Empowering DC Supplies and AC Sources
by Tom Lecklider, Senior Technical Editor
You can’t physically bend a benchtop DC power
supply or AC source, but in all other senses of the word, they
are flexible with a capital F. Sure, you still can buy a simple,
dumb, single-output supply. But other than to save a few
dollars, why would you? Even a $300 DC supply offers remote
programmability; on-board memory to store V-I settings; constant
power output characteristics; and overvoltage, overcurrent, and
overtemperature output protection.
Courtesy of Agilent Technologies
Generally, these capabilities
and advanced features such as digital closed-case calibration
and fast pulse outputs have been developed to address real
customer requirements. Of course, if one company’s power
supplies or sources are gaining market share, their features
soon will be copied. Nevertheless, there is an underlying demand
for most of the newer functions that is dictated by specific
applications.
For example, semiconductor testing needs
short power pulses to avoid self-heating that would alter the
device characteristics being measured. And, there is continuing
emphasis on battery life in portable equipment. To minimize
power consumption, many small design changes are made, and at
each step, it’s important to accurately measure the load voltage
and current.
In addition, power supplies and sources have
benefited from newer semiconductor devices and control
techniques that have made possible higher efficiency, faster
transient response, and greater peak power in a smaller or
lighter-weight package. At first glance, these characteristics
and a steadily increasing list of standard features such as
programmability, a universal input voltage range, and accurate
regulation may appear to be only evolutionary. Some things are,
but among the specifications are some truly innovative advances.
Noteworthy Innovations
Constant Power
When you specify a power supply, you have to
address the present requirements as well as how they may change
in the future. The move from 5-V digital ICs to 3.3 V is a good
example. If you bought a conventional 0-V to 5-V, 10-A, 50-W
supply to handle the 5-V ICs, you still are limited to 10 A for
the 3.3-V devices. Newer supplies with a constant power mode
support the maximum output current for a 50-W rating or 15.15 A
in this example.
There are limits, so don’t expect 100 A at
0.5 V from the same 50-W supply. But within the manufacturer’s
specifications, you can trade volts for amps, keeping the
maximum power constant. Kepco’s Series KLP features Hyperbolic
Power™ technology, a descriptive name for a constant power mode.
For Model KLP 75-33-1.2K, the supply’s 1,200-W maximum power can
support any combination of voltage and current between 75 V @ 16
A and 36 V @ 33.3 A. The -1.2K suffix indicates a built-in LAN
port, a standard feature of the LXI-compatible KLP Series.
To View the DC Power Sources Comparison Chart click here.
To View the AC Power Sources Comparison Chart click here.
Chroma ATE’s 62000P Series has a
constant-power operating envelope also reflected in the model
number. The 1200W/80V/60A can provide 1,200 W output power at
any combination of voltage and current bounded by 80 V @ 15 A
and 20 V @ 60 A. And, B&K Precision’s 100-W Model 9110 offers a
wide operating area from 60 V @1.66 A to 20 V @ 5 A.
The Agilent Technologies N675XA and N676XA DC
Supplies use the term autoranging to describe the constant power
mode. For all constant power supplies, the locus of allowed V-I
combinations is a hyperbola.
The DC Power Supply Comparison Chart that
accompanies this article includes the maximum voltage available
from the lowest to highest voltage model in a series. Similarly,
the current corresponds to the maximum available from the models
in that range. For supplies with a constant power mode, the
power is listed as 100 W max, for example. The maximum
qualification is not used with conventional supplies.
Universal Input
DC supplies and AC sources often can accept
more than one range of input voltage. Typically, a choice must
be made between 115-V or 230-V nominal ranges although this may
be an automatic selection made by the instrument when power is
applied. Larger supplies and sources generally operate from
three-phase power. Sometimes, as with the Pacific Power Source
AC sources, many input voltages are available—100, 110, 120,
200, 208, 220, 230, and 240 to list just the single-phase
voltages.
In the AC Sources Comparison Chart, several
series of products have been listed. In one case, the entry for
a series was subdivided into groups of models. This approach
allowed a distinction to be made between lower-power units that
can operate from either 115-V or 230-V and higher-power units
that must have a 230-V input. In other cases, the products in a
series are treated as a group, but because two output voltage
ranges were available, specifications were listed for both
conditions.
Agilent’s N67XX DC Supplies feature a
universal input rated from 86 VAC to 264 VAC at 50 Hz to 400 Hz.
There are no switches to set or fuses to change when the
supplies are used in different countries or in test systems
running from different AC voltages. Lambda America’s compact ZUP
Series operates with inputs from 85 V to 265 VAC. Equipment with
a truly universal input range works well at either extreme and
anywhere in between.
