Hello,

Kai, thank you for your answer. You really have a lot of experience (and luckily for us) willing to share. I will try your method. But for now I changed inductance 220uH that comes on output of MC34063. I had 220uH/400mA and it looked like resistor. I assumed with higher currents it will heat up a bit, but I guess it goes in some kind of saturation too. I dissasembled car phone charger (it is built with MC34063) and put inductor to my circuit. Results were drastically improved. Maybe because of different type of inductance? It is a toroid (I'm not sure if this is an English word - well it looks like donut :)) and it does the job better. Tommorrow I am taking it to work to measure inductivity. I will try to add additional filter 1uH like datasheet says. But even now results are better. I have 100mV ripple at full current (I dit not measure it but I fill 1800mAh NIMH cells). Ripple is present on the output of L4940V5. I will try tantal.

I use low drop reg. because in case of power failure I draw 5 volts from 4 nimh cells.

Again thank you for helpful answer. I will investigate options you proposed.

regards, Damjan