After my success at testing the proto boards with my meters, I wrapped them in some masking tape and set up some lights for testing. I have the 24v supply coming from my bench supply and going go an RJ45 connector. The ESP pwm signal is going out on a couple of pins:

I have the modules connected off of a 100 foot ethernet cable to the first one and have the second one daisy chained off that one via a 25 foot cable. I was worried about voltage drop for the 3.3v pwm signal, but since the opamps take next to nothing as far as current draw on the inputs this has not been an issue. I don't see there being more than 125 feet of cable used anyway, even if multiple lights are daisy chained. I would test but I do not have any longer ethernet cables available. 

The modules worked great, but I noticed some pretty bad flickering around 15-20% brightness. I took one of the modules back to my bench and checked the output with an oscilloscope and found that the voltage signal was rippling along with the pwm signal. I thought the opamp was supposed to smooth out the pwm signal into a steady voltage but apparently I was wrong. 

I found that a 100uF capacitor across the voltage output pins was enough to even it out so I installed one on each of my test lights. When the final boards are made this will of course be on the board itself.

Everything is working great and I can control the brightness of the two panels with no noticeable difference in brightness even with long cables between them.  The next step will be redesigning the boards for the new arrangement. Since the opamps are not on the controller board now, that frees up enough space that I might be able to use one of those 4 button RF receivers like are used in some wireless light switches to be able to have the manual override buttons be wireless instead of hardwired.