[need help] Make LM3488 regulate output from 0V

[Unixon]

Hi,

I have a piece of hardware with LM3488 in non-isolated forward configuration (common ground) with 1:10 step-up transformer (ATB3225) powered from 5V input. LM3488 is used as a digitally controlled voltage source in [0-150V] range to drive a PZT actuator. The problem is the lowest stable output voltage this solution is capable of producing is about 12V (8% full scale, heavily depends on output loading) at f_sw=100kHz and I need to go down to 5V or less. At f_sw=400kHz the lowest stable Vout is ~25V. Cout~1-10uF, Rfb~0.5MOhm.

I'm probably hitting the minimum "on time" T_on=325ns limit, or something like that and I'm already at the minimum switching frequency. Reaching this limit results in low frequency like 50Hz triangle wave oscillations at full Vout amplitude instead of stable constant Vout.

Loading the output with additional capacitance or introducing an LC-filter after secondary transformer winding breaks regulation much faster than lowers stable output voltage. So, the only other attempts to calm LM3488 down I can think of is limiting the primary side current with a higher value current sense resistor and/or lowering the switching frequency beyond the spec limit to reach a lower minimal duty cycle.

Will LM3488 tolerate lower switching frequencies like, say, 30kHz?

[Unixon]

Update:

This topic describes similar issue with different topology:
https://e2e.ti.com/support/power-management/f/196/t/333595?Reg-LM3488-Boost-convertor-output-oscillation-without-minimum-load

Here's the piece of LM3488 circuit: (see attachment)
C33,C34 are replaced with 3x0.47uF caps.

Regulation above minimum Vout threshold of ~12V is perfect, but below this level I also get low frequency full amplitude oscillations probably for just the same reasons.
Is it possible to bring this minimum output voltage down?

p.s. Most of this circuit previously belonged to DRV2700, but I had to drop that IC because of its naturally high ripple at certain Vout ranges.