Conducted Emissions Margin Calculator

How far under a conducted-emissions limit your measured or predicted spectrum sits, at every frequency you care about, with the worst one called out and the extra attenuation a stated design margin would need. MIL-STD-461G’s CE102 and all four CE101 limit curves are built in, read from the standard’s own figures, with their relaxation rules; any other limit, including the copyrighted ones, you enter as breakpoints.

Last updated 24 Sep 2026

Margin against a conducted-emissions limit

Limit breakpoints + your levels -> margin

Which limit are you working to?The five MIL-STD-461G sets carry the standard’s own numbers, read from its figures; the breakpoint boxes below then fill themselves and lock, so you can see exactly what is being used. MIL-STD-461G is a work of the US Government and may be reproduced. DO-160 and CISPR 25 are copyrighted and never will be here, which is what the last option is for. CE101 applies to surface ships, submarines, Army aircraft including flight line, and Navy aircraft only when the platform carries anti-submarine warfare equipment working between 30 Hz and 10 kHz. It does not apply to spacecraft, to Air Force aircraft or to ground installations.

Unit your limit and your measured levels are inCE102 is a VOLTAGE limit in dBuV, measured at the LISN’s signal port. All four CE101 figures are CURRENT limits in dBuA, measured with a current probe. Pick a built-in set and this box is set for you and locked, so the page can never put one unit’s limit against the other’s trace. The two are 20 log Z apart, which is 34 dB in 50 ohm, so mixing them is not a small error.

Nominal EUT source voltage, AC and DCUsed by CE102, whose figure carries a five-row relaxation table, and by CE101-4, whose figure selects curve #1 above 28 V and curve #2 at 28 V or below. THERE IS NO RELAXATION BELOW 28 V: a 12 V, 5 V or 3.3 V source gets the basic curve unrelaxed. The table lists five voltages and no rule for anything else, so the last option leaves CE102’s relaxation to you and to your procuring activity.

Equipment size class, for the two AC ship and submarine figuresThe threshold is the figure’s, not this page’s: 1 kVA input power on Figure CE101-2 (60 Hz), and 0.2 kVA on a multi-phase source or 2 amperes on a single-phase source on Figure CE101-3 (400 Hz). Note the direction — the larger class gets the LOWER limit on both figures.

Load current at the power frequency (A)Used only by the three CE101 ship and submarine figures, each of which prints its own relaxation rule. CE101-1 (DC): at or below 3 A use the curve as drawn, between 3 and 185 A relax by 20 log (I/3), at or above 185 A relax by the printed 35 dB. CE101-2 and CE101-3: above 1 A relax by 20 log (I), where I is the load current at the power frequency. CE101-4 prints no current relaxation.

How many breakpoints does your limit curve have?Count the corners on the limit line, including its two ends. Most conducted emissions limits are two or three segments on a log-frequency axis. With a built-in set selected this is set from the figure.

Breakpoint 1 frequency (kHz)Used only when you are entering your own limit. Pick a limit set above and this box is filled from that figure and locked; pick “my own breakpoints” and you type it. Read the breakpoint off the limit figure in your own copy of the standard.

Breakpoint 1 level (dBuV, or dBuA for a current limit)The limit’s value at that frequency, in whichever unit the limit is drawn in. Margin is the difference between two numbers in the same unit, which is why the unit selector above exists and why it locks itself when you pick one of the built-in sets.

Breakpoint 2 frequency (kHz)Used only when you are entering your own limit. Pick a limit set above and this box is filled from that figure and locked; pick “my own breakpoints” and you type it. Read the breakpoint off the limit figure in your own copy of the standard.

Breakpoint 2 level (dBuV or dBuA)

Breakpoint 3 frequency (kHz)Only used when the curve has three or more breakpoints.

Breakpoint 3 level (dBuV or dBuA)

Breakpoint 4 frequency (kHz)Only used when the curve has four or more breakpoints.

Breakpoint 4 level (dBuV or dBuA)

Breakpoint 5 frequency (kHz)Only used when the curve has five breakpoints.

Breakpoint 5 level (dBuV or dBuA)

Limit relaxation or offset (dB)With a built-in set selected this is the figure’s own relaxation, worked out from the source voltage or the load current above, and you cannot type over it. It stays yours in two cases: when you are entering your own limit, and when you have told CE102 that the source voltage is not one of the five its table lists — because the figure gives no rule for those and inventing one is exactly what this page will not do.

