WARRINGTON - Monday 17 May - Alain CHAROY: Power Converters For Particle Accelerators

Download as pdf or txt
Download as pdf or txt
You are on page 1of 49

EMC

WARRINGTON - Monday 17 May - Alain CHAROY


AEMC
POWER CONVERTERS FOR PARTICLE ACCELERATORS - [email protected]

Introduction Differential Mode Immunity Differential Mode Emissions Common Mode Emissions Electromagnetic Radiations

EMC terms and lab assessment conditions


Electromagnetic emission
Standardized test

Electromagnetic susceptibility
Standardized test

Given EM disturbance Defined Measuring equipment Standardized Measuring conditions

The best situation is to undertake EMC tests & to solve EMC problems before they appear in situ

Given EM disturbance Defined Disturbance source Standardized Measuring conditions

Emission-test level

Immunity-test level

Emission-test limit

Immunity-test limit

Single emitter
Deterministic situation

Single victim
Deterministic situation

EMC best controlling conditions


Electromagnetic emission
Standardized test

Electromagnetic susceptibility
Standardized test

In situ test

Given EM disturbance Defined Measuring equipment Standardized Measuring conditions

Given EM disturbance Defined Measuring equipment Defined Measuring conditions

Given EM disturbance

Given EM disturbance Defined Disturbance source Standardized Measuring conditions

An EMC Validation is desirable to plan in situ after system installation before problems appear

Defined Disturbance source Defined Measuring conditions

Emission-test level

Emission level

Immunity level

Immunity-test level

Emission-test limit

Emission limit

Immunity limit

Immunity-test limit

Single emitter
Deterministic situation

Single victim
Deterministic situation

EMC troubleshooting
Electromagnetic emission
Standardized test

Electromagnetic susceptibility
Standardized test

In situ test

Given EM disturbance Defined Measuring equipment Standardized Measuring conditions

Given EM disturbance Defined Measuring equipment Defined Measuring conditions

Given EM disturbance Defined Measuring equipment Adapted Measuring conditions

Given EM disturbance Non-defined Disturbance sources Adapted Measuring conditions

Given EM disturbance Defined Disturbance source Defined Measuring conditions

Given EM disturbance Defined Disturbance source Standardized Measuring conditions

Emission-test level

Emission level

Disturbance level

Interference level

Immunity level

Immunity-test level

Emission-test limit

Emission limit

Disturbance limit

Interference limit

Immunity limit

Immunity-test limit

Single emitter
Deterministic situation

Probability situation
Superposition of disturbance

Single victim
Deterministic situation

Common Mode & Differential Mode

ICM
2 2

ICM
Common Mode Path

Equipment

VCM
Ground (chassis)

IDM IDM VDM


Equipment

Differential Mode Path

The 5 kinds of disturbances generated by a converter


5

INPUT

3 VDM

ZC1

ZC2

VDM

4 OUTPUT

Z
Safety wire Chassis ground

1 1 2 3 4 5
Input-to-Chassis Common Mode Input-to-Output Common Mode Input Differential Mode Output Differential Mode Electromagnetic radiations (E & H)

How to measure CM & DM currents ?

ICM 2 ICM 2

Current probe (Clamp)

CM current measurement

ICM

IDM IDM
DM current measurement

2.IDM

Typical input current of a 5 kVA filtered converter

RTCA DO160D Power lines category B

DIFFERENTIAL MODE

COMMON MODE

Frequency

How to measure a disturbing voltage ?


50 H 250 H

EUT
To Spectrum Analyzer or 50 load 220 nF 8 F 1 F 1k 5

Line
Bleeder 100 k

Ground
CISPR 50
100 50

// (50 H + 5

) LISN
LISN 50 H

LISN impedance

LISN 5 H 10 6

10 kHz

100 kHz

1 MHz

10 MHz

30 MHz

A concealed key point: the switching dynamic impedance


V
100 kV
k M k 10 0 10 1 k

10 kV
P = 10 kW

Circuits in series
1 W M

Increasing severity

High impedance zone

Z P = 0 10 kW

P =

E field dominates, so:


Reduce parasitic capacitors Limit high V/ t trace lengths Choose low r materials (air !) Use lower V circuits in series Increasing severity
V diode critical

1 kV

P = 0 10 W

P = 1 kW

100 V
P P = 1 W = 10 W

10 V
10 0 10 1 = = Z = Z Z Z = 0. 1

1V 10 mA 100 mA 1A 10 A 100 A 1000 A


Circuits in parallel

Low impedance zone

H field dominates, so:


Reduce ESR and ESL Limit high I/ t loop areas Choose sandwich geometries Use lower I circuits in parallel

Any switching circuit should be positioned in this plane

Introduction Differential Mode Immunity Differential Mode Emissions Common Mode Emissions Electromagnetic Radiations

