I am looking at: https://en.wikipedia.org/wiki/Euler%27s_equations_(rigid_body_dynamics)
\begin{align} I_1\dot{\omega}_{1}+(I_3-I_2)\omega_2\omega_3 &= M_{1}\\ I_2\dot{\omega}_{2}+(I_1-I_3)\omega_3\omega_1 &= M_{2}\\ I_3\dot{\omega}_{3}+(I_2-I_1)\omega_1\omega_2 &= M_{3} \end{align}
and think of a sphere, where \begin{align} I_1 = I_2 = I_3 = I \end{align}
thus \begin{align} I_1\dot{\omega}_{1} &= M_{1}\\ I_2\dot{\omega}_{2} &= M_{2}\\ I_3\dot{\omega}_{3} &= M_{3} \end{align}
Assuming my model has inputs as: \begin{align} {\omega}_1 = 1 && \dot{\omega}_1 = 0\\{\omega}_2 = 0 && \dot{\omega}_2 = 0\\ {\omega}_3 = 0 && \dot{\omega}_3 = 1 \end{align}
We see: \begin{align} {M}_1 &= 0 \\ {M}_1 &= 0 \\ {M}_3 &= I \dot{\omega}_3 \end{align}
which is only the effect of inertia as if no rotation about 1 had been applied.
This indicates, that the gyroscopic effect vanishes in case of a sphere, and inverts when a disk turns into a rod (I_1 < I_3), where 1 is the axis of constant rotational speed.
I wasn't able to find any evidence of my conclusion on the internet. Am I missing something?