How many measured or predicted points do you want to check?

Point 1 frequency (kHz)Enter the points in increasing frequency order; the chart joins them in that order and the page refuses the arithmetic if they are out of sequence.

Point 1 level (dBuV or dBuA)What the receiver read, or what a prediction gives, IN THE UNIT SHOWN IN THE UNIT BOX ABOVE — dBuV against CE102, dBuA against any of the CE101 curves. This is the level at the EUT’s POWER INPUT through the LISN, not on any rail inside the unit. A CE101 level comes off a current probe clamped 5 cm from the LISN, not off the LISN’s own signal port.

Point 2 frequency (kHz)

Point 2 level (dBuV or dBuA)

Point 3 frequency (kHz)

Point 3 level (dBuV or dBuA)

Point 4 frequency (kHz)

Point 4 level (dBuV or dBuA)

Point 5 frequency (kHz)

Point 5 level (dBuV or dBuA)

Point 6 frequency (kHz)

Point 6 level (dBuV or dBuA)

Design margin you are working to (dB)6 dB is the usual programme requirement and it is a CONVENTION, not a rule from any of these standards — it exists to cover unit-to-unit spread, the receiver’s own uncertainty and the difference between your bench and the chamber. Some programmes ask for 10 or 12 dB; check your own specification.

Impedance used for the unit conversions (ohm)Only used for the conversions at the bottom, never for the margin. 50 ohm is the receiver’s input impedance and the value the conversions are conventionally quoted against. The MIL-STD-461G LISN’s own impedance is nothing like flat — Figure 7 of the standard takes it from about 5 ohm at 10 kHz to 47.6 ohm above 2 MHz, with a stated tolerance of plus or minus 20 per cent — and the converter emissions estimator computes it from the network in Figure 6 if you need it.

The arrangement a conducted-emissions limit refers to, drawn as a layout rather than as a circuit you would build. The EUT sits on a bonded ground plane and each power lead reaches it through its own LISN. For CE102 the measurement receiver looks at one lead at a time through the LISN's coupling capacitor into 50 ohm, and the limit is a POTENTIAL in dBuV. For CE101 the receiver is fed instead from a current probe clamped on the lead 5 cm from the LISN, and the limit is a CURRENT in dBuA; the probe appears on the drawing when a CE101 set is selected. Everything this page computes is a margin at THIS point — the power input of the unit — and not on any rail inside it. The worst-entered frequency, the limit there and the margin are carried live from the results.

-18.0dBExample

the MIL-STD-461G CE102 basic curve at a nominal 28 V source, against four predicted levels from an unfiltered 250 kHz converter: 250 kHz at 84, 500 kHz at 72, 1 MHz at 64 and 5 MHz at 52 dB above a microvolt, with a 6 dB design margin

A limit is a straight line between breakpoints on a log axis

limit(f) = L₁ + Σ_i s_i · clamp(log₁₀f − x_i, 0, x_i+1 − x_i)    s_i = (L_i+1 − L_i) / (x_i+1 − x_i)
margin(f) = limit(f) − level(f)    positive means below the line
extra attenuation = max(0, design margin − worst margin)
V(µV) = 10^dBµV/20    dBm = dBµV − 90 − 10·log₁₀Z    dBµA = dBµV − 20·log₁₀Z

x i: the base-10 logarithm of breakpoint i’s frequency. Interpolating on a log axis is not a choice — it is how the limit is drawn, and interpolating linearly in frequency instead can be tens of decibels wrong in the middle of a decade
s i: the slope of segment i in decibels per decade. A flat segment has slope zero; the common falling segment of a conducted emissions limit has a slope of roughly minus 20
margin: positive when you are below the line. This page never calls a positive margin a pass, because a limit is met by the standard’s own method on the standard’s own setup
Z: the impedance the conversions are quoted against. 50 ohm is the receiver’s input impedance and gives the familiar 107 dB between dBµV and dBm, and 34 dB between dBµV and dBµA

