had trouble with glare / a double image -- especially visible in the first image -- which may both be a result of improper focus, but I wasn't sure how to resolve that. The contrast in my images is obviously much more poor than that of the one that Dale provided.
I think I'll need a more robust testing setup. Mine is simply on my floor, which is the sturdiest surface I have that is sufficiently long, but the floor is quite shaky. I also have no good method to precisely position the knife edge; I was moving it by hand.
So my understanding is that I'll need to be able to take cleaner images, do so for several knife edge locations, and then use the data reduction software to process results quantitatively. I think that will be doable, though my images are probably far from sufficient quality as they are, and it's not too clear why.