The pump is just set at 30khz and the mosfet driver is been driven by me(through switches),I am using simple electrolytic capacitors(470uf/35v) and in4148x2(using in parallel insted of single in5818)because they arn't available at my place.
I'm not sure if the simple electrolytic will work properly in such high frequency.
20KHz is more than enough for the pump I wouldn't use any higherm this is 50% duty right?
Can you try using a pump frequency of about 5-10KHz to see if you get the same results?
Sorry For the late reply,Yeah the charge pump is stable with a voltage of 19.6v output with a 30khz frequency,When hooked up with a single mosfet driver gives a voltage drop of 1 v only.Changed the value of both the caps to 470uf/35v each.
I still think it is the quality of the capacitors and a lower value low ESR capacitor used at a lower frequency of 20KHz would give you very good results.
Have you tried with a pump frequency of 5-10KHz to see if you get better results ?
I have tried different combinations of frequency and the best result is on 30khz and for 15 khz the voltages became low of about 15vdc.so 30khz is the sweet spot for charge pump.
This is nowhere near the simulation result, using 20KHz at the pump with the 10uf and a resistor load of 100 ohm (almost as much current as all four drivers) I got a ripple of 0.5v.
Do you have any MKT capacitors that you can use , even if you have 2.2 or 3.3uf try to add them in parallel to the 10uF capacitors, also add a MKT in the input capacitor (the 470uf connected to the 12v of the pump totem pole)
A rough estimation suggests, that the dimensioning of the charge pump will be O.K. with 10 uF ceramic caps or good low ESR electrolytic. But switching times have to be considered, so I suggest a simulation.