Worked example

the MIL-STD-461G CE102 basic curve at a nominal 28 V source, against four predicted levels from an unfiltered 250 kHz converter: 250 kHz at 84, 500 kHz at 72, 1 MHz at 64 and 5 MHz at 52 dB above a microvolt, with a 6 dB design margin
The CE102 basic curve is 94 dBµV at 10 kHz, 60 dBµV at 500 kHz and flat at 60 dBµV to 10 MHz. Between the first two breakpoints that is a slope of -20.012 dB per decade, which is the drawn 20 dB per decade to within 0.02 dB — 94 − 20·log(50) is 60.0206, so the figure's own three numbers agree with each other
28 V is the table's basic-curve row, so the relaxation is 0 dB and the curve is used as drawn. There is no relaxation below 28 V either: a 12 V or 5 V bus gets this same curve
250 kHz sits on the sloping segment, 1.3979 decades above 10 kHz, so the limit there is 66.02 dBµV and a level of 84 is 17.98 dB ABOVE it
500 kHz is the knee itself at 60, so 72 is 12.00 dB above; 1 MHz and 5 MHz are both on the flat at 60, where 64 is 4.00 dB above and 52 is 8.00 dB below
The worst of the four is -18.0 dB at 250 kHz — above the limit, not below it
To reach a 6 dB design margin there the noise has to come down by 23.98 dB. That is the attenuation requirement a filter design starts from, and it is needed at the converter's fundamental
As a check on the units, 84 dB above 1 microvolt is 15.85 mV, which in 50 ohm is -22.99 dBm and 50.02 dB above 1 microamp. That last conversion is the one that matters when you switch to a CE101 curve, because CE101 is drawn in dBµA and the two units are 33.98 dB apart at this impedance

The five MIL-STD-461G limit sets this page carries

Read from the standard’s own figures by rendering them at 300 dpi and tracing the drawn curves against the printed gridlines. Where the ink and the printed markers disagree — they do by one or two per cent in frequency on three of these figures, which is two or three pixels of drafting error — the printed marker is what is implemented. CE101 does NOT apply to spacecraft, Air Force aircraft or ground installations; check the standard’s requirement matrix before choosing one of its curves.
Set	Applies to	The curve, in the figure’s own units
CE102, Figure CE102-1 (para 5.5.1)	10 kHz to 10 MHz, all power leads including returns, every platform	94 dBµV at 10 kHz, −20 dB/decade to 60 dBµV at 500 kHz, flat to 10 MHz. Relaxed by the figure’s table: 28 V basic, 115 V +6, 220 V +9, 270 V +10, 440 V +12 dB
CE101-1, Figure CE101-1 (para 5.4.1)	Surface ships and submarines, DC power, 30 Hz to 10 kHz	95 dBµA from 30 Hz to 2.6 kHz, then straight to 76 dBµA at 10 kHz. Relaxed by 20·log(I/3) between 3 A and 185 A, by the printed 35 dB above
CE101-2, Figure CE101-2	Surface ships and submarines, 60 Hz power; from 120 Hz	a (60 Hz, 120) → b (1.92 kHz, 90) → c (10 kHz, 76) dBµA under 1 kVA; d (120 Hz, 90) → b → c at 1 kVA and above
CE101-3, Figure CE101-3	Surface ships and submarines, 400 Hz power; from 800 Hz	Curve #1, under 0.2 kVA or 2 A: 120 dBµA at 400 Hz falling 20 dB/decade to 92 dBµA at 10 kHz. Curve #2, at or above: flat 90 dBµA
CE101-4, Figure CE101-4	Navy ASW aircraft and Army aircraft including flight line	Above 28 V: 110 dBµA to 1 kHz then to 90 at 10 kHz. 28 V or below: 100 dBµA to 1 kHz then to 80 at 10 kHz

Which standard, and what may be reproduced

The site’s standing position is that a copyrighted standard’s tables and curves are not reproduced and that the reader types values from their own copy. MIL-STD-461G is the one document here that may lawfully be reproduced, and it now is — but only because every number came from reading its figures rather than from a summary of them. An unverified limit line is worse than none, because you would have no reason to doubt it.
Limit	Status	What this page does
MIL-STD-461G CE102, CE101	US Department of Defense interface standard; a work of the US Government, distributed without charge	Reproduced. All five limit curves and both relaxation rules are built in, read from the standard’s figures and checked against the markers printed on them
RTCA DO-160 Section 21	Copyrighted; sold by RTCA	Method and category letters named, values entered by you from your copy
CISPR 25, CISPR 32	IEC documents, copyrighted	Method and class named, values entered by you from your copy
CISPR 17	IEC document, copyrighted	Named as the source of the 50 ohm / 50 ohm insertion-loss convention only
A programme-specific limit	Your customer’s document	Exactly the same treatment: type the breakpoints

