ADDC02803SC/ADDC02805SA

REV. A

–15–

Circuit Setup for EMI Test

Figure 23 shows a schematic of the test setup used for the EMI

measurements discussed above. The output of the converter is

connected to a resistive load designed to draw full power. There

is a 0.1

µF capacitor placed across this resistor that typifies by-

pass capacitance normally used in this application. At the input

of the converter there are two differential capacitors (the larger

one having a series resistance) and two small common-mode

capacitors connected to case ground. The case itself was con-

nected to the metal ground plane in the test chamber. For the

RE102 test, a metal screen box was used to cover both the con-

verter and its load (but not the two meters of input power lead

cables). This box was also electrically connected to the metal

ground plane.

With regard to the components added to the input power lines,

the 100

µF capacitor with its 1 Ω series resistance is required to

achieve system stability when the unit is powered through the

LISNs, as the MIL-STD-461D standard requires. These

LISNs have a series inductance of 50

µH at low frequencies,

giving a total differential inductance of 100

µH. As explained

earlier in the System Instability section, such a large series

source inductance will cause an instability as it interacts with the

converter’s negative incremental input resistance unless some

corrective action is taken. The 100

µF capacitor and 1 Ω resis-

tor provide the stabilization required.

It should be noted that the values of these stabilization components

are appropriate for a single converter load. If the system makes

use of several converters, the values of the components will need

to be slightly changed, but not such that they are repeated for

every converter. It should also be noted that most system appli-

cations will not have a source inductance as large as the 100

µH

built into the LISNs. For those systems, a much smaller input

capacitor could be used.

Increasing Margin Between Specification Limit and

Measured Results

With regard to the 2

µF differential-mode capacitor and the two

82 nF common-mode capacitors, these components were in-

cluded in the test setup to augment the performance of the

power supply’s internal EMI filter. The values were chosen to

achieve the results shown in Figures 20 and 22. To increase the

margin between the specification limits and the measured emis-

sions, larger external component values could be used.

To do this it is useful to know that most of the emissions below

10 MHz, whether conducted or radiated, are due to differential-

mode currents flowing in the input power leads. To make the

emissions in this frequency range smaller, the differential ca-

pacitor value should be increased above 2

µF. Conversely, most

of the emissions above 10 MHz are due to common-mode cur-

rents; to make them smaller the common-mode capacitors

should be increased above the 82 nF value. In both cases it is

important to minimize the parasitic inductance of the capaci-

tors; the use of several smaller capacitors connected in parallel is

one way to achieve this.

Using larger valued capacitors than those shown in Figure 23 is

a good solution if an additional 6 dB–10 dB of margin is de-

sired. If, however, in an extremely sensitive application it is

desired to increase the margin by 20 dB or more, it may be

better to add both differential- and common-mode inductors to

the external components to make a higher order filter.

RELIABILITY CONSIDERATIONS

MTBF (Mean Time Between Failure) is a commonly used

reliability concept that applies to repairable items in which

failed elements are replaced upon failure. The expression for

MTBF is

MTBF = T/r

where:

T = total operating time

r = number of failures

In lieu of actual field data, MTBF can be predicted per

MIL-HDBK-217.

MTBF, Failure Rate and Probability of Failure: A proper

understanding of MTBF begins with its relationship to lambda

(

λ), which is the failure rate. If a constant failure rate is assumed,

then MTBF = 1/

λ, or λ = 1/MTBF. If a power supply has an

MTBF of 1,000,000 hours, this does not mean it will last

1,000,000 hours before it fails. Instead, the MTBF describes

the failure rate. For 1,000,000 hours MTBF, the failure rate

during any hour is 1/1,000,000, or 0.0001%. Thus, a power

supply with an MTBF of 500,000 hours would have twice the

failure rate (0.0002%) of one with 1,000,000 hours.

What users should be interested in is the probability of a power

supply not failing prior to some time t. Given the assumption of

a constant failure rate, this probability is defined as

R(t )

where R(t) is the probability of a device not failing prior to some

time t.

If we substitute

λ = 1/MTBF in the above formula, then the

expression becomes

R(t )

= e

–t

MTBF

This formula is the correct way to interpret the meaning of

MTBF.

If we assume t = MTBF = 1,000,000 hours, then the probability

that a power supply will not fail prior to 1,000,000 hours of use

is e

supply will last 1,000,000 hours before it fails. The probability

that the power supply will not fail prior to 50,000 hours of use is

e

ure is e

Temperature and Environmental Factors: Although the

calculation of MTBF per MIL-HDBK-217 is a detailed process,

there are two key variables that give the manufacturer signi-

ficant leeway in predicting an MTBF rating. These two vari-

ables are temperature and environmental factor. For users to

properly compare MTBF numbers from two different manu-

facturers, the environmental factor and the temperature must

be identical. Contact the factory for MTBF calculations for

specific environmental factors and temperatures.