[ List Subjects ][ Main Page ]
Subject: Re: DC/DC Converter - Charge Pump
[ View Followups ][ Post Followup ]
Date: 05/31/00 at 3:21 AM
Posted by: Dominique Bovey
In Reply to: Re: DC/DC Converter - Charge Pump posted by Roberto Casas on 05/31/00 at 12:57 AM:
The maximum current you will get @5V form this source will be naturally 500uA,
assuming a conversion efficiency of 100%.
Given the low levels of current the circuit must use to work, and the "high" level of the
input voltage it seems that excludes IC-based designs, so discrete transistors must be used.
Several possibilities I see:
1) Set up a 2-transistor astable multivibrator (running at a few kHz or tens of kHz) to chop the
50V, with a duty cycle of 10%. Smooth the output (at one of the collectors) with an LC low
pass, and you will get your 5V. But the efficiency will probably be not more than 40-50%.
2) a discrete transistor approach of a "buck" step-down regulator could be a solution, but
then there is a regulation loop you will have to "tweak".
3) A discrete charge pump, transferring the charge from an input capacitor to a 10 times
bigger capacitor output capacitor, because the charge is Q=C1*Uin=C2*Uout (total discharge),
and the current being I=Q*f (f switching frequency).
In that case you will have to design two "floating" switches
preferably MOS based. Run the circuit at like 20kHz, then Q=25nC, C1=0.5nF, C2=5nF.
take at least 10-20 times these value for low ripple.
4) there are certainly other solutions!
In all cases if you want to reduce the output voltage ripple at the minimum I advise you to
1) make the circuit as compact as possible (SMD, short and thick PCB tracks where the switch
currents flow). Remember , Uspike=L*di/dt, L being 10-20nH/cm of wire or PCB track.
2) output a little more than 5V (like 5.5V), and regulate to 5V using a micropower low drop-out
linear regulator (check at www.linear.com or www.maxim-ic.com for example)
Hope it helps,
Post a Follow-up: