Wiki

This version is outdated by a newer approved version.DiffThis version (06 Apr 2018 18:06) is a draft.
Approvals: 0/1

This is an old revision of the document!


Activity . Power Conversion: Switched Capacitor Methods

Objective:

The objective of this activity is to understand the operation of several switched-capacitor power conversion circuits.

Background:

A high quality power supply is essential for nearly any electronic circuit. If you have never had a circuit fail because of a power integrity issue, it is likely because you are starting with a high-quality power supply. But what if the only supply you have available is nominally the correct voltage, but varies over time or temperature - your 9V-battery powered op-amp circuit may work fine when you first power it up, but begin to fail in new and interesting ways as the battery voltage drifts down to 8V, 7V, 6V, 5V, etc. This particular situation can be remedied with a linear voltage regulator, supplied with a voltage higher than the regulated output voltage. An LM7805 will provide a stable, 5V source when powered from a 9V battery. This is done by adjusting the drive to a power transistor, such that the voltage drop across the transistor results in the correct output voltage.

Linear regulators have widespread application and are well deserving of their own activity, but there are some things that a linear regulator just can't do:

“Boost” a lower voltage to a higher voltage “Invert” a positive voltage to a negative voltage Reduce a high voltage to a lower voltage any more efficiently than could be achieved with a power resistor in series with the supply, at an equivalent output current. This is where other techniques must be employed. One such technique is “switched-capacitor” voltage conversion, and conceptually involves little more than capacitors, switches, and some form of control to the switches.

Side note - A quick survey of switching regulators will reveal that the vast majority involve inductors, rather than capacitors. There are reasons for this, but for understanding power conversion, there is merit to starting with capacitors. Capacitors are easier for many students to grasp (literally), at least in terms of energy storage. A large capacitor charged from a battery stores the energy as a static electric charge. You can disconnect the battery, and hold the stored energy in your hand for a “human perceptible” amount of time - seconds, minutes, even hours or days. This is generally not possible with inductors (unless they are superconducting) - they store energy in a field induced by a moving current. One could attempt to quickly switch out a battery from an inductor, and short the inductor such that the current continues to circulate, but the current will rapidly decrease to zero (for any inductor that you could hold in your hand, anyway.)

Materials:

ADALM2000 Active Learning Module
Solder-less breadboard
Jumper wires
PC running LTspice and Scopy
LT1054 Switched-Capacitor Voltage Converter
Electrolytic capacitors from ADALP2000 parts kit
Solderless breadboard
5V USB power supply
Voltmeter (optional, can use M2K in Voltmeter mode.)

XX.0 Human-switched capacitor power conversion

The introduction alluded to being able to hold a capacitor with stored energy in your hands. Let's do a quick experiment and make a human-switched-capacitor voltage inverter.

Say you have a 9V battery, and you want to make a split +/-9V supply. Hold a 330uF cap to a 9V battery (OBSERVE PROPER POLARITY!), then connect cap (+) to battery (-), measuring from battery (-) to cap (-).

=re-take photos=

This “flips” the positive voltage across the capacitor below ground, and you now have a split supply… until the capacitor discharges, either through your circuit, or through internal leakage.

Similarly, if you have a 9V battery need a higher voltage, say 18V, you can charge the capacitor again, then connect capacitor (-) to battery (+). This “stacks” the positive voltage across the capacitor on top of the battery's own 9V, producing 18V… until the capacitor discharges, either through your circuit, or through internal leakage.

If you do this fast enough, your circuit won't notice the switching and will see a continuous, steady voltage, but it would quickly get tiring for the human switcher.

=re-take photos=

With a basic idea of switched capacitors, let's proceed to something more practical.

XX.1 Voltage Inversion

Some op-amp circuits can operate on a single supply, with the op-amp negative supply pin connected to ground. However there are applications that benefit from the use of a “split” supply with voltages of opposite polarity referred to circuit ground, for example, with the op-amp's supply pins connected to positive 5V and negative 5V. Such a supply can be created by using two 9-V batteries, an LM7805 positive regulator, and an LM7905 negative regulator. But this extra battery is an inconvenience - when is the last time you saw a product that required TWO 9V batteries?

This is where a switched-capacitor inverter can come in handy - A single battery can now provide both a positive supply (direct connection to the positive terminal), and a negative supply (through an inversion circuit.) Before testing real circuits, let's get a feel for how the circuit works by runnign an idealized simulation. This simulation essentially automates the procedure from Experiment 0.

The LTspice file is included here: (link to ltspice file)

The key element in this simulation is a switch:

which is assigned a value of my_sw. Any switch in the schematic with value my_sw will have properties defined by the spice directive:

.model my_sw sw(Vt= 0.5 Vh=-.25 Ron=0.01 Roff=100Meg)

Which says:

  • When the ( + ) input is greater than the ( - ) input by more than 0.5V, the switch resistance is 0.01 ohms
  • When the ( + ) input is not greater than the ( - ) input by more than 0.5V, the switch resistance is 100 million ohms

Why not zero ohms and infinity ohms? SPICE can get confused when the range of values in a simulation covers too great a range. In this simulation, we're charging and discharging capacitors, and without some simulated resistance, currents will approach infinity. The Vh parameter is the hysteresis at the input, which ensures that the switch changes state properly.

The 0.5V threshold was chosen so that the switch can be controlled by an intuitive voltage level, where 0V = OFF and 1V = ON.

The switches are controlled by two pulse generators, with outputs labeled clk and clk_bar.

Right-clicking will bring up the parameter windows:

Each source outputs 3 microsecond pulses with a 10 microsecond period. The only difference is that V3 is delayed by 5us - this produces a “non-overlapping clock”, which allows switches to be alternately turned on and off, and never on at the same instant. Running the simulation, and probing q and q_bar, shows what's going on:

The last step in making the idealized simulation is to put the pieces together, to do what human hands did in experiment zero. When clk is asserted (at 1V, closing S1 and S2), the capacitor is charged to +5V:

When clk_bar is asserted, the left terminal of C1 that was charged to 5V is then grounded (to zero volts), which drives the right terminal negative:

Run the simulation, and probe vout, clk, and clk_bar.

Notice that the output voltage does not immediatly reach its final value, but takes several “steps.” This is because while C1 is initially charged to 5V, C2 is initially discharged. Thus the charges in each capacitor are:

q(C1) = 5V * 1uF = 5uc

q(c2) = 0V * 1uF = 0uc

When S3 and S4 close, C1 and C2 are placed in parallel, so the 5uc is then divided among two 1uF capacitors, resulting in a voltage of -2.5V. Subsequent charge / discharge cycles drive the output voltage closer to its final value of nearly -5V. The 1k load resistor prevents the output from ever reaching exactly -5V, but if the switching is fast enough, it can come close.

Questions:

Add questions here:

Return to Lab Activity Table of Contents

university/courses/electronics/switched-cap-power-supplies.1523030818.txt.gz · Last modified: 06 Apr 2018 18:06 by Mark Thoren