It works something like this:
When the switch is closed an extra current flows
through the inductance and stores energy there. The capacitor
supplies the load with current during this time.
After the switch closes the capacitor is charged
by the energy stored in the inductance and an extra current starts
flowing through the load, causing the output voltage to rise
(energy is supplied directly from the input source also as long as
the diode is forward biased). During this time, the system behaves
like a RLC-circuit, so, after a while, the current decreases. The
switch is then closed again and the cycle repeats. One could say
that charge is pumped from input to output, increasing the output
voltage up to the point where there is an equilibrium between the
discharging of the capacitor while the switch is closed and the
charging by the inductor while the switch is open.
The output voltage equals (ton / toff
+ 1)×Uin and is controlled by PWM of the switching action.
To implement this, I have used the LM2577T-ADJ
from National Semiconductor. It
operates conform the given discription and is connected like so:
|R1 and R2
||Voltage devider for monitoring
||20Kohms pot. (Bourns)
||0.1µF, 63V MKS condensator
||Use a good quality coil!
||160µH toroïd (2.5A, 70mohms,
||Current higher than output
||FR603 60V reverse breakdown, 3A
|Rc and Cc
||Pole-zero compensation network
||2200ohms, 5% and 1µF, 63V elco
||Get a low ESR type!
||2200µF, 16V elco (Telecon)
You can download the PCB
design here (only 4Kb). It's in CorelDraw 3.0 format (zip
by Oscar den Uijl,