The relaxation rules, exactly as the figures print them

The CE102 table is five discrete rows and the figure gives no rule for a voltage it does not list; a greater-or-equal mark is printed to the left of the voltage column, and it sits between the 220 V and 270 V rows rather than against any one of them, so it does not settle the question either. This page therefore hands an unlisted voltage back to you. The five printed values do follow 10·log(V/28) to within 0.16 dB — 6.14, 8.95, 9.84, 11.96 against 6, 9, 10 and 12 — which is worth noticing and is not a rule; it is certainly not 20·log(V/28), which would give 12.3 dB at 115 V. CE101-1’s second and third rules do not meet: 20·log(185/3) is 35.80 dB, so the relaxation steps DOWN by 0.8 dB as the load current crosses 185 A. Both are implemented as printed, steps and all.
Figure	Condition	Relaxation
CE102-1	Nominal EUT source voltage (AC and DC) 28 V	Basic curve
CE102-1	115 V	6 dB
CE102-1	220 V	9 dB
CE102-1	270 V	10 dB
CE102-1	440 V	12 dB
CE101-1	Load current 3 A or less	Use the limit curve as shown
CE101-1	Load current between 3 and 185 A	20·log(I/3)
CE101-1	Load current 185 A or more	35 dB
CE101-2 and CE101-3	Fundamental current — the load current at the power frequency — greater than 1 A	20·log(fundamental current)
CE101-4	None printed	—

What moves a conducted-emissions margin, and by roughly how much

None of these is in the arithmetic above, and every one of them is why a design margin exists. The last two are the ones that turn a comfortable bench result into a chamber failure, because neither is a component value.
Change	Direction	Typical size
Full load instead of light load	worse	5 to 15 dB at the switching harmonics
Lowest input voltage instead of nominal	worse	a few dB; duty cycle and peak current both move
Ceramic filter capacitor at its rated voltage	worse	a Class II ceramic can lose half its capacitance, which is 6 dB
Cold start, before the ferrite warms up	varies	core permeability and capacitor ESR both move with temperature
A harness routed beside the switching stage	worse	unbounded; this is coupling, not conduction, and no filter fixes it
Peak detector instead of average	worse	several dB on broadband content; the standards specify which
Chassis bond made with a wire instead of a face	worse	the Y capacitors’ return path is the bond; a wire is an inductor

Margin, not a verdict

A conducted-emissions limit is a line on a chart with decibels up the side and a logarithmic frequency axis along the bottom. Between its breakpoints it is straight, so the whole curve is described by a handful of corner points. This page carries five of them outright — MIL-STD-461G’s CE102 and all four of its CE101 figures, with the relaxation rules printed alongside them — and takes yours for anything else.

Where the built-in numbers came from. Not from memory, and not from a summary. Each figure’s page was rendered at 300 dpi and the drawn curve traced column by column against the gridlines, then checked against the markers the figures themselves print: the 94 and the 60 on CE102-1, the 95 and the 76 and the 2.6 on CE101-1, the a, b, c and d points and the 1.92 on CE101-2, the 92 on CE101-3, the two flat levels on CE101-4. Where the ink and the markers disagree the markers win, and the disagreements are recorded: the drawn knee on CE101-1 sits at 2.53 kHz against a printed 2.6, and both dashed markers on CE101-2 sit 1.4 per cent high. That is drafting error of two or three pixels, not a different limit, but it is the sort of thing that has to be looked at rather than assumed away.

