DC-DC Converter Applications

DC-DC CONVERTER APPLICATIONS
Contents
terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
connecting DC-DC converters in series . . . . . . . . . . . . . . 133
connecting DC-DC converters in parallel . . . . . . . . . . . . . 134
recommended values for paralleled DC-DC Converters . . . . . 135
filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
2 - APPLICATIONS
recommended values for filtered DC-DC converters . . . . . . . 136
limiting inrush current . . . . . . . . . . . . . . . . . . . . . . . 137
maximum output capacitance . . . . . . . . . . . . . . . . . . . 137
settling time . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
isolation capcitance and leakage current . . . . . . . . . . . . . 138
overload protection . . . . . . . . . . . . . . . . . . . . . . . . 139
input voltage drop-out (brown-outs) . . . . . . . . . . . . . . . 142
no load over voltage lock-out . . . . . . . . . . . . . . . . . . . 142
long distance supply lines . . . . . . . . . . . . . . . . . . . . . 143
LCD display bias . . . . . . . . . . . . . . . . . . . . . . . . . . 144
EIA-232 interface . . . . . . . . . . . . . . . . . . . . . . . . . 144
3V/5V logic mixed supply rails . . . . . . . . . . . . . . . . . . 144
isolated data acquisition system . . . . . . . . . . . . . . . . . 147
EMC considerations . . . . . . . . . . . . . . . . . . . . . . . . 147
DC-DC Converters 2-127

Contents
power supply considerations . . . . . . . . . . . . . . . . . . . . 148
interpretation of DC-DC converter EMC data . . . . . . . . . . . . 148
conducted and radiated emissions . . . . . . . . . . . . . . . . . 150
line impedance stabilisation network (LISN) . . . . . . . . . . . . 151
shielding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
line spectra of DC-DC converters . . . . . . . . . . . . . . . . . . 151
temperature performance of DC-DC converters . . . . . . . . . . 154
transfer moulded surface mount DC-DC converters . . . . . . . . 155
custom DC-DC Converters . . . . . . . . . . . . . . . . . . . . . 159
2-128
Terminology
terminology centages. For example, assume a 12V input,

5V output device exhibited a 0.5V change at
The data sheet specification for DC-DC con- the output for a 1.2V change at the input,
verters contains a large quantity of information. line regulation would be 1 %/%.
This terminology is aimed at ensuring the user
is interpreting the data provided correctly and
obtaining the necessary information for their
output voltage accuracy :
circuit application. The proximity of the output voltage to the
specified nominal value. This is given as a
absolute maximum ratings : tolerance envelope for unregulated devices
with the nominal input voltage applied. For
The absolute maximum ratings are the limits example, a 5V specified output device at
2 - APPLICATIONS
to which the devices can be stressed without 100% load may exhibit a measured output
causing permanent and irreparable damage. voltage of 4.75V, i.e. a voltage accuracy of
These limits are not the normal operating or -5%).
functional limits of the devices and operating
at the absolute maximum ratings will pro-
duce different parametric results to those
input and output ripple :
quoted in the data sheet. The amount of voltage droop at the input or
output between switching cycles. The value
input voltage range : of voltage ripple is a measure of the storage
ability of the filter capacitors.
The range of input voltage that the device can
tolerate and maintain functional performance.
input to output isolation :
load voltage regulation : The dielectric breakdown strength test be-
tween input and output circuits. This is the
The change in output voltage over the isolation voltage the device is capable of
specified change in output load. Usually withstanding for a specified time, usually 1
specified as a percentage of the nominal second. (NMV series is 1 minute)
output voltage, for example if a 1V change
in output voltage is measured on a 12V
output device, load voltage regulation is
insulation resistance :
8.3%. For unregulated devices the load The resistance between input and output
voltage regulation is specified over the load circuits. This is usually measured at 500V DC
range 10% to 100% of full load.
efficiency at full load :
line voltage regulation : The ratio of power delivered from the device
The change in output voltage for a given to power supplied to the device when the part
change in input voltage, expressed as per- is operating under 100% load conditions.

Terminology
temperature drift : mean time to failure (MTTF) :

The change in voltage, expressed as a These figures are calculated expected device
percentage of the nominal, per degree lifetime figures using the hybrid circuit model
change in ambient temperature. This pa- of MIL-HDBK-217F. The hybrid model has
rameter is related to several other temperature various accelerating factors for operating
dependent parameters, mainly internal com- environment ( E), maturity ( L), screening
ponent drift. ( Q), hybrid function ( F) and a summation
of each individual component characteristic
temperature above ambient : ( C). The equation for the hybrid model is
then given by;
The temperature rise developed by the device
under full load conditions. This is related to
efficiency. = ∑ (NC C) (1+0.2 E) L F Q
(failures in 106 hours)

switching frequency :
The nominal frequency of operation of the The MTTF figure is the reciprocal of this
switching circuit inside the DC-DC converter. value.
The ripple observed on the input and output
pins is usually twice the switching frequency In the data book all figures for MTTF are
due to full wave rectification and the push- given for the ground benign (GB) environ-
pull configuration of the driver circuit. ment ( E=0.5), this is considered the most
appropriate for the majority of applications
in which these devices are likely to be
no load power consumption :
designed in. However, this is not the only
This is a measure of the switching circuits operating environment these devices can be
requirement to function, it is determined with used for and hence those users wishing to
zero output load and is a limiting factor for incorporate these devices into a more severe
the total efficiency of the device. environment can calculate the predicted
MTTF from the following data.
isolation capacitance : The MIL-HDBK-217F has military environ-
The input to output coupling capacitance. ments specified, hence some interpretation of
This is not actually a capacitor, but the parasitic these is required to apply them to standard
capacitive coupling between the transformer commercial environments. Table 1 gives ap-
primary and secondary windings. Isolation proximate cross references from MIL-HDBK-
capacitance is typically measured at 1MHz 217F descriptions to close commercial
to reduce the possibility of the on-board filter equivalents. Please note that these are not
capacitors affecting the results. implied by MIL-HDBK-217F but are our
interpretation, also we have reduced the
number of environments from 14 to 6
which are most appropriate to commercial
2-130
Terminology
E MIL-HDBK-271F Commercial Interpretation or

Environment
Symbol Description Examples
Non-mobile, temperature and Laboratory equipment, test

Ground
GB humidity controlled environments instruments, desktop PC’s,
Benigh
readily accessible to maintenance static telecomms
Equipment installed in wheeled or In-vehicle instrumentation,

Ground
GM tracked vehicles and equipment mobile radio and telecomms,
Mobile
manually transported portable PC’s
Sheltered or below deck

Naval Navigation, radio equipment and
NS equipment on surface ships or
Sheltered instrumentation below deck
2 - APPLICATIONS
submarines
Pressurised cabin compartments

