Agilent Technologies AN 372-1 User Manual

Agilent AN 372-1

Power Supply Testing

Application Note

An electronic load offers a broad range of
operating modes, providing versatile loading
configurations needed for characterizing and
verifying DC power supply design specifications.

2

As regulated-power supply technology evolves,
testing methods for design verification and product
function require more sophisticated electronic equip­ment. The different power supply architectures
and output combinations also dictate the need for
versatile test instruments that can accommodate a
broad range of specifications. As a result, one test­ing requirement that has been growing in impor­tance is the method of loading the power supply
under test. The need for a higher degree of load
control due to test sophistication, such as the need
for computer programmability, has increased the
demand for electronic load instruments. The follow­ing examination of the most common power supply
architectures or topologies clearly illustrates the
growing need for higher performance and versatility
in electronic loads and power supply test equipment.

An Overview of Power Supply Topologies

Of all the possible power supply topologies, linear
and switching regulation techniques are the most
common design implementations. Linear power
supplies are typically used in R&D environments
and in production test systems because they pro­vide high performance, low PARD (ripple and noise),
excellent line and load regulation, and superior
transient recovery time specifications. However,

they are relatively inefficient when compared to
switching power supplies, and tend to be large
and heavy due to the heat sinks required to contin­uously dissipate power from the series transistors
and due to the magnetics used in this design. Typi­cally, linear power supplies provide a most effec­tive solution in lower power applications, and are
often used as subassemblies in various products.

Switching power supplies address the disadvan­tages of linear power supplies (namely the low effi­ciency and relatively large size and weight), and
are therefore a more effective and less costly solu­tion for high power applications. The relative dis­advantages occur in three areas when compared to
linear power supplies: slower transient recovery
time, higher PARD, and lower reliability. Switching
power supplies are used in a wide variety of indus­tries and environments, and are commonly found
as subassemblies in products such as computers,
computer peripherals, and copiers. Recent power
supply designs combine the best features of switch­ing and linear topologies. Below, Table 1 compares
the typical specifications for linear and switching
topologies.

Introduction

Table 1

Regulation Load Line Transient
Technique Regulation Regulation Response PARD Efficiency

Switching 0.05 – 0.5% 0.05 – 0.5% 1 – 20 ms 5 – 20 mVrms 65 – 85%

20 – 150 mVp-p

Linear 0.005 – 0.1% 0.005 – 0.1% 20 – 200 µs 0.25 – 5 mVrms 30 – 50%
(Series Pass) 1.0 – 15 mVp-p

3

Power supplies are used in a wide variety of prod­ucts and test systems. As a result, the tests per­formed to determine operating specifications can
differ from manufacturer to manufacturer, or from
end user to end user. For instance, the tests per­formed in an R&D environment are primarily for
power supply design verification. These tests
require high performance test equipment and a
high degree of manual control for bench use. In
contrast, power supply testing in production envi­ronments primarily focus on overall function based
on the specifications determined during the prod­ucts design phase. Automation is often essential
due to large volume testing, which requires high
test throughput and test repeatability. Power sup­ply test instruments must then be computer pro­grammable. For both test environments, measure­ment synchronization is necessary to perform some
tests properly and to obtain valid data. In addition,
considerations such as test set reliability, protec­tion of the power supply under test, rack space,
and total cost of ownership may be of equal impor­tance to the power supply test set designer. Proper
selection of testing instrumentation will provide
the best combination of measurement sophistica­tion and test set complexity.

Power Supply Testing Instrumentation

The power supply testing methods and configu­rations discussed in this application note are cer­tainly not the only means of obtaining the desired
measurements. However, certain instruments are
essential to all tests, regardless of the implementa­tion. Some commercially available turnkey power
supply test systems incorporate custom board level
instrumentation and hand wiring. However, power
supply test systems based on standard products
afford greater benefits. These systems are more
reliable and provide repeatable, high performance
measurements because of their low noise environ­ment. A system which utilizes standard instrumen­tation is modular, allows configuration flexibility
based on performance needs, and is easier to
upgrade. In addition, the service, replacement, or
calibration of separate instruments in the system
can be performed with minimal system down-time.

The tests covered in the following section are con­figured with standard instrumentation: electronic
loads, digital oscilloscopes, digital multimeters,
true rms voltmeters, wattmeters, and AC power
sources.

Electronic loads can facilitate power supply testing
in several ways. They are typically programmable,
although most require external DAC programmers.

