problem in regulated dickson charge pump

[slamnas]

Hi, I'm new on this site is the first I filed a problem.
I am designing a regulated charge pump 'Dickson topology' (Specification: power supply = 3.3 V, 5 V regulated output, Cload=10pF) I used two stages (2 capacitors 1pF pumping with a clock frequency is betwin 500khz and 1MHz . I use the simulator spectreS. The technology is TSMC035.
I have a problem in the level regulation. voltage drop when using a voltage divider with two resistors in values depend on the feedback voltage. Comparator 2-stage OTA with hysterisis works well alone.
files attached below shows the result of pumping without regulation and with control

 

[keith1200rs]

I cannot tell from the circuit - I cannot see a value on the resistors, but are you sure you aren't just loading the pump too much with the resistors? With 2 stages you will struggle to get a usable current I would have thought (well under 1uA).

Keith.

 

[slamnas]

thank you for your reply. Anyway I do not know how to regulate a charge pump. Here are the specifications: (VDD = 3.3V; Cload = 10pF; Voutreg = 5V ripple voltage less than 30mV clock frequency is 500kHz and the pump must deliver a current greater than or equal to 40mA). Can anyone help me?

 

[keith1200rs]

That is a huge amount of current! Are you planning off chip capacitors? You are going to need some large capacitors for that.

This is a useful presentation about various topologies:

www.ocf.berkeley.edu/~eng/classes/EE290cPresentation.ppt

Also, JSSC Vol32 No 6 June 1997 page 852 has some useful information - changing the frequency according to load although with only 10pF output capacitance you will not get the ripple you seek without linearly regulating. However, the 10pF decoupling limit doesn't make sense when you will most likely need capacitors around 1uF for the pump!

Keith.

 

[slamnas]

thank's for your helps

 

[neddie]

Your charge pump caps are WAY to small for a output of 40mA
Increace them to a couple of uF.

 

[slamnas]

but I saw in a book for a DC-DC boost converter, it uses a PWM comparator whose input (-) is the output of the error amplifier and the input (+) is a ramp. Is this the case for a switched-capacitor converter. And how to build the ramp? And what values Min and Max of the ramp?
My goal is to obtain from an input ranging from 2.8 to 3.3 V an output regulated 5V at 50 mA

 

[keith1200rs]

You cannot avoid the laws of physics. You want to move a certain amount of power through some capacitors at a given frequency. The amount of power you want means you need capacitors in the uF region.

Keith

 

[slamnas]

PWM comparator or ERROR AMPLIFIER in regulated the output for switched capacitor converter (charge pump)?