■ FIGURE 10
current and equivalent series resistance (ESR). ESR is a
measure of how much power a capacitor will dissipate
as current flows in and out of it. When a DC voltage is
applied to a filter cap, any AC ripple on the DC will
cause an AC ripple current. The interaction between ESR
and ripple current produces heat in the capacitor.
For C2, we can use a general-purpose aluminum
electrolytic capacitor. The relatively slow rise and fall times
of the voltage applied to C2 causes only a small amount
of ripple current. The value of C2 needs to be relatively
large to prevent the circuit from oscillating as you adjust
the output voltage over its full range. I used 1,200 μF.
Look again at Figure 9. That square wave shape means
high ripple current, so C1 must have very low ESR to prevent
the capacitor from failing due to heat. (I’ve seen capacitors
get so hot they burned my finger.) I used a Nichicon type
HE low-impedance aluminum electrolytic for C1.
The output can be adjusted from a minimum
of about 1.2 volts to a maximum close to the
input voltage. The standard version LM2576 has
a maximum input of 40 volts. The HV version has
a maximum input of 60 volts. For this design,
Vin is limited by the 35 VDC capacitors. Vin can
be supplied by a fixed power source such as a
battery. For many applications, I like to use
unregulated, wall mounted DC “power blocks;” inexpensive
surplus units can be obtained from many vendors.
Assembly is straightforward. The only thing that takes
a bit of care is mounting the LM2576 to the PCB. The IC
uses the TO-220-5 staggered-lead hole pattern as shown in
Figure 10. While there is a pre-formed version of the IC,
they’re not always available. Most likely, you will have to
bend the LM2576 leads to fit the pattern.
First, give the board a careful visual inspection. Look
for the usual suspects:
The Feedback Path
The potentiometer R1 across the output supplies the
feedback required by the LM2576 to keep the output voltage
constant at the selected value. The value of R1 is important.
If it’s too big, then output voltage will drop as you draw current.
If it’s too small, you’re wasting power. A 2K value works well.
I used a multi-turn trimpot. A single-turn trimpot will work, but
it will be tricky to set the output voltage to a specific value.
• Bad solder joints
• Broken copper traces
• Shorted copper traces
• Reversed polarity on capacitors
• Diodes in backwards
• Component leads touching (especially the LM2576 leads)
Once you’re sure there are no obvious problems,
connect a 470Ω, 1/2 watt resistor to the output terminals
as a test load. Connect 12 volts DC to the input terminal.
Connect a voltmeter across the
load resistor and adjust the
PARTS LIST trimpot to get five volts output. If
all is well, replace the load with a
ITEM DESCRIPTION SUPPLIER 10Ω, 5 watt resistor and verify
❑ Printed Circuit Proto-board or custom RadioShack #276-150 that the output is still five volts. If
Board PCB (see text) it isn’t, re-do the visual inspection;
❑ Inductor 220-330 μH rated at 1A you must have missed something.
DC minimum (see text) That’s all there is to it.
Two contacts, .200 spacing
2 k W trimpot Switching regulators are not
330 μF @ 35 VDC, low as difficult as you might have ESR aluminum electrolytic, thought. NV
radial leads (see text)
❑ C2 1,200 μ @ 35 VDC, general-purpose aluminum electrolytic,
❑ Terminal Blocks
Molex #39890-0302 orequivalent
Bourns #3296W-202LF orequivalent
Wakefield 230-75AB orequivalent
The PCB and/or complete kit
for this project can be
purchased through the
Nuts & Volts Webstore at
or call our order line at