Aside from the convenience of not having to
manually configure a supply for a particular input voltage, a
universal input range has an important benefit. If the power
supply or source is used with a 230-V input, it will continue to
operate correctly down to about 40% of the nominal value. No
selectable input configuration offers that degree of immunity to
AC supply dips or sags.
Power Factor Correction
IEC 61000-3-2 sets limits on harmonic
currents for products that draw less than 16 A per phase but
more than 75 W. Professional equipment that consumes more than 1
kW is excluded. However, many DC supplies and AC sources that
use switch-mode technology to reduce size, weight, and cost are
covered by the standard.
Typically, switching supplies create a raw DC
bus by rectifying the input AC voltage. The usual
capacitor-input configuration draws current pulses near the
peaks of the voltage sine wave, producing harmonic distortion.
To reduce the distortion, power factor correction (PFC)
circuitry attempts to make the input current resemble a sine
wave. Active PFC schemes can be very effective, producing power
factors near unity. Passive PFC may be less expensive on
lower-power products but also is less effective and not suitable
for high-power applications.
Data Acquisition
DC supplies and AC sources with a readback
feature are capable of responding to a controller or PC with a
measured value of output voltage and current. To do this, the
equipment must have a built-in measurement system with a digital
output. A readback capability is not new, but the responsiveness
of today’s faster measurement systems is.
Agilent’s recently introduced Model N67XX DC
Power Analyzer takes advantage of the 50-kS/s sampling rate of
the measurement ADC in the N67XX supplies. Up to four N67XX
power supplies can be housed in the analyzer mainframe and
controlled from a front panel with an integral color display.
Individual outputs can be programmed to turn on/off in a
particular sequence with specified timing and levels.
In addition to fast measurements, the N67XX
supplies also have sufficient output bandwidth to function as
arbitrary waveform generators (Arbs), directly generating
user-defined disturbances. Supplemental characteristics now
specify a 3-dB frequency bandwidth at five output levels for
many of the supplies in the series. For example, the 35-V Model
N6774A 300-W Supply achieves 125-Hz bandwidth at 0.35-V or 0.7-V
pk-pk but only 40 Hz at 3.5 V and 20 Hz at 35 V.
As useful as the power analyzer is, its
performance depends entirely on the power supplies it uses.
Agilent promotes the instrument as combining a DMM,
oscilloscope, Arb, and datalogger, but these capabilities are
only as good as those of the individual power modules. You do
not get a 10-MHz bandwidth general-purpose scope although the
10-kHz bandwidth corresponding to a 50-kS/s sampling rate is
helpful.
If you’re only concerned about the changes
occurring to the power driving your DUT, the power analyzer can
do the whole job. It’s likely that you would use it to control
the power sequencing and deliberate disturbances needed to test
a DUT at the corners of its input power specification. However,
you will need a conventional scope to monitor points of interest
in the DUT circuitry to determine unusual behavior such as
high-frequency oscillation. Only sampled supply output voltage
and current can be viewed on the analyzer display.
Fast Transient Response
Fast transient response is important in many
applications, portable device development being a very popular
one at the moment. Many of these products can transition almost
instantaneously from a near-zero-power sleep mode to a
full-power-on state. The trade-off encountered when designing a
wide-bandwidth power supply output stage is stability for a
range of loads. Digital control is attractive because it can
tune the output characteristics to the load, reducing transient
voltage droop and the time to recover to large load changes.
Agilent’s Kevin Cavell, product manager for
power products, explained some of the techniques used to achieve
the required performance: "Our most recent products use digital
control to protect the end customer’s load through digital-based
monitoring of key facets of the power supply. We also use
digital control in the most complex way, which includes digital
regulation of the output voltage and current through both linear
and digital feedback systems. And, digital control is used for
power management for redundancy and budgeting in limited power
environments."
Specifications
Accuracy
Although many data sheets for DC supplies and
AC sources are well written and comprehensive, some are not. A
common failing is to assume that the reader will understand the
meaning of 0.1% accuracy. If every one meant 0.1% of the actual
value, that would be fine. Unfortunately, it’s very convenient
to create overall specifications for a series of power supplies
or sources based on percentage of rated output or range or full
scale—the same things.
It’s typical for a higher voltage output to
have greater ripple and noise than a much lower-level output.
Specifying ripple, noise, or accuracy based on a full-scale
output covers the whole series of products in a simple one-line
statement. A problem occurs when it’s unclear if the percentage
relates to full scale or the actual output value, and it can
make a big difference.
Periodic and Random Deviations
On another topic, consider the periodic and
random deviations (PARD) Vrms and Vpk-pk values listed for
several DC supplies. PARD generally is measured in a 20-MHz
bandwidth.