Aircraft Typical conditions in cargo
and cock-pit, in flight
Inhabited AIC compartments which can be
entertainment and non-safety
Cargo occupied by aircrew
critical applications
Earth orbital. Vehicle in neither Orbital communications satellite,

Space
SF powered flight nor in atmospheric equipment only operated once
Fight
re-entry in-situ
Severe vibrational shock and very

Missile Severe conditions relating to
ML high accelerating forces, satellite
Launch missile launch
lauch conditions
Table 1 : Interpretation of Environmental Factors
applications. For a more detailed under- factor is the only variable;

standing of the environments quoted and the
hybrid model it is recommended that a full
= k (1 + 0.2 E)
copy of MIL-HDBK-217F is obtained.
It is interesting to note that space flight and
The MTTF values for other environment
ground benign have the same environment
factors can therefore be calculated from the
factors. It could be suggested that the act of
ground benign figure quoted at each tem-
achieving space flight should be the deter-
perature point in the data book. Hence pre-
mining environmental factor (i.e. missile
dicted MTTF figures for other environments
launch).
can be calculated very quickly. All the values
The hybrid model equation can therefore be will in general be lower and since the majority
rewritten for any given hybrid, at a fixed of the mobile environments have the same
temperature, so that the environmental factor a quick divisor can be calculated for

Isolation
each condition. Therefore the only calcula- range once the converter is started.
tion necessary is to divide the quoted MTTF
There are almost limitless applications for
figure by the divisor given in table 2.
DC-DC converters since these components
provide the basic power supply function on
E E
many circuit boards. These application
Environment Divisor notes will concentrate on using the converters
Symbol Value
in some specific applications as well as high-
Ground Benign GB 0.5 1.00 lighting good design practice for the circuit
Ground Mobile GM 4.0 1.64 around the DC-DC converter. Where detailed
designs are given, they are for illustrative
Naval Sheltered NS 4.0 1.64 purposes to demonstrate a principle, even
Aircraft Inhabited though the circuit has been built to solve a
AIC 4.0 1.64 particular circuit or design problem.
Cargo
Space Flight SF 0.5 1.00

isolation
Missile Launch ML 12.0 3.09
One of the main features of the majority of
Table 2 : Environment Factors Newport Components DC-DC converters is
their high galvanic isolation capability. This
allows several variations on circuit topography
noise : by using a single DC-DC converter.
Input conducted noise is given in the line The basic input to output isolation can be
conducted spectra for each DC-DC con- used to provide either a simple isolated
verter (see EMC issues for further details). output power source, or to generate different
Noise is affected significantly by pcb layout, voltage rails and/or dual polarity rails (see
measurement system configuration, termi- figure 1).
nating impedance etc. and is difficult to
quote reliably and with any accuracy other These configurations are most often found
than via a spectrum analysis type plot. There in instrumentation, data processing and
will be some switching noise present on top other noise sensitive circuits where it is
of the ripple, however, most of this is easily necessary to isolate the load and noise
reduced by use of small capacitors or filter presented to the local power supply rails
inductors as shown in the application notes. from that of the entire system. Usually local
supply noise appears as common mode
noise at the converter and does not pollute
temperature derating curve : the main system power supply rails.
The component will operate over a wider
The isolated positive output can be con-
temperature range if less power is drawn
nected to the input ground rail to generate
from the output and the device is already
a negative supply rail if required. Since the
running. The temperature derating curve
output is isolated from the input the choice of
shows the operating power-temperature
2-132
Connecting DC-DC Converters in Series
a/ Single Output (NME, LME, NML etc) a/ Non-Isolated Dual Rails
VCC +VO
VIN
DC VO DC 0V
DC -VO
GND DC 0V
GND
b/ Non-Isolated Negative Rail

b/ Dual Output (NMA, NMH, NMV etc)
VCC
+VO
VIN +VO DC
DC 0V
0V DC
-VO
2 - APPLICATIONS
GND DC -VO GND
c/ Dual Isolated Outputs (NMD)

c/ Twinl Isolated Outputs (NMD) VCC
+VO
VO1
DC (VO+VIN)
VIN
DC 0V1 DC 0V
V O2 GND
GND DC
0V2
Figure 1 : Standard Isolated Configurations Figure 2 : Alternative Supply Configurations
reference for the output side can be relatively connecting DC-DC converters in
arbitrary, for example an additional single series
rail can be generated above the main supply
rail or offset by some other DC value (see Galvanic isolation of the output allows
figure 2). multiple converters to be connected in
series simply by connecting the positive
Regulated converters need more consideration
output of one converter to the negative of
than the unregulated types for mixing the
another (see figure 3). In this way non-
reference level. Essentially the single supply
standard voltage rails can be generated,
rail has a regulator in its +VO rail only,
however, the current output of the highest
hence referencing the isolated ground will
output voltage converter should not be ex-
only work if all the current return is through
ceeded.
the DC-DC and not via other external com-
ponents (e.g. diode bias, resistor feed). When converters are connected in series,
Having an alternative return path can upset additional filtering is strongly recommended
the regulation and the performance of the as the converters switching circuits are not
system may not equal that of the converter. synchronised. As well as a summation of
the ripple voltages, the output could also

Connecting DC-DC Converters in Parallel
VCC VCC
+VO +VO
DC DC
0V
DC DC
2VO
+VO
DC DC
DC
0V DC
0V
GND GND
Figure 3 : Connecting DC-DC Converters in Figure 4 : Diode Coupled Paralleled DC-DC

Series Converters
produce relatively large beat frequencies. A When connecting converter outputs it should
capacitor across the output will help, as will be remembered that the switching will not
a series inductor (see filtering). be synchronous, hence some form of cou-
pling should be employed. One possible
connecting DC-DC converters in solution is to use a diode feed, this is suitable
mainly for 12V and 15V output types only
parallel where the diode voltage drop (typically 0.6V)
If the available power output from a single will not significantly affect the circuit func-
converter is inadequate for the application tionality (see figure 4). With 5V and 9V
then multiple converters can be paralleled supplies the diode drop is generally too
to produce a higher output power. large to consider as a suitable means of
connecting paralleled converters.
It should be noted that it is always preferable
to parallel multiple converters of the same This method also has a beat frequency that
type. For example if a 2.5W converter is will superimpose itself over the ripple of the
required, then either 2 NMHs should be two converters, this can be reduced by using
used or 3 NMAs, not an NMH and NMA. an external capacitor at the paralleled output.
The reason for this is that the output voltages The preferred method of connecting converters
are not sufficiently well matched to guarantee in parallel is via series inductors on the
that an NMH would supply twice as much as output (see figure 5). This configuration not
an NMA and the situation could occur where only has a lower loss of voltage than the
there was only 1W being drawn from the diode method, but by suitable choice of
NMH and 1.5W from the NMA. Even with inductor and an additional external capaci-
parallel converters of the same type, loading tor, the beat frequency can be significantly
will be uneven, however, there is only likely reduced, as will the ripple from each con-
to be around a 10% difference in output load verter. Suitable values are given in table 3,
when the output voltages are well matched. these typically reduce the beat and ripple
frequencies by a factor of 10.
2-134
Recommended Values for Paralleled DC-DC Converters
be done when essential and a higher power