Why the other limits are still typed in. RTCA’s DO-160 and the IEC’s CISPR 25 and CISPR 32 are sold documents, and reproducing their curves would be reproducing the part of them that has value. MIL-STD-461G is different: a US Department of Defense interface standard, a work of the US Government, distributed without charge. That is why one of these is built in and the others never will be. The breakpoint mode is not a second-best — it is the same arithmetic, and it is what a programme-specific limit needs too.

dBµV and dBµA are not interchangeable. CE102 is a voltage limit read at the LISN’s signal port. All four CE101 figures are current limits read from a probe clamped on the lead. The two are 20·log of the impedance apart — 34 dB in 50 ohm — so a dBµV reading put against a dBµA limit is not slightly wrong, it is wrong by more than any margin anyone argues about. Selecting a built-in set locks the unit control to that figure’s unit, and every conversion this page prints follows it, so the page cannot put one unit’s limit beside the other’s trace. What it cannot do is stop you typing the wrong numbers into the level boxes, which is why the level printed in the other unit is there below: if that number looks absurd, yours is in the wrong one.

What a positive margin is, and what it is not. It is the distance between a number you measured and a number on a figure. It is not a compliance statement. A limit is met by measuring the equipment under test the way the standard says — its setup, its ground plane, its bonding, its receiver bandwidths, its dwell times, its detector, at the operating points the test plan calls out — and a bench measurement with a spectrum analyser and a clip lead is not that. What a margin is good for is engineering: knowing how much attenuation to design for, knowing which frequency is driving the filter, and knowing whether the last change helped.

Applicability comes before the curve. CE102 applies to every platform in the standard’s matrix. CE101 applies only to surface ships, submarines, Army aircraft including flight line, and Navy aircraft that carry anti-submarine warfare equipment working between 30 Hz and 10 kHz — not to spacecraft, not to Air Force aircraft, not to ground installations. And for an AC application CE101 starts at the second harmonic of the power frequency, which is 120 Hz on the 60 Hz figure and 800 Hz on the 400 Hz one, even though both figures draw their upper curve from the fundamental. Picking the wrong figure is a bigger error than any interpolation on this page.

The 6 dB. Six decibels of design margin is what most programmes ask for and it appears in none of these standards as a requirement. It is there to absorb the things the measurement does not see: unit-to-unit spread, the receiver’s own uncertainty, capacitor derating with DC bias and temperature, the difference between your bench and the chamber, and the fact that a handful of spot frequencies is not a swept measurement. Space and launch programmes often specify more. It is a number from your specification, not from physics, which is why this page asks for it rather than assuming it.

Where the measurement is. Conducted emissions are measured on the leads entering the unit, through a LISN, one lead at a time. That is worth repeating because the noise a designer can see on an oscilloscope is usually on an internal rail, and the two are not the same measurement. A reader chasing ripple on a 3.3 V point-of-load rail very often has a problem at the 28 V bus input, and a filter on the 3.3 V rail does nothing about it. Once you have an attenuation requirement from this page, the filter designer turns it into components and the insertion-loss page says what those components actually deliver once their parasitics are in; where the spectrum comes from before hardware exists, the converter emissions estimator predicts the differential-mode part of it through the standard’s own LISN network.

Frequently asked questions

Where do the MIL-STD-461G curves on this page come from?

From the standard’s own figures — CE102-1 and CE101-1 through CE101-4 — read by rendering each figure at 300 dpi, calibrating on the printed gridlines and tracing the drawn curve, then checking the result against the numeric markers the figures themselves carry. MIL-STD-461G is a work of the US Government, distributed without charge, so its curves may lawfully be reproduced. DO-160 and CISPR 25 are copyrighted and are not, which is what the breakpoint mode is for.

How is the CE102 limit adjusted for the EUT source voltage?

By the table printed on Figure CE102-1 and by nothing else: 28 V is the basic curve, 115 V relaxes it by 6 dB, 220 V by 9 dB, 270 V by 10 dB and 440 V by 12 dB. There is no rule for voltages the table does not list, and there is NO RELAXATION BELOW 28 V — a 12 V, 5 V or 3.3 V source gets the basic curve unrelaxed. The five rows follow 10·log(V/28) to within 0.16 dB, which is worth noticing and is not a rule; it is certainly not 20·log(V/28), which would give 12.3 dB at 115 V where the table says 6.

Which CE101 figure applies to me?

CE101-1 for surface ships and submarines on DC power; CE101-2 for the same platforms on 60 Hz; CE101-3 for the same platforms on 400 Hz; CE101-4 for Army aircraft including flight line and for Navy aircraft that carry ASW equipment working between 30 Hz and 10 kHz. If your platform is not one of those, CE101 does not apply at all — it is absent for spacecraft, Air Force aircraft and ground installations. On an AC platform the requirement starts at the second harmonic of the power frequency.

