Carrying on from the above today....
A nifty way of check this is with the X5S facing the "USAF Resolving Power Chart" available from Edmund Scientific for $40 here:
36" L x 24" W, Resolving Power Chart | Edmund Optics
If you center it up and perpendicular about 9 feet away from the X5S with the Olympus 45mm lens set to f/1.8 (wide open), turn the Manual Focus on the GO 4 screen with a stylus, or better using a Focus Wheel to examine the focus. Have the Focus Peaking in GO 4 camera menu set to High where the red lines appear around the edges of contrasty objects for a sharp focus.
If you scroll the focus wheel, you can see the image on one side become red and then it rolls the red across the other target squares if the sensor alignment is off as you turn the wheel. Turning the wheel back and forth and you can see how the red focus peaking rolls across the target if it's bad. It shouldn't do that at all.
If it all pops into red focus peaking at once with all four corner targets red at the same time, then congratulations as it's aligned correctly as in the attached screenshot. Prior to me shimming my X5S bayonet, only one side would go red while the other side remained untouched until I turned the focus wheel and soon that side became red while the first faded out. The attached screenshot of the iPad Air 2 shows all four corners red now at once so I'm good to go.
For the tech inclined, the tolerance of the bayonet-to-sensor distance is +/- 0.0012" if one looks at a Canon service manual. Pretty tight, and likely much tighter than DJI's tolerance level.
Given one side was less sharp on mine, I went with the DJI supposed tolerance (Ack!) of 0.005" and put that thickness of brass sheet under the two rightmost bayonet screws facing the sensor. You can see the two shims I cut out with scissors at the 1 and 5 o'clock positions (One is under a sensor protection pad.) in the photo where the four screws for the bayonet go. My final shims were 0.005" thick with a #44 drill (0.0860") screw hole. At the right tip of the screwdriver is a bent piece of brass shim that I put under the lens at various positions to determine the evenness of the focus plane. A cut Pec-Pad (Lens tissue.) is covering the sensor just in case. I used a 0.002" shim first, but the 0.005" did a much better job (It's tight in the testing portion placed underneath the lens against the 0.002" too!) and it showed me how much DJI erred on their assembly. Much better now.
Granted, I could have done it far better if I had my old Pearl Optical Autocollimator from decades ago when I repaired this junk for the majors, but trial and error works too - albeit much slower.
Fwiw, I recommend one getting a screwdriver made for these things rather than buggering up the soft heads. Most camera screws are JIS (Japanese Industrial Standard) screws and not the US Phillips design. Using a USA Phillips will result in a cam-out of the screw head (i.e. Reaming out the head.). The NIWA's PD-30 was the JIS industry standard for decades and the following site sells their copy and the blades too. Not cheap, but good, and they last!
Info on cross-point screw heads:
Cross-head screws - a caution: - PentaxForums.com
PD-30 copy of JIS driver:
Handle:
GRES_PD30 PRO Screwdriver Handle | SPT / C&C
Blade set:
Screwdriver Blades: | SPT / C&C
If you do this, might also want to apply some moly grease paste (Honda Moly #60 Paste.) to the bayonet and the flat spring behind it. I used a dab to hold the shims in place prior to put the bayonet and locking ring back on too. The moly paste helps immensely with easing the task of taking the lenses on and off and causing less metal-to-metal shards to come off the bayonets being dry. Just don't go overboard with the stuff, and wipe off any excess.