Product Output Inductance
Range Voltage ( H) single converter is always a preferable
solution. There should always be a correc-
LME 3.3 33 tion factor of the maximum power rating to
NMA allow for mismatch between converters and
NMD 5 47
a select at full load test is recommended to
NME 9 100 ensure the output voltage is matched to
NMH
NML within 1% or 2%. In general a factor of 0.9
12 220
NMV should be used to provide a power safety
15 330 margin per converter (e.g. 2 NMH convert-
ers paralleled should only be used up to a
power level of 3.6W, not their 4W maxi-
2 - APPLICATIONS
Table 3 : Parallel Output Inductors
mum). At most three DC-DC converters can
be paralleled with a high level of confidence
recommended values for paralleled in the overall performance. If the circuit needs
more power than three converters in parallel,
DC-DC Converters then a single converter with a much higher
The capacitance value used (COUT) should power rating should be considered.
be approximately 1 F per parallel channel Regulated output DC-DC converters should
(i.e. for 2 parallel single output converters, not be paralleled since their output voltage
2 F between the common positive output would need to be very accurately matched
and 0V). to ensure even loading (to within the tolerance
The same comments can be applied to the of the internal linear regulator). Paralleling
input circuit for converters whose inputs are regulated converters could cause one of the
paralleled and similar values for inductance parts to be overloaded. If a high power
and input capacitance should be used as regulated supply is required, it would be
shown above. better to parallel unregulated converters and
add an external linear regulator.
In general paralleling converters should only
VCC
+VO
LIN
DC
LOUT
DC
Cin COUT
DC LOUT
LIN
DC 0V
GND
Figure 5 : Fully Filtered Paralleled DC-DC Converters

Filtering
filtering The DC resistance of the inductor is the final

consideration that will give an indication of
All Newport Components isolated DC-DC the DC power loss to be expected from the
converters have a fixed characteristic inductor.
frequency at which the device operates.
This fixed frequency allows filtering that is The value of inductor and capacitor to use
relatively simple compared to pulse-skipping is given in the table below for the majority of
types. In a pulse skipping converter a large Newport Components DC-DC converters.
range of frequencies are encountered as the The capacitance is chosen to form a pi filter
device adjusts the pulse interval for loading to match the input or output capacitor of the
conditions. DC-DC converter. The inductor is chosen to
cause heavy attenuation of the characteristic
When reducing the ripple from the converter, frequency when combined with the given
at either the input or the output, there are capacitors.
several aspects to be considered. Newport
Components recommend filtering using
simple passive LC networks at both input
recommended values for filtered
and output (see figure 6). A passive RC DC-DC converters
network could be used, however, the power
Those converters that feature an internal linear
loss through a resistor is considered too
regulator (eg NMF, NMXSO) do not require
high. The self resonant frequency of the
external filtering at the output but can benefit
inductor needs to be significantly higher
from the above filter combinations at the input
than the characteristic frequency of the
(see table 4).
DC-DC (typically 100kHz for Newport
Components DC-DC converters). The DC
current rating of the inductor also needs
consideration, a rating of approximately
twice the supply current is recommended.
LIN LOUT
VIN +VO
DC CO
CIN 0V
CO
GND
DC -VO
LOUT
Figure 6 : Input and Output Filtering
2-136
Limiting Inrush Current
Product Input Inductance/ Output Inductance/

Range Voltage Capacitance Voltage Capacitance
3.3 33 H / 1.5 F 3.3 33 H / 1.5 F

LME
NMA 5 47 H / 1.0 F 5 47 H / 1.0 F
NME
NMH 12 220 H / 1.0 F 9 100 H / 1.0 F
NMD
NMV 24 470 H / 470nF 12 220 H / 1.0 F
48 680 H / 180nF 15 330 H / 1.0 F
Table 4 : Recommended Values for Filtered DC-DC’s
2 - APPLICATIONS
limiting inrush current V
i=
R
Using a series inductor at the input will limit
 − t 
the current that can be seen at switch on (see V o l t a g e : V = Vin ( i − e x p   )
 RC 
figure 7). If we consider the circuit without VIN
the series inductor, then the input current is

given by ;
V  −t 
Current : i = exp  
V  − t R  RC 
i= exp  
R  RC 
time
When the component is initially switched on
(i.e. t=0) this simplifies to ;
Figure 7 : Input Current & Voltage at Switch
V On
i=
R
maximum output capacitance
This would imply that for a 5V input, with say A simple method of reducing the output
50m track and wire resistance, the inrush ripple is simply to add a large external
current could be as large as 100A, this could capacitor. This can be a low cost alternative
cause a problem for the DC-DC converter. to the LC filter approach, although not as
A series input inductor therefore not only effective. There is also the possibility of
filters the noise from the internal switching causing start up problems if the output
circuit, but also limits the inrush current at capacitance is too large.
switch on. With a large output capacitance, at switch
on there is no charge on the capacitors and

Settling Time
the DC-DC converter immediately experi- of the coupling between input and output
ences a large current demand at its output. circuits. Providing this is the largest coupling
The inrush current can be so large as to source, a calculation of the leakage current
exceed the ability of the DC-DC converter between input and output circuits can be
and the device can go into an undefined estimated.
mode of operation. In the worst case scenario
Assuming we have a known isolation capaci-
the device can give a lower than expected
tance (CIS - refer to DC-DC converter data)
DC output with a very high ripple. The
and a known frequency for either the noise or
DC-DC converter may survive this condition,
test signal, then the expected leakage current
however, the circuit being supplied is unlikely
(IL) between input and output circuits can be
to function under this supply scheme.
calculated from the impedance.
Newport Components recommend a maxi-
The general isolation impedance equation
mum safe operating value of 10 F for the
for a given frequency (f) is given by ;
output per channel. When used in conjunction
with a series output inductor this value can
be raised to 47 F should extremely low 1
Zf =
j 2π f Ci s
ripple be required.
settling time For an NMA0505S, the isolation capaci-