Immediately obvious is the distinction
between bulk DC supplies and more closely regulated precision
supplies. Chroma Systems Solutions’ Series 62000B quotes up to
200-mV pk-pk PARD. The 62000B Series of 1.5-kW modules can be
configured to provide up to 120 kW. In contrast, precision
supplies may have only a few millivolts of PARD whether measured
as rms or pk-pk.
It’s interesting to note the wide range of
ratios between listed rms and pk-pk PARD values. A general rule
of thumb for relating rms and pk-pk is to use a factor of about
6:1. Mathematically, there is no direct relationship, because if
the PARD truly is random, eventually the maximum pk-pk value
could be very large. Practically, 6:1 is a reasonable value.
Nevertheless, for companies that list both rms and pk-pk PARD,
the factor varies greatly.
Chroma’s 62000P Series quotes 15-mV rms and
100-mV pk-pk, close to 6:1. In contrast, Kepco’s KLP Series has
10-mV rms and 125-mV pk-pk, a 12:1 ratio. Protek’s Model 6006S
has 5-mV rms vs. 100-mV pk-pk, a ratio of 20:1. Perhaps the
larger ratios indicate that more of the noise is related to the
switching frequency and consequently not random. Very narrow
noise pulses add little to the rms value but greatly influence
the pk-pk measurement.
Output Flexibility
There’s nothing special about the polarity of
a DC supply unless you need the other one. Many supplies specify
±Vout, and they mean it. The outputs are floating and can be
configured for either polarity of operation. In contrast, some
multi-output supplies may connect all the low output terminals
to the same ground reference, eliminating the flexibility you
have with separate supplies. It’s a small point that may not
even be highlighted in the data sheet other than to list the
output rating with a ± prefix.
A similarly easy-to-overlook specification
relates to supply output parallel operation. Especially given
the trend to higher-current, lower-voltage ICs, paralleling N
power supply outputs often is done. Active current sharing
ensures that each supply provides 1/N of the total. Many
supplies can be operated in parallel, but only a careful reading
of the data sheet will reveal how the output currents will be
distributed.
Standard vs. Optional Features
Confusion between what is and isn’t claimed
and under what conditions is especially prevalent in AC source
data sheets. Some of these products can provide virtually any
waveform you may need at a wide range of power levels. The
problem is that you may require a special controller or software
to access all the capabilities.
Understandably, a manufacturer wants to
highlight the functionality, accuracy, flexibility,
programmability, and applicability of its product. And, many
times, new features are provided in software or through a
controller firmware upgrade. But all the great new capabilities
often displace basic specifications, making the data sheet less
informative than it should be. Make certain that the unit
fundamentally meets your requirements, and then determine if the
extra capabilities might be useful in your application.
Summary
Table 1 summarizes
a long-term trend toward smaller, lighter DC supplies and AC
sources. Advances in switch-mode power control are largely
responsible, together with microprocessors, better semiconductor
processes, and improved high-frequency transformer and choke
materials. Although this table shows basic changes, it doesn’t
indicate the greatly increased capabilities of these products
that were gradually added over the years.
Table 1. 30 Years of 5-kW DC Power Supply Progress
Courtesy of Lambda Americas
According to Lambda Americas’ Product Manager
for High Power John Breickner who developed the table, "Users
increasingly are using digital control via serial, GPIB, or LXI
interfaces. They usually write their own system software but
turn to the power supply manufacturer for drivers as the
building blocks peculiar to a specific supply. It’s most
convenient for users to have a series of supplies or sources
with identical drivers and the same command structures."
Certainly, it can be advantageous to use the
most efficient, smallest, and lowest-cost product that will do
the job. On the other hand, surprises may be lurking in a test
application just waiting to catch the unwary engineer.
Mitchel Orr, sales manager at Pacific Power
Source, cautioned, "The linear amplifier type of AC source still
has its place, outperforming many switch-mode designs with its
lower output impedance and faster load transient response. The
uninformed test engineer frequently selects a product that’s too
large or small for his application and may not understand the
important differences between switch-mode and linear
technology."
Table 2 gives a direct comparison of the
company’s AMX linear and ASX switch-mode AC sources and
emphasizes the points Mr. Orr made. To be fair, the transient
response figures are only typical values, but the AMX Series
boasts a 50-kHz small-signal bandwidth compared to 5 kHz for the
ASX Series. The AMX also can drive a more diverse range of load
impedances with its high 6:1 peak-to-rms current capacity. The
ASX is a fine series of products and appropriate for many
applications. Nevertheless, there are differences in the
characteristics that linear and switch-mode technologies
support, and they could be important in your application.
Table 2. Linear vs. Switch-Mode AC Source Comparison