Very strange… Using DXD9 client I tested various MFD refresh rates (1x and 10x par sec) as well as several MFD sizes and never noticed that. Anyway, here is a patch that should solve the problem.
For the reentry process, I’ve chosen to use the Retrograde mode to accelerate the descent and have a more or less predictable descent rate, which is fine for an autopilot… If I use the “head down” mode, the capsule somehow skips on atmosphere, which is normal and perfect for high speed reentry (e.g. return from the moon) as it limits the number of Gs, but more difficult to control from a pure automation point of view. At least it was my opinion, but not so sure now… The result is a not so accurate reentry autopilot, and I would be happy to get some good advices for implementing a better one. The key questions for designing such a tool are:
How to estimate the braking burn point (distance to target or long/lat)?
How to control the descent rate for a more accurate landing point?
And, ideally, how to control the lateral course and maintain the direction to target?
I spent a good amount of hours testing various empirical solutions for all this stuff, but I reached my limits, as I’m not – at all – an aerospace engineer. So all suggestions are welcome…