Too high meaning greater than the recommended 90 degree shoulder abduction at impact discussed by Ellenbecker?
I think one problem with implementing Ellenbecker's tip is that it is not all that physically easy to get into such a steep shoulder tilt position at impact.
Especially for rec players.
We may be taking joint angles defined for other purposes, such as measuring range of motion, into complex athletic motions that they were not originally intended for. Shoulder 'abduction' as defined is a problem as my earlier reply discussed. I received a reply from an expert, which is posted in reply #34 .
This site has an on-topic discussion of the shoulder.
http://bestperformancegroup.com/?page_id=921
Shoulder girdle and humerus (upper arm bone).
The shoulder girdle can be tilted in relation to the spine. That will change the angle between the arm and side of the body or spine/thorax = will change the abduction angle. But when the upper body tilts, the arm may still have the same exact angle relative to the scapula. See discussion in reply #34.
Raul_SJ, you want to get a word of text such as "90°" that simplifies a very complex issue. It does not. I think your question points out an inconsistency in this issue regarding the definition and applicability of shoulder 'abduction'.
Ellenbecker has put our attention on this issue with a lot of detail. I believe that he also explains tilting the shoulders adequately around minute 9. In a 20 minute video, he can't give an anatomy lesson. We can find pictures and videos of high level serves to further understand this issue in detail, I think that viewing pictures is necessary.
My opinion - If you look at the shoulder girdle and draw a line between the balls on each humerus, you have a reference line. That line can be tilted by trunk motions and maybe other motions. As that line tilts, the tilt alone directly affects the 'abduction angle' - as defined and used - the angle between the arm and the side of the body.
In usage, this is probably how shoulder abduction is measured.
Most important for injury issues is probably the positions of the humerus and acromion, details unknown. The acromion and shoulder joint itself are part of the scapula, and the humerus is in contact with the shoulder joint. The scapula is also connected to the clavicle by a joint so that the scapula can move on its own relative to the line of the shoulder girdle (see illustration reply #34). It can be left to future generations to figure this all out.........and exactly what causes the injury..... Fortunately, we have now important guidance in Ellenbecker's video - probably enough information to avoid many shoulder injuries, otherwise, we would have nothing.
My view - Since I can't see the scapula very accurately, I look at the line between the shoulders and extend it to form an angle with the upper arm bone. I would estimate at impact for a high level serve that the upper arm bone forms an average angle of 10-20° above shoulder line extended. That is a lose estimate, but, say, arm up 40° does not look right to me. I don't think in terms of abduction angle on the serve because of the shoulder tilt. If the shoulders were not tilted as in the service impact and the arm-shoulder line angle were 10-20° up, the 'abduction' angle would be 100-110° - more than the 90°. You should view the Ellenbecker video and high level serve videos and make your own estimates. Camera viewing angle is important for estimating angles.
We can see what many high level servers, using internal shoulder rotation, are doing in pictures and videos and copy this orientation if we choose. For servers using unknown serving techniques, Waiter's Tray, etc. or rec players with physical limitations, consider the information, but you are on your own for injury risk issues.
Look at the shoulder angles of players at your courts.