tance is 18pF, hence the isolation impedance
The main reason for not fitting a series to a 50Hz test signal is ;
inductor internally is that many applica-
tions require a fast power on time (there is 1
Z5 0 = = 177 M Ω
also a size constraint with our miniature j 2π 50 18 pF
parts). When the power on voltage is a
controlled fast ramp, then the output can If using a test voltage of 1kVrms, the leakage
respond within 500 s of the input reaching current is ;
its target voltage (measured on a range of
NMA and NMH components under full out-
Vt e s t 1000V
put load without external filters). The use of iL = = = 5.65 µ A
Zf 177M Ω
external filters and additional input or output
capacitance will slow this reaction time. It is
therefore left to the designer to decide on the It can be easily observed from these simple
predominant factors affecting their circuit; equations that the higher the test or noise
settling time or noise performance. voltage, the larger the leakage current, also
the lower the isolation capacitance the lower
the leakage current. Hence for low leakage
isolation capcitance and leakage
current, high noise immunity designs, high
current isolation DC-DC converters should be
The isolation barrier within the DC-DC con- selected with an appropriate low isolation
verter has a capacitance which is a measure capacitance.
2-138
Overload Protection
overload protection If there is an intelligent power management

system at the input, using a series resistor
Although the use of filtering will prevent (in place of the series inductor) and detecting
excessive current at power-on, under normal the voltage drop across the device to signal
operating conditions, there is no protection the management system can be used. A
against an output circuit taking excessive similar scheme can be used at the output to
power or even going short-circuit. When determine the output voltage, however, if
this happens the DC-DC converter will take the management system is on the input
a large input current to try to supply the side, the signal will need to be isolated from
output, eventually the converter will overheat the controller to preserve the system isolation
and destroy itself if this condition is not barrier (see figure 10).
rectified (short circuit overload is only guar-
2 - APPLICATIONS
anteed for 1s on an unregulated part). The thermal dissipation in a series resistor
on the output can also be used to determine
There are several ways to prevent overload at overloading and preserve the isolation barrier.
the outputs destroying the DC-DC converter, If a thermistor or other thermally sensitive
the simplest being a straight forward fuse, device is mounted close to the resistor this
sufficient tolerance for inrush current is can be used to indicate an overload condition.
required to ensure the fuse does not blow on System temperature will also need to be
power-on (see figure 8). Another simple known to provide a suitable offset for different
scheme that can be applied is a circuit breaker. operating environments.
There is also the potential to add some There are several other current limiting tech-
intelligence to the overload scheme by niques that can be used to detect an over-
either detecting the input current, or the load situation, the suitability of these is left
output voltage (see figure 9). The simplest to the designer. The most important thing to
implementation for overload protection at consider is how this information will be used.
the input is to have the device supplied via If the system needs to signal to a controller
a linear regulator with an internal thermal the location or module causing the overload
shutdown facility. This does however reduce some form of intelligence will be needed. If
the overall efficiency significantly. the device simply needs to switch off, a
VIN
DC
FUSE
DC
GND
Figure 8 : Simple Overload Protection

Overload Protection
VCC
REG
DC
DC
GND
a/ Linear Regulator with Internal Thermal Shutdown
VCC RIN
DC
VOL DC
GND
b/ Series Resistor for Input Current Measurement
VCC
ILIMIT
DC
R1 DC
R2 RGND
GND
c/ Ground Current Monitor

Choose current limit (ILIMIT) and ground resistor (RGND) so that : 0.7V = RGND x ILIMIT.
Figure 9 : Input Monitored Overload Protection
2-140
Overload Protection
VCC
+VO
DC
DC 0V
GND
RD
RO Opto-Isolator
2 - APPLICATIONS
VOL
a/ Opto-Isolated Overload Detector

On overload +VO falls and the LED switches off, the VOL. line is then pulled high.
VOL
NTC Thermistor
VCC
DC VO
RO
DC 0V
GND
b/ Thermal Overload Detector
Figure 10 : Ouput Monitored Overload Protection

Input Voltage Drop-Out
simple fuse type arrangement will be adequate. source for the converter, the diode blocking
other circuits from draining the capacitor
All Newport Components DC-DC converters
over the supply rail. When combined with
which include an internal linear regulator
an in line inductor this can also be used to
have a thermal overload shut-down condition
give very good filtering. The diode volt drop
which protects these devices from excessive
needs to be considered in the power supply
overload. If this condition is to be used to
line under normal supply conditions, a low
inform a power management system the
drop Schottky diode is recommended (see
most suitable arrangement is the output
figure 11).
voltage detector (see figure 10a), since this
will fall to near zero on shut-down. A thermal
probe on the case of the DC-DC converter no load over voltage lock-out
is also a possible solution.
Unregulated DC-DC converters are ex-
pected to be under a minimum of 10% load,
input voltage drop-out (brown-outs) hence below this load level the output voltage
is undefined. In certain circuits this could be
When the input voltage drops, or is momen-
a potential problem.
tarily removed, the output circuit would suffer
similar voltage drops. For short period input The easiest way to ensure the output voltage
voltage drops, such as when other connected remains within a specified tolerance is to
circuits have an instantaneous current de- add external resistors so that there is always
mand, or devices are plugged in or removed a 10% loading on the device (see figure 12).
from the supply rail while ’hot’, a simple This is rather inefficient in that 10% of the
diode-capacitor arrangement can prevent power is always being taken by this load,
the output circuit from being effected. hence only 90% is available to the additional
circuitry.
The circuit uses a diode feed to a large
reservoir capacitor (typically 47uF electrolytic) Zener diodes on the output are another
which provides a short term reserve current simple method. It is recommended that
ZDX60
LIN
DC OUTPUT
47 F
CIRCUIT
DC
Figure 11 : Input Voltage Drop-out
2-142
Long Distance Supply Lines
R10%
DC
DC
R10%
2 - APPLICATIONS
R2
DC
DC
Figure 12 : No Load Over Voltage Lock-out
these be used with a series resistor or inductor long distance supply lines
as when the Zener action occurs a large
current surge may induce signal noise into When the supply is transmitted over a cable
the system. there are several reasons why using an
isolated DC-DC converter is good design
practice (see figure 13). The noise pick up
and EMC susceptibility of a cable is high
NMA0512D NME2405
CABLE
VIN
DC DC TARGET
CIRCUIT
GND DC DC
Figure 13 : Long Distance Power Transfer

