Here is my approach for a (rough) loop gain calculation:

• The loop is opened at the base node of T2

• The voltage gain at the node of T2 is

  G=v_c2/v_b2=-(1/2)gm_2 * r_load * 2=-gm_2 * r_load. (the factor 2 results from the current mirror)

• The dynamic load resistance is r_load=hie_7(1+gm_7 * R_e) with R_e=R1+(R2||r_in2)

• Because of the voltage follower (T7) we can set v_c2=v_e7 and the voltage v_f across the feedback resistance r_f=(R2||r_in2) can be found using the voltage division between r_f and R1.

• Therefore, the loop gain is G_loop=G * [R2||r_in2]/[R1+(R2||r_in2)]

Here is my approach for a (rough) loop gain calculation:

• The loop is opened at the base node of T2

• The voltage gain at the collector node of T2 is

  G=v_c2/v_b2=-(1/2)gm_2 * r_load * 2=-gm_2 * r_load. (the factor 2 results from the current mirror)

• The dynamic load resistance is r_load=hie_7(1+gm_7 * R_e) with R_e=R1+(R2||r_in2)

• Because of the voltage follower (T7) we can set v_c2=v_e7 and the voltage v_f across the feedback resistance r_f=(R2||r_in2) can be found using the voltage division between r_f and R1.

• Therefore, the loop gain is G_loop=G * [R2||r_in2]/[R1+(R2||r_in2)]