At last, I have something to share with you all. Some details as to how I mount my autopilot boards to my helis. Mounting these boards on helis has been a touchy subject, but so far I have had pretty decent luck. I am yet to experience 'the leans' with this method - the unpredictable variation of AHRS estimate caused by vibration. This also gives me a chance to show off my hand modeling skills ;)
Pictured below is my most recent build - the all new KDS Chase 360. Its a brilliant design by Aussie designer Glen Kimpton. The design itself incorporates a dual-stage belt drive train, which is unusual for a 450 class of heli. This means that not only can it handle massive power (my motor a KDE 459XF rated at > 725W continuous), but its also extremely smooth. The heli also includes an innovative battery tray, which 'clips' in rather than being mounted with Velcro.
The autopilot, a PX4FMU + PX4IO combo is both lightweight and powerful. It fits between the side frames with 1.5mm to spare either side. This one was a clone, which sadly dies with only 4 minutes of time in the air. Next board will be genuine for sure.
The mount was designed to bolt to the heli frame and then hold the board with foam two layers of tape from above. The board is then secured to the mount with a rubber band (not pictured) for added security in case the tape loses adhesive strength at the wrong time. I 3D printed the mount at 40% infill put of black PLA. I find that PLA is great for these mounts because it isn't as rigid as other plastics so doesn't seem to resonate as badly. I also like PLA because its extremely cheap and biodegradable - I can make hundreds of revisions and never feel guilty :)
The effect of vibration on the board is better than expected, logs showing approx < +/- 1g for xyz accelerometers in the air @3000rpm. Will upload a log to back this up.