Your diagram unfortunately shows the xfmr secondary being short-circuited! Lotsa smoke, maybe. Not good smoke like pipe tobacco, either.
As gneill says, put diode in each secondary leg (not the center-tap). Cathodes connected together, anodes to each secondary high (pointing to the right). Then, the 100 ohm CT resistor is too big - at max output current = 7.5V/120 = 62.5 mA, the CT voltage is already at - 6.25V so each secondary high winding can't go above -6.25 + 115/(2*5)V. Moreover, these voltages are squirrely - combination of sine and chopped-off full-wave rectified sine.
I would ground the CT and move R to between the diode cathodes and your zener/resistive load. To avoid overheating the zener, 7.5*(11.5-7.5)/R = 1W. This limits the zener to 1W peaks in the no-load condition.
Unfortunately, your output voltage will not be constant; the zener will not always be turned on. To achieve that goal, you need a capacitor shunting your load. I would guess RC = 8ms with worst-case R = 120 so C = 68uF. This would be a polarized cap so make sure you get the + end on the output & the - end to ground, not vice-versa.