Yes Lauri, it is exactly what I discovered also ... inverted installation was not so perfect and outboard also ... 45 degrees installation canceled both effects and it worked perfectly ... I would sa it was my best powertrain (I had two such models) ... I do not know, why, but I was not so happy with larger motors :-))
This has been a problem with schneurle engines in particular since pretty much day one. In fact it was the bane of our existence when trying to run schneurle motors without a pipe. You still hear it happening at times - listen from the intersection of the square 8 through the outside loop, 95% of the time you will hear it speed up.
That's why I run a RO-Jett vice the alternatives. It does this far less than any other engine I have tried - nearly undetectable even when you are listening/feeling for it to happen. I don't want to go into the details I have been told or discovered myself over the years as they could arguably be considered proprietary, but I will note that the cast-case version doesn't work the same as the bar-stock version as far as this effect is concerned.
Just about everything Dave/Ted/I have done over the years with engines has been to get rid of this effect. It's why we have gone "off the farm" on engine setups.
Near as I can tell, there are two things going on, and you have to do something about both of them to get it right - the engine scavenges different based on the direction of the G loading, and the flow of fuel/air mixture in the venturi changes as the direction of the air changes from maneuvering.
Another interesting observation os that this effect diminishes as the air gets thinner. What works just dandy in a Midwest summer day can be utterly unworkable here at sea level, to the point that many people have not been willing to admit that this effect even exists. Just like the Fox burp - and I think it is closely related.
Ted discovered the tilting effect back when he was trying various schneurle engines on the Imitation back in the late 70'/early 80's. He had round RC radial mount that was drilled to permit rotating it to 16 different positions (22.5 degrees). In most cases with inverted mounts, it runs much faster in outsides than insides no matter how you shim the tank. I think Ted told me that he found that his 40FSR did as expected in an inverted mount, ran faster on insides than outsides with the cylinder at 45 degrees right roll, and ran symmetrically at 22.5 degrees right roll.
The problem with the tilting engine solution is that it isn't the same from engine to engine. One engine might want 22.5 degrees, another might want 45, another might want 90 degrees.
Brett