This capability enables finer control over loading
values during testing, and can provide the test set
operator with valuable status information. These
loads are often designed with FETs, which provide
increased reliability over less sophisticated solu­tions consisting of relays and resistors. Also, these
products offer a selection of operating modes:
constant current (CC), constant voltage (CV), and
constant resistance (CR). The more sophisticated
electronic loads provide all three modes in one
product for optimum testing flexibility. They pro­vide a versatile solution for testing both DC voltage
and current sources. A final advantage is provided
by loads with readback over the bus. This can elim­inate the need for digital multimeters for voltage
and current measurements in some tests. As men­tioned, there are varying degrees of electronic load
sophistication. The Agilent Electronic Load family
provides all of the most sophisticated features and
high level performance in one box.

Several other instruments are required for power
supply testing. The performance criteria (accuracy,
resolution, stability, bandwidth, etc.) vary for each
test. In general, the measurement capability of the
instruments should ensure an error no greater than
10% of the measured specification. Table 2 on the
next page provides a guideline for instrument per­formance levels for each test discussed in this
application note.

An Overview of Power Supply Testing Needs

4

Load Transient Recovery Time

A constant voltage DC power supply is designed
with a feedback loop which continuously acts to
maintain the output voltage at a steady-state level.
The feedback loop has a finite bandwidth, which
limits the ability of the power supply to respond
to a change in the load current. If the time delay
between the power supply feedback loop input and
output approaches a critical value at its unity gain
crossover, the power supply will become unstable
and oscillate. Typically, this time delay is measured
as an angular difference and is expressed as a degree
of phase shift. The critical value is 180 degrees of
phase shift between the loop input and output.

Power Supply Tests

Table 2

Load Transient Current Limit Efficiency and
Recovery Time Load Effect Characterization PARD Power Factor Start-Up

Electronic Load t

rise

≤15 µs 1% programming 1% programming 1% programming 1% programming 1% programming

accuracy accuracy accuracy accuracy accuracy

Trigger output to CC or CR mode CR or CC mode CC or CR mode CC or CV mode CR mode

the oscilloscope Low PARD

Digital t

sample

≤100 ns N/A N/A t

sample

≤25 ns N/A t

sample

≤1 µs

Oscilloscope DC to 20 MHz Record length 1 K

minimum bandwidth samples minimum

Record length

≥2 K samples

100 µ/Div (linears)

1 mV/Div (switchers)

Digital N/A 51/2 Digits 51/2 Digits N/A N/A N/A
Multimeter ±0.005% accuracy ±0.005% accuracy

Wattmeter N/A N/A N/A N/A 1% accuracy with N/A

crest factors to
10:1 in current
waveforms

Regulated >1% regulation >1% regulation N/A >1% regulation >1% regulation >1% regulation
AC Source Adjustable peak Adjustable peak Adjustable peak Adjustable peak Adjustable peak

and frequency and frequency and frequency and frequency and frequency

Power factor Phase control

measurement
capability

RF rms N/A N/A N/A 100 µV Full scale N/A N/A
Voltmeter DC to 20 MHz

minimum bandwidth

Figure 1. Load Transient Recovery Time

Load transient recovery time measurements require an electronic
load with a risetime and falltime at least five times faster than the
power supply under test.

5

For a step change in load current, a marginally
stable CV power supply will have a ringing voltage
output. This defeats the purpose of the power sup­ply’s regulation circuitry and can be damaging
to voltage-sensitive loads. An example of a voltage­sensitive load is the logic circuitry in a computer.
In this case, a computer manufacturer that pur­chases power supplies from an external source may
consider verifying the load transient recovery spec­ification of the power supply subassembly. This
test can also reveal critical manufacturing flaws
that can cause instability, such as a defective out­put filter capacitor or loose capacitor connections.

Test Overview/Procedures

CV Load Transient Recovery Time is a dynamic
measurement of the time required for the output
voltage of a CV power supply to settle within a
predefined settling band following a load current
induced transient (see Figure 1). The response is
typically measured in microseconds or milliseconds,
and varies in value depending on the topology of
the power supply under test. The electronic load

used in this test should have a risetime at least
five times faster than the power supply under test,
and should be able to operate in CC mode (or CR
mode) up to the maximum current rating of the
power supply. Measuring the load transient recov­ery time requires the load to have the capability
to pulse between two different values in CC or CR
mode. For continuous load transient testing, the
repetition rate of the pulses should be slow enough
so that the power supply feedback loop can recover
and stabilize after each applied transient.

Figure 2 shows a typical test system for making load
transient recovery time measurements. Measure­ment of V

out

of the power supply can be made with
a digitizing oscilloscope as the load input pulses
are applied. Synchronization of the measurement
is crucial in obtaining proper measurements. There­fore, a common trigger should start the electronic
load and oscilloscope measurements.

Figure 2. Load Transient Recovery Test Configuration and V

OUT

Measurement Results for a CV Power Supply