For the connector, you should only need 6 pins, yes? Why not hard wire a 6 conductor phone cable out the back of the display, with an RJ-11 phone connector on the end. Then if I want to make a small hole, I can cut the connector, feed the cable through, and crimp on another RJ-11 once through the panel.

Now that I have canceled the light sensor option I have gained some room for a RJ-25 plug on the PCB... I planned to use this kind of cable anyway... But it is 2.5x7 mm, so the hole through the panel will be only marginally smaller...

I understand about the port mounting now, thanks. One further question; is the sensor at the port, or do you need to run tubing back to the PCB? Much easier to run a wire to the PCB than a tube...

The sensor I2C data isn't supposed to travel more than inches away... And it is a differential gauge... So you should run a static tube to the sensor+ wiring to the panel... More complex than pulling another tube along the existing pitot and/or static... And the panel is probably less than 3-4 feets away... I'm using relatively flexible PU tubing for pneumatic applications.... It's stiff enough to be pushed everywhere, will not be pinched by chance, will not wear out and can still be bent much easier than nylon... This will be a part of the kit, I think.

And having about the same volume of air between the 2 ports and the sensors might induce a comparable damping of both sensors... I don't know if it has a real influence, but it can't hurt... Electronic damping has been introduced in the software as I manage to get the AOA data about 600 times/second.... A 1/10 second average value looks still instant and filters most of the unsignificant turbulence...

By monitoring the rate of change of this dampened value, I can predict if a stall is likely to be reached within 2 seconds... And display the corresponding warning before the critical AoA is reached, giving you some time to react before loosing control...