feedback voltage is fixed internally at 2.5 volts and current sense is fixed internally to 1 volt
This may be a translation error. (I don't know your location) The voltages are not
fixed. It is true that when the supply is running
correctly the feedback voltage is 2.5V but it is not fixed.
Look at the block diagram.
The voltage on FB is compared to 2.5V.
....If FB is less than 2.5V COMP goes up in voltage.
....If FB is more than 2.5V COMP goes down in voltage.
....If FB is equals 2.5V COMP stays at the same voltage.
....The speed of COMP is set by a capacitor on the outside.
(COMP-1.4V)/3 this voltage goes to a voltage comparator that compares the voltage at CS.
....If CS is below (COMP-1.4V)/3 output=on
....If CS is above (COMP-1.4V)/3 output=off
........Output= off during part of the oscillator time period.
The voltage on CS comes from the current in the FET, Transformer. It should be a ramp. It should start at 0. There is a "spike" when the FET turns on. Then the current (voltage) should ramp up to a voltage (COMP-1.4V)/3 where the FET turns off and the current drops to zero. This ramp is because: (supply voltage, inductance in the transformer, time.) Without this ramp the IC can not work, or will be at maximum duty cycle!