Question regarding patent US4357582 (dielectric resonator oscillator)

In this patent it is stated that it is difficult to properly design oscillators shown on Fig1. to 5. Also other problems mentioned, and that new design can overcome those problems. Mainly solution is to use perpendicular feedback lines instead of classical parallel lines arrangement (in parallel feedback DRO). But i can't understand why there can be any improvement over classical parallel feedback lines. We just take gate stub and place it perpendicular to drain stub.

But practically i observed this difference, and perpendicular lines arrangement gives more freedom in placing DR. What is the source of this improvement? I only have idea, that with perpendicular arrangement there are more possible paths through magnetic coupling is possible, so phase condition is easier to obtain.

http://www.freepatentsonline.com/4357582.html

I'm not sufficiently knowledgeable in this area to say which of you is correct.

However I have come to suspect that patent articles often omit one or two crucial details (in the description or the diagram). There could be advantages for the inventor in doing so.

There is also the incentive for the inventor to say 'existing methods are problematic', and then claim these problems are solved by his invention.

So your understanding, and your observations, might very easily qualify you to fill in with comments, clarifications and corrections.