What does a positive margin mean?

That the level you entered is below the limit, by that many decibels, at that frequency. It does not mean the equipment complies with anything. Compliance is a measurement made to the standard’s own method on the standard’s own setup, with its receiver bandwidths and detectors, at the operating points the test plan requires — and it is a test house that says so, not a calculator.

Why 6 dB of design margin?

Convention, not a rule from any of these standards. It covers unit-to-unit spread, measurement uncertainty, capacitor derating with DC bias and temperature, and the gap between a bench measurement and a chamber. Many space and launch programmes specify more than 6 dB, and some commercial work uses less. Take the number from your own specification.

How do I convert dBuV to dBm or to dBuA?

dBm = dBuV minus 90 minus 10 log of the impedance, which is dBuV minus 107 in 50 ohm. dBuA = dBuV minus 20 log of the impedance, which is dBuV minus 34 in 50 ohm. Both are on this page, computed against whatever impedance you enter. The second one matters because CE101 is a current limit and CE102 is a voltage limit, and comparing one against the other without the impedance is a 34 dB error. Note that the LISN’s own impedance is not 50 ohm at audio frequencies — it is about 5 ohm at 10 kHz — so 34 dB is a convention for quoting, not a conversion between the two requirements.

My limit has a slope. Do I enter it as a slope or as breakpoints?

As breakpoints. Give the frequency and level at each end of the sloping segment and the page interpolates a straight line between them on the logarithmic frequency axis, which is exactly how the limit is drawn. Interpolating linearly in frequency instead — which is what happens if you treat a log chart as a linear one — can be tens of decibels wrong in the middle of a decade.

Can I use this for radiated emissions limits too?

The arithmetic is identical: a piecewise-linear limit on a log axis, a measured level, a difference. What is not identical is everything around it. Radiated limits are in dBuV/m at a stated distance, they depend on antenna factors and chamber corrections you have to apply before you get a level to type in, and a margin against a radiated limit says nothing about the conducted one or the other way round. RE102 is deliberately not built into this page.

Related calculators

References

MIL-STD-461G, Requirements for the Control of Electromagnetic Interference Characteristics of Subsystems and Equipment, 11 December 2015, superseding MIL-STD-461F. A US Department of Defense interface standard and a work of the US Government, distributed without charge. The limit sets on this page are taken from paragraph 5.4 (CE101: 5.4.1 applicability, 5.4.2 limits, 5.4.3 test procedure) and paragraph 5.5 (CE102: 5.5.1 applicability, 5.5.2 limits, 5.5.3 test procedure), with the curves read from Figures CE101-1, CE101-2, CE101-3, CE101-4 and CE102-1 and the relaxation rules from the annotation boxes printed inside those same figures. Table V, the requirement matrix, is the source of the applicability statements about which platforms CE101 and CE102 apply to.
RTCA DO-160G, Environmental Conditions and Test Procedures for Airborne Equipment, Section 21 Emission of Radio Frequency Energy. Copyrighted and sold by RTCA; its category letters and limit curves are not reproduced here. Cited for the method and the category structure only.
CISPR 25, Vehicles, boats and internal combustion engines — Radio disturbance characteristics — Limits and methods of measurement for the protection of on-board receivers. An IEC document, copyrighted; its five severity classes and its voltage and current methods are named here, its limit values are not reproduced.
CISPR 17 Edition 2.0, 2011-06, Methods of measurement of the suppression characteristics of passive EMC filtering devices. Defines the asymmetrical (common-mode), symmetrical (differential-mode) and unsymmetrical test circuits, with impedances referenced to 50 ohm; the first edition’s worst-case methods were deleted in this edition as no longer used in industry. Verified from the standard’s published preview; cited for the measurement convention, with no values reproduced.
Ott HW. Electromagnetic Compatibility Engineering. Wiley, 2009. Chapter 13 Conducted Emissions: 13.1.1 Line Impedance Stabilization Network, 13.2.1 Common-Mode Emissions, 13.2.2 Differential-Mode Emissions. The chapter and section titles were checked against the author’s own published contents listing.