The characteristics of PWM off for a long time are checked using the evaluation DC2296A board.The rise time was measured with 4 LEDs, but the rise time is delayed to 50 usec. The PWM off time during that time is 30 seconds.
If the PWM frequency is confirmed to be 100Hz, it rises at 2 to 3 usec.So, when I checked the voltage of the device at the time of PWMoff, the voltage of Vout and Vc was 0V.
The evaluation board is a circuit as shown.Is there a way to improve it?
When using the LT3932 for flash applications, the first "flash" pulse is used to help the part find where to regulate the following pulses, but isn't a perfect pulse itself (hence the long rise time). In a well designed system, the following pulses after this first pulse should all be well regulated and square in waveform shape. This first pulse is typically not used in a real system and will only need to occur once while the system maintains regulated input power.
For your application, I would recommend first changing the inductor L1 value from 1.2uH to something more like 4.7uH or 6.8uH with similar saturation characteristics (>2A ISAT). I think that the output capacitor could also be swapped out for a 25V ceramic cap (same value or 10uF). Another thing to try would be to remove C15 and remove/short FB2.