LCD Display Bias
compared to a pcb track, by isolating the EIA-232 interface

cable via a DC-DC converter at either end,
any cable pick-up will appear as common In a mains powered PC often several supply
mode noise and should be self cancelling at rails are available to power an RS232 interface.
the converters. However, battery operated PC’S or remote
equipment having an RS232 interface
Another reason is to reduce the cable loss by added later, or as an option, may not have
using a high voltage, low current power the supply rails to power an RS232 interface.
transfer through the cable and re-converting Using an NMA0512 is a simple single chip
at the terminating circuit. This will also solution, allowing a fully EIA-232 compatible
reduce noise and EMC susceptibility since interface to be implemented from a single
the noise voltage required to affect the rail 5V supply rail and only 2 additional components
is also raised. (see figure 15).
For example, compare a system having a 5V
supply and requiring a 5V, 500mW output 3V/5V logic mixed supply rails
at a remote circuit. Assume the connecting
cable has a 100 resistance. Using an There has been a lot of attention given to
NME0505D to convert the power at either new I.C.’s and logic functions operating at
end of the cable, with a 100mA current the what is rapidly emerging as the standard
cable will lose 1W (I2R) of power, the NME supply level for notebook and palmtop com-
would not be suitable since this is its total puters. The 3.3V supply is also gaining
power delivery, hence there is no power rapid acceptance as the defacto standard for
available for the terminating circuit. Using personal telecommunications, however, not
an NMA0512D to generate 24V and an all circuit functions required are currently
NME2405D to regenerate 5V, only a 21mA available in a 3.3V powered IC. The system
supply is required through the cable, a cable designer therefore has previously had only
loss of 44mW. two options available; use standard 5V logic
or wait until the required parts are available
in a 3V form, neither being entirely satisfac-
LCD display bias tory and the latter possibly resulting in lost
An LCD display typically requires a positive market share.
or negative 24V supply to bias the crystal. There is now another option, mixed logic
The NME0524 (custom) converter was de- functions running from separate supply rails.
signed specifically for this application. A single 3.3V line can be combined with a
Having an isolated 0V output this device can range of DC-DC converters from Newport
be configured as a +24V supply by connecting Components to generate voltage levels to
this to the GND input or a -24V supply by run virtually any standard logic or interface
connecting the +VO output to GND (see figure IC. The Newport Components range in-
14). cludes dual output parts for powering ana-
logue bipolar and amplifier functions (NMA
series) as well a single output function for
2-144
3V/5V Logic Mixed Supply Rails
NME0524
DC
DC -24V
(Upto 42mA)
Figure 14 : LCD Display Bias
2 - APPLICATIONS
+12
EIA-232
PORT 5V VCC
VDD
DCD
DB9S CONNECTOR
DSR NMA0512
RX +VO
DC
RTS 0V
TX
-VO DC
CTS
DTR
RI GND
SN75C185
Figure 15 : Optimised RS232 Interface
localised logic functions (LME, NME series). Newport Components has another variation
A typical example might be an RS232 interface on this theme and has developed two 5V to
circuit in a laptop PC using a 3.3V interface 3.3V step down DC-DC converters
chip (such as the LT1330) which accepts (LME0503 and NME0503). These have
3.3V logic signals but requires a 5V supply been designed to allow existing systems to
(see figure 16). start incorporating available 3.3V I.C.’s
without having to redesign their power supply.

3V/5V Logic Mixed Supply Rails
3.3V / V CC
LME0305D
3 8 +5V
DC
1 DC 7 OV
GND
GND VCC
1 F 100nF
V+ V-
1 2 28
3 26
3.3V + +
220nF 200nF
4 27
14
TX1 25 5 TX1
3.3V LOGIC RS232

RX1 24 6 RX1
17
LT1330
GND
Figure 16 : RS232 Interface with 3V Logic
This is particularly important when trying to voltage than is available from the supply
reduce the overall power demand of a system, rail, or for a single localised function. Using
but not having available all of the functions a fully isolated supply is particularly useful in
at the 3.3V supply. interface functions and systems maintaining
separate analogue and digital ground lines.
The main application for this range of devices
are system designers who want to provide
some functionality that requires a higher
2-146
Isolated Data Acquisition System
NMV0505
5V LOGIC CIRCUIT 5V 5V +5V
47 H
OPTO ISOLATORS DC
4K7 1K2 1 F
DATA DC
DATA
+5V
1k2 +5V VCC
CS
4K7 1K2
5V
Vref
CS
4K7 1K2
ZN509 470 F
2 - APPLICATIONS
AIN
STATUS STATUS
+5V CLK GND
1K2 4K7
CLK
SFH610
Figure 17 : Isolated Serial ADC System
isolated data acquisition system EMC considerations

Any active system requiring isolation will When used for isolating a local power supply
need a DC-DC converter to provide the and incorporating the appropriate filter cir-
power transfer for the isolated circuit. In a cuits as illustrated above, DC-DC converters
data acquisition circuit there is also the need can present simple elegant solutions to
for low noise on the supply line, hence good many EMC power supply problems. The
filtering is required. range of fixed frequency DC-DC converters
is particularly suitable for use in EMC prob-
The circuit shown (see figure 17) provides a
lem situations as the stable, fixed switching
very high isolation barrier by using an NMV
frequency gives easily characterised and
converter to provide the power isolation and
easily filtered output.
SFH610 opto-isolators for the data isolation.
An overall system isolation of 2.5kV is The following notes give suggestions to
achieved. avoiding common EMC problems in power
supply circuits. A more extensive discussion
on other aspects of EMC is available in the
Newport Components EMC Design Guide-
lines book.

Power Supply Considerations
power supply considerations interpretation of DC-DC converter

eliminate loops in supply lines
EMC data
(see figure 18). Electromagnetic compatibility (EMC) of elec-
decouple supply lines at local boundaries trical and electronic products is a measure
(use RCL filters with low Q, see figure 19). of electrical pollution. Throughout the world
there are increasing statutory and regulatory
place high speed sections close to the requirements to demonstrate the EMC of
power line input, slowest section furthest end products. In Europe the EC directive
away (reduces power plane transients, 89/336/EEC requires that any product sold
see figure 20). after 1 January 1996 complies with a series
isolate individual systems where possible of EMC limits, otherwise the product will be
(especially analogue and digital systems) prohibited from sale within the EEC and the
on both power supply and signal lines seller could be prosecuted and fined.
(see figure 21). Although DC-DC converters are generally
An isolated DC-DC converter can provide a exempt from EMC regulations on the
significant benefit to reducing susceptibility grounds that these are component items, it
and conducted emission due to isolating both is the belief of Newport Components Ltd.
power rail and ground from the system supply. that the information on the EMC of these
The range of DC-DC converters available components can help designers ensure their
from Newport Components all utilise toroidal end product can meet the relevant statutory
power transformers and as such have negli- EMC requirements. It must be remembered
gible EMI radiation (they also incorporate the however, that the DC-DC converter is un-
recommended pcb layout suggestions as likely to be the last component in the chain
stated in Newport Components EMC Guide- to the mains supply, hence the information
lines Data book). quoted needs interpretation by the circuit
designer to determine its impact on the final
Isolated DC-DC converters are switching EMC of their system.
devices and as such have a characteristic
switching frequency which may need some The notes given here are aimed at helping
additional filtering. Some commercial con- the designer interpret the effect the DC-DC
verters offer a pulse skipping technique, converter will have on the EMC of their end
which although offering a flat efficiency product by describing the methods and ra-
response gives a very wide spectral range tionale for the measurements made. Where
of noise, since it does not have a fixed possible CISPR and EN standards have been
characteristic frequency. Newport Compo- used to determine the noise spectra of the
nents devices feature a fixed frequency con- components, however, all of the standards
verter stage which is stable across its full reference mains powered equipment and
loading and temperature curve, hence it is interpretation of these specifications is nec-
very easy to filter the switching noise using essary to examine DC supplied devices.
a single series inductor.
2-148
Power Supply Considerations
VCC
PSU CCT1 CCT2
GND
VCC
2 - APPLICATIONS
PSU CCT1 CCT2
GND
Figure 18 : Eliminate Loops in Supply Line
VCC
CCT1 CCT2
GND
Figure 19 : Decouple Supply Lines at Local Boundaries

Conducted and Radiated Emissions
LOCAL
LOW SPEED
POWER
P HIGH MEDIUM CIRCUIT
INPUT
S SPEED SPEED
U CIRCUIT CIRCUIT
DC CIRCUIT
FILTER
Figure 20 : Place High Speed Circuit Close to PSU
VCC
DC DC
CCT1 CCT2
DC DC
GND
Figure 21 : Isolate Individual Systems
conducted and radiated emissions are those transmitted over wire connecting
circuits together and covers the frequency
There are basically two types of emissions spectrum 150kHz to 30MHz. Radiated are
covered by the EC directive on EMC, radi- those emissions transmitted via electromag-
ated and conducted. Conducted emissions netic waves in air and cover the frequency
2-150
Line Impedance Stabilisation Network (LISN)
spectrum 30MHz to 1GHz. Hence the EC supply to the DUT. The same approach has
directive covers the frequency spectrum been used in the testing of DC-DC converters
150kHz to 1GHz, but as two separate and and the DC supply to the converter was
distinct modes of transmission. filtered to ensure that no noise from the PSU
as present at the measuring instrument.
Newport Components range of DC-DC con-
verters feature toroidal transformers within the A line impedance stabilisation network
component. These have been tested and (LISN) conforming to CISPR 16 specification
proved to have negligible radiated noise. The is connected to both positive and negative
low radiated noise is primarily due to toroidal supply rails and referenced to mains earth
shaped transformers maintaining the mag- (see figure 22). The measurements are all
netic flux within the core, hence no magnetic taken from the positive supply rail with the
2 - APPLICATIONS
flux is radiated by design. Due to the excep- negative rail measurement point terminated
tionally low value of radiated emission only with 50W to impedance match the measure-
conducted emissions are quoted. ment channels.
Conducted emissions are measured on the
input DC supply line. Unfortunately no shielding
standards exist for DC supplies as most
At all times the DUT, LISN’s and all cables
standards cover mains connected equipment.
connecting any measurement equipment,
This poses two problems for a DC supplied
loads and supply lines are shielded. The
device, firstly no standard limit lines can be
shielding is to prevent possible pick-up on
directly applied, since the DC supplied
cables and DUT from external EMC sources
device does not directly connect to the
(e.g. other equipment close by). The shielding
mains, also all reference material uses the
is referenced to mains earth (see figure 22).
earth-ground plane a reference point. In a
DC system often the 0V is the reference,
however, for EMC purposes, it is probably line spectra of DC-DC converters
more effective to maintain the earth as the
All DC-DC converters are switching devices,
reference, since this is likely to be the reference
hence, will have a frequency spectra. Fixed
that the shielding or casing is connected to.
input DC-DC converters have fixed switching
Consequently all measurements quoted are
frequency, for example the NMH range of
referenced to the mains borne earth.
converters has a typical switching frequency
of 75kHz. This gives a stable and predictable
line impedance stabilisation network noise spectrum regardless of load condi-
(LISN) tions.
It is necessary to ensure that any measure- If we examine the noise spectrum closely
ment of noise is from the device under test (see figure 23) we can see several distinct
(DUT) and not from the supply to this device. peaks, these arise from the fundamental
In mains connected circuits this is important switching frequency and its harmonics (odd
and the mains has to be filtered prior to labelled line spectra) and the full rectified

Line Spectra of DC-DC Converters
POWER SUPPLY
50 Termination
-- LISN
LOAD
DC
DC
+ LISN
To Spectrum Analyser
Figure 22 : Filtered Supply to DC-DC Converter
spectra at twice the fundamental switching The EC regulations for conducted interfer-
frequency (even labelled line spectra). ence covers the bandwidth 150kHz to
Quasi-resonant converters, such as the 30MHz, considering a converter with a
Newport range, have square wave switching 100kHz nominal switching frequency, this
waveforms, this produces lower ripple and would exhibit 299 individual line spectra.
a higher efficiency than soft switching There will also be a variation of absolute
devices, but has the drawback of having a switching frequency with production vari-
relatively large spectrum of harmonics. ation, hence a part with a 90kHz nominal
frequency would have an additional 33 lines
Figure 23 : Individual Line Spectra
2-152
Line Spectra of DC-DC Converters
over the entire 30MHz bandwidth. Absolute ing for variances in switching frequency due
input voltage also produces slight variation to production variation or differences in
of switching frequency (see figure 24). absolute input voltage (see figure 25).
Hence, to give a general level of conducted
The conducted emissions are measured
noise, we have used a 100kHz resolution
under full load conditions in all cases, under
bandwidth (RBW) to examine the spectra in
lower loads the emission levels do fall, hence
the data sheets. This wide RBW gives a
full load is the worst case condition for
maximum level over all the peaks, rather
conducted line noise.
than the individual line spectra, this is easier
to read as well as automatically compensat-
2 - APPLICATIONS
Figure 24 : Frequency Voltage Dependency
Figure 25 : NMS Spectrum

Temperature Performance of DC-DC Converters
temperature performance of DC-DC rating, their current carrying capacity derates

converters as temperature exceeds 100°C.
Therefore this allows the DC-DC converters
The temperature performance of the DC-DC
to be used above their specified operating
converters detailed in this book is always
temperature, providing the derating of
better than the quoted operating temperature
power delivery given in the specification is
range. The main reason for being conser-
adhered to. Components operating outside
vative on the operating temperature range
the quoted operating temperature range
is the difficulty of accurately specifying
cannot be expected to exhibit the same
parametric performance outside this tem-
parametric performance that is quoted in the
perature range.
specification.
There are some limiting factors which provide
An indication of the stability of a device can
physical barriers to performance, such as
be obtained from the change in its operating
the Curie temperature of the core material
frequency as the temperature is varied (see
used in the DC-DC converter (the lowest
figure 26). A typical value for the frequency
Curie temperature material in use at Newport
variation with temperature is 0.5% per °C, a
Components is 125°C). Ceramic capacitors
very low value compared to other commercial
are used almost exclusively in the DC-DC
parts. This illustrates the ease of filtering of
converters because of their high reliability and
Newport Components DC-DC converters
extended life properties, however, the absolute
since the frequency is so stable across load
capacity of these can fall when the temperature
and temperature ranges.
rises above 85°C (ripple will increase). Other
considerations are the power dissipation
within the active switching components, al-
though these have a very high temperature
Figure 26 : Typical Switching Frequency vs. Temperature
2-154
Transfer Moulded Surface Mount DC-DC Converters
transfer moulded surface mount The leads are tinned with a 60:40 lead-tin
DC-DC converters (Pb:Sn) solder finish. This is a standard lead
finish and compatible with virtually all solder
production guideline application note mixes used in a production environment.
The recent introduction by Newport Compo-
nents Ltd (NCL) of a new and innovative component placement
method of encapsulating hybrid DC-DC The TM ranges are designed to be handled
converters in a transfer moulded (TM) by placement machines in a similar way to
thermoset epoxy plastic has enabled a new standard SOIC packages. The parts are
range of surface mount (SM) DC-DC convert- available either in tubes (sticks) or in reels.
ers to be brought to market which addresses The parts can therefore be placed using
2 - APPLICATIONS
the component placement with SOIC style machines with either vibrational shuttle,
handling. gravity feeders or reeled feeders.
With any new component there are of course The vacuum nozzle for picking and placing
new lessons to be learnt with the mounting the components can be the same as used for
technology. With the new TM range of DC- a standard 14 pin or 18 pin SOIC (typically
DC converters, the lessons are not new as a 5mm diameter nozzle). An increase in
such, but may require different production vacuum pressure may be beneficial due to
techniques in certain applications. the heavier weight of the hybrid compared
to a standard SOIC part (a typical 14 pin
component materials SOIC weighs 0.1gm, the NMETM DC-DC
converter weighs 1.3gm). It is advisable to
The body of the TM product range is a high
consult your machine supplier on choice of
thermally conductive thermoset epoxy
vacuum nozzle if in doubt.
plastic.The advantage of thermoset materials
in this application is that the body does not If placing these components by hand,
deform under post-cure heat cycles (i.e. tweezer on the central body area where
under high temperature reflow conditions). there are no component pins. Tweezering on
Consequently there are no precautions re- the pins can cause bending and the pin
quired to protect the body during reflow. co-planarity could be compromised.
Other manufacturers components using
thermoplastics may deform or require a heat component alignment
shield during the reflow process.
The components can be aligned by either
The lead frame is a copper material, hence optical recognition or tweezering. If using
has a high conductivity and reduces the tweezer alignment it should be ensured that
internal resistance of tracking within the DC- the tweezers are aligning on the component
DC converters. Hybrid designs which use body and not on the pins. The components
film deposition for tracking (or printed inks) themselves are symmetrical in the body,
feature higher losses within the DC-DC con- hence relatively easy to align using either
verter due to their higher resistance. method.

solder pad design In general these parts will exceed the reflow
capability of most IC and passive compo-
The TM range of DC-DC converters are
nents on a PCB and should prove the most
designed on a pin pitch of 1.27mm (0.05")
thermally insensitive component to the re-
with 1mm pad widths and 1.75mm pad
flow conditions.
lengths. This allows pads from one part to
be used within a PCB CAD package for
forming the pad layouts for other NCL TM adhesive requirements
parts. These pads are wider than many If SM components are going to be wave
standard SOIC pad sizes (0.64mm) and soldered (i.e. in a mixed through hole and
CAD packages may not accommodate these SM PCB) or are to be mounted on both sides
pins with a standard SOIC pad pattern. It of a PCB, then it is necessary to use an
should be remembered that these compo- adhesive to fix them to the board prior to
nents are power supply devices and as such reflow. The adhesive prevents the SM parts
need wider pads and thicker component being ’washed off’ in a wave solder, and
leads to minimise resistive losses within the being ’vibrated off’ due to handling on a
interconnects. doubled sided SM board.
Pad patterns for each component are As mentioned previously, the NCL range of
included in the relevant chapter. These SM DC-DC converters are heavier than
should be followed where appropriate. standard SOIC devices. The heavier weight
One of the benefits of the NCL approach is is a due to the size (volume) and internal
that PCB layout can be produced for dual hybrid construction. Consequently the parts
component usage. For example the NMATM place a larger than usual stress on their
dual output DC-DC converter pad layout solder joints and leads if these are the only
can accommodate the NMETM product to method of attachment. Using an adhesive
give a single positive output voltage only, between component body and PCB can re-
without any PCB tracking changes. duce this stress considerably. If the final
system is to be subjected to shock and vibra-
tion testing then using adhesive attachment
solder reflow profile is essential to ensure the parts pass these
The TM range of components supplied by environmental tests.
NCL are designed to withstand a maximum
The TM range of DC-DC converters from
reflow temperature of 280C in accordance
NCL all have a stand-off beneath the com-
with CECC 00802. If multiple reflow profiles
ponent for the application of adhesive to be
are to be used (i.e. the part is to pass though
placed without interfering with the siting of
several reflow ovens), it is recommended
the component. Method of adhesive dis-
that lower ramp rates be used than the
pensing and curing, plus requirements for
maximum specified in CECC 00802, con-
environmental test and in-service replace-
tinual thermal cycling to this profile could
ment will determine suitability of adhesives
cause material fatigue if more than 5 maxi-
rather than the component itself. However,
mum ramp cycles are used.
having a thermoset plastic body, thermoset
2-156
epoxy adhesive bonding between board and would be the preferred choice, these types
component is the recommended adhesive ’wick’ under the component body and offer
chemistry. a good all round adhesion from a single
dispensed dot.
If the reflow stage is also to be used as a cure
for a heat cure adhesive, then the compo- The patterns shown allow for the process
nent is likely to undergo high horizontal spread of the stand-off on the component,
acceleration and deceleration during the but do not account for the thickness of the
pick and place operation. The adhesive must PCB tracks. If thick PCB tracks are to be used
be sufficiently strong in its uncured (green) a grounded copper strip should be laid
state in order to keep the component accu- beneath the centre of the component (care
rately placed. should be exercised to maintain isolation
2 - APPLICATIONS
barrier limits). The adhesive should not
adhesive placement retard the pins reaching their solder pads
during placement of the part, hence low
The parts are fully compatible with the 3 viscosity adhesive is recommended.
main methods of adhesive dispensing; pin
transfer, printing and dispensing. The The height of the adhesive dot, its viscosity
method of placing adhesive will depend on and slumping properties are critical. The dot
the available processes in the production must be high enough to bridge the gap
line and the reason for using adhesive between board surface and component, but
attachment. For example if the part is on a low enough not to slump and spread, or be
mixed though-hole and SM board, adhesive squeezed by the component and so con-
will have to be placed and cured prior to taminate the solder pads.
reflow. If using a SM only board and heat If wishing to use a greater number of dots of
cure adhesive, the reflow may be used as the smaller diameter (common for pin transfer
cure stage. If requiring adhesive for shock methods) the dot pattern can be changed by
and vibration but using a conformal coat, following a few simple guidelines. As the
then it may be possible to avoid a separate number of dots is doubled their diameter
adhesive all together and the coating pro- should be halved and centres should be at
vides the mechanical restraint on the com- least twice the printed diameter from each
ponent body. other, but the dot height should remain at
Patterns for dispensing or printing adhesive 0.4mm. The printed dot should always be
are given for automatic lines. If dispensing positioned by at least its diameter from the
manually after placement the patterns for nearest edge of the body to the edge of the
UV cure are easily repeated using a manual dot. The number of dots is not important
syringe (even if using heat cure adhesive). provided good contact between adhesive
If dispensing manually dot height and size and body can be guaranteed, but a mini-
are not as important and the adhesive mum of 2 is recommended.
should be applied after the components
have been reflowed. When dispensing after
reflow, a chip underfill formulation adhesive

cleaning
The thermoset plastic encapsulating material
used for the NCL range of surface mount
DC-DC converters is not fully hermetically
sealed. As with all plastic encapsulated
active devices, strongly reactive agents in
hostile environments can attack the material
and the internal parts, hence cleaning is
recommended in inert solutions (e.g. al-
cohol or water based solvents) and at room
temperature in an inert atmospheres (e.g.
air or nitrogen).
A batch or linear aqueous cleaning process
would be the preferred method of cleaning
using a deionised water solution.
2-158
CUSTOM DC-DC CONVERTERS
custom DC-DC converters procedures as standard products.
In addition to the standard ranges shown in Newport Components custom parts are
this data book, Newport Components have used in many applications which are very
the capability to produce custom DC-DC specific to the individual customer, however,
converters designed to your specific require- some typical examples are;
ments. In general, the parts can be rapidly ECL Logic driver
designed using computer based CAD tools Multiple cell battery configurations
to meet any input or output voltage require- Telecommunications line equipment
ments within the ranges of Newport Com- Marine apparatus
ponents standard products (i.e. up to 48V at Automotive electronics
either input or output). Prototype samples LCD display power circuitry
2 - APPLICATIONS
can also be produced in short timescales. Board level instrumentation systems
Custom parts can be designed to your speci- To discuss your custom DC-DC converter
fication, or where the part fits within a standard requirements, please contact Newport
series, the generic series specification can be Components technical support desk or your
used. All custom parts receive the same local distributor.
stringent testing, inspection and quality

DC-DC Converter Applications

Uploaded by

Copyright:

Available Formats

DC-DC Converter Applications

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

DC-DC Converter Applications

Uploaded by

Copyright:

Available Formats

DC-DC CONVERTER APPLICATIONS

connecting DC-DC converters in series . . . . . . . . . . . . . . 133

connecting DC-DC converters in parallel . . . . . . . . . . . . . 134

recommended values for paralleled DC-DC Converters . . . . . 135

limiting inrush current . . . . . . . . . . . . . . . . . . . . . . . 137

maximum output capacitance . . . . . . . . . . . . . . . . . . . 137

settling time . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

isolation capcitance and leakage current . . . . . . . . . . . . . 138

overload protection . . . . . . . . . . . . . . . . . . . . . . . . 139

input voltage drop-out (brown-outs) . . . . . . . . . . . . . . . 142

no load over voltage lock-out . . . . . . . . . . . . . . . . . . . 142

long distance supply lines . . . . . . . . . . . . . . . . . . . . . 143

LCD display bias . . . . . . . . . . . . . . . . . . . . . . . . . . 144

EIA-232 interface . . . . . . . . . . . . . . . . . . . . . . . . . 144

3V/5V logic mixed supply rails . . . . . . . . . . . . . . . . . . 144

isolated data acquisition system . . . . . . . . . . . . . . . . . 147

EMC considerations . . . . . . . . . . . . . . . . . . . . . . . . 147

DC-DC Converters 2-127

power supply considerations . . . . . . . . . . . . . . . . . . . . 148

interpretation of DC-DC converter EMC data . . . . . . . . . . . . 148

conducted and radiated emissions . . . . . . . . . . . . . . . . . 150

line impedance stabilisation network (LISN) . . . . . . . . . . . . 151

line spectra of DC-DC converters . . . . . . . . . . . . . . . . . . 151

temperature performance of DC-DC converters . . . . . . . . . . 154

transfer moulded surface mount DC-DC converters . . . . . . . . 155

custom DC-DC Converters . . . . . . . . . . . . . . . . . . . . . 159

terminology centages. For example, assume a 12V input,

DC-DC Converters 2-129

temperature drift : mean time to failure (MTTF) :

(failures in 106 hours)

E MIL-HDBK-271F Commercial Interpretation or

Non-mobile, temperature and Laboratory equipment, test

Equipment installed in wheeled or In-vehicle instrumentation,

Sheltered or below deck

Pressurised cabin compartments

Earth orbital. Vehicle in neither Orbital communications satellite,

Severe vibrational shock and very

Table 1 : Interpretation of Environmental Factors

applications. For a more detailed under- factor is the only variable;

DC-DC Converters 2-131

Space Flight SF 0.5 1.00

a/ Single Output (NME, LME, NML etc) a/ Non-Isolated Dual Rails

b/ Non-Isolated Negative Rail

c/ Dual Isolated Outputs (NMD)

Figure 1 : Standard Isolated Configurations Figure 2 : Alternative Supply Configurations

DC-DC Converters 2-133

Figure 3 : Connecting DC-DC Converters in Figure 4 : Diode Coupled Paralleled DC-DC

be done when essential and a higher power

Figure 5 : Fully Filtered Paralleled DC-DC Converters

DC-DC Converters 2-135

filtering The DC resistance of the inductor is the final

Figure 6 : Input and Output Filtering

Product Input Inductance/ Output Inductance/

3.3 33 H / 1.5 F 3.3 33 H / 1.5 F

48 680 H / 180nF 15 330 H / 1.0 F

Table 4 : Recommended Values for Filtered DC-DC’s

the series inductor, then the input current is

DC-DC Converters 2-137

settling time For an NMA0505S, the isolation capaci-