Regarding the AR636, HH search doesn't say anything about SAFE so I will try to make what I have work here.
Based on what you said to do, the receiver appears to be the problem.
Here's some pictures: first pic: where the servo arms end up after startup w/ throttle application, second pic is rudder position. After picking up the airplane and introducing some intentional movement (yaw, pitch, & roll [wing installed] shows counter-actions taking place), third pic shows servo arm positions, and forth shows rudder position. Last picture shows rudder servo hooked up to servo driver using 6v with no evidence of drifting - tape marked with Sharpie to index servo arm center point. I can't reproduce the extreme deflection I've seen other times like I saw at the flying field. I'm slightly paranoid that the computer hackers are conspiring to make life miserable for us, using our PI to introduce bugs in the devices we use these days, watching our every move by satellites, devices, and smart phones and pointing and laughing at us...
What I do know is that the rudder servo always drifts to the left when hooked up to the rx. I don't think this situation threatens the life of the airplane - its just not ideal. If the drifting continues to only change a few degrees during a typical 10 minute flight, thinking about splitting the difference by setting up prior to flight favoring a little right rudder - as the flight progresses, it would improve as the rudder approached center and if it went to a few degrees left - no big deal.
However - thinking out loud here - if the system continues to migrate left, continuous repositioning of the output arm/adjustment of clevis is finite and becomes another problem.
On a better note, I had a chance to fly a club member's little Apprentice last Friday evening and was equally impressed at how great it flies - as well as it's bigger brother. Only caveat was that I couldn't trim out a tendency to turn right with application of left trim on his DX18 - both rudder and aileron. Not sure if stock setup but rudder was coupled to his aileron stick.
