Your LST is near midnight (very convenient :-), so your hour angle is around 4 hours east of the Meridian. At 15 degrees an hour, that corresponds to 60 degrees.
So, I see nothing wrong there to keep your mount's horizon limit from kicking in (since the 60 degree slew moves the RA axis west). Especially when your latitude is also so large that you can rotate the RA axis by a full 360 degrees and you will still never reach horizon! (My latitude is +45.58 degrees, by the way; not quite as far north as you).
Unfortunately, the Avalon protocol is one that I never included in my simulator (I have included protocols for the LX-200, some iOptron, SynScan, some Celestrons, EQMOD and my own RainbowAstro only.
(By the way, the M-zero and the RainbowAstro RST-150h were the two final candidates I had in mind when I was upgrading from my Tak EM-11 mount -- because I am getting too old, I wanted something lighter (all Tak mounts are way heavier than their counterparts), and also a mount that can handle a bit more payload. While I was weighing the options, RainbowAstro announced the RST-135, and I ordered it before it was released :-). I also hounded the poor guys at ZWO in the early days to get the RST-135 implemented and debugged (thank you, Messrs. Jiang, Zhou and Ji).
Doesn't the StarGo also have an LX-200 protocol simulator, Jarno? If so, have you tried that to see if the use of the LX-200 simulator is enough to fool ASIAIR into slewing? (That being said, I see postings on the ASIAIR Facebook page of the Meade mounts failing too.)
Another suggestion is to actually start with a much smaller DEC, like +75º (closer to zenith). Even if it works, this is not a great solution, since the polar alignment accuracy is not as good as being within a few degress of the pole. But it might help diagnose what the heck is happening. You might also try to start at different hour angles. I always start with my dovetail positioned about 60 degrees east of top dead center-- this way, after the 60 degrees slew, the imaging train is completly clear of my altitide and azimuth bolts. I don't remember once where it refuses to make the 60 degree slew.
Oh, try one more thing too, Jarno:
The RA and DEC are only where the mount thinks where the OTA is pointed at. We know that the ASIAIR takes one plate solution before it moves the 60 degrees. Next time you are out, see if you can do an image capture and a plate solve in the Preview window to see where it thinks the OTA is pointed at that starting location. This might offer us a clue and a better chance for a work around.
Lots of people complain about their mounts (not just your Avalon) not moving during Polar Alignment, or even worse, moving erratically (and in reverse) and smashing the imaging train into the pier. For the lights of me, I can't come up with a scenerio why that can happen.
I was earlier hoping the simulator can log the command(s) that the ASIAIR sends to the mount to start the 60 degree slew. If the ASIAIR sends a bogus command (for example, a DEC value greater than 90.0000 degrees or RA that is over 23h 59.99m) the mount could well refuse to move (as it should). Unfortunately, I did not implement the StarGo protocol. The math around the pole when using the RA and DEC coordinate system is very sensitive (a singularity at the pole itself). It is very easy to blindly compute a destination that is over 90 degrees of DEC or 24 hours of RA.
Chen