I'd think the decision (serial or parallel) depends on the source impedance. Try and simulate it!
Hi,
I had seen some strange behaviour when I simulated chargepump (dickson) constructed using MOSFETs.
What is the switching frequency to be used? How should we select Cap value?
I have chosen 500KHz as the clock frequency and cap is 1pf.
I tried b varying the clock frequency from 250KHz to 1GHz. This charge pump is used for low power RFID applications@ 865MHz.
I am aware that there are -ve implications when we go for higher frequency. but I experimented to see the behaviour.
Moreover, I changed the width of the MOSFET from 240nm to 1um.
Strangely, I saw that as the width is increased, output is decreased. I assumed that as width of MOSFET is increased, more will be the drain current and so output voltage must be higher. Surprisingly, I found the reverse, wherein, output voltage started decreasing as the width is increased. Why is it happening so?
I changed the capacitor value from 0.01pf to 1uf. Here also, I agree that cap should be small in order to see faster response time. But still, I experimented to see the results. To my surprise, I found that as cap is increased, response time is decreased. I don't understand why is it happening so.
Can you please explain the logic behind this?
Thanks!