Why does collimated light form an airy disk pattern on a camera sensor?

I’m using a Raspberry Pi camera v2.1 (lens removed) to make some observations on a laser. I have the laser coupled into a single-mode fiber which outputs a smooth gaussian spread onto the camera. From the end of the fiber, the camera has an angular size of about .02 radians. When the laser is turned on, the camera sees an airy disk interference pattern, much like what you might see in a Fabry-Perot interferometer, though it’s not perfectly centered. I’ve been busting my brain trying to figure out what’s causing that pattern and I keep coming up blank. Here’s what I’ve considered so far:

At first I thought it might be thin-film interference in the glass covering the photoarray, but the fringes are too small for any reasonable thickness of glass. I did the calculations, and it would need to be like 8cm thick.

Then I thought maybe it was just single-slit diffraction from the fiber, but again, the fringes are too small for the aperture and distance.

Finally, realized that, if it was interference in the glass, it’s too thick for thin-film, so I thought it might be a sort of thick-film interference, where the incident ray at each angle will have an infinite series of rays at smaller angles, which each reflect a certain number of times before all ending at the same spot. I did those calculations and still came up short.

The effect seems very similar to the airy disk interference in a Fabry-Perot etalon, and it seems like that type of effect should happen, since the glass has two parallel surfaces, but I can’t figure it out for the life of me.

UPDATE: I’ve looked at the beam without the camera, and I’ve looked at the camera with the laser off, and there are no rings in either case. Also, the pattern phase shifts slightly when I tune the laser to slightly different frequencies, and significantly if the camera is moved at all.

I’m not at the lab now, but I can get a picture to post tomorrow.