thin lens OPD sign?

[mperrin]

Issue by mperrin
Friday Apr 22, 2016 at 21:39 GMT
Originally opened as mperrin/poppy#164

---

While working on #129 I ran into something I've now manage to confuse myself about: The sign of the defocus term for the thin lens class.

Right now the convention is that a poppy.ThinLens with a positive number of waves of defocus has a negative OPD at its center, and a positive OPD at its edge. I.e. it is a diverging optic such as a biconcave lens or a convex mirror. See figure.

This seems backwards to me, but maybe I've just gotten turned around in my head about signs. Questions...

• Should we flip the sign behavior here? If so that's a significant change in a sense.
• Should we leave the behavior as-is, and just document it more clearly?
• Do we have to worry about sign convention consistency between ThinLens and fresnel.QuadraticLens?

Need to think about these more when it's not the end of Friday.

[mperrin]

Comment by mperrin
Monday Jun 06, 2016 at 20:17 GMT

---

Yeah I'm convinced this is the case based on double checking against simulated WFS images from ITM. Right now we have the signs the wrong way around in the weak lens images, and I believe that flipping the sign in poppy is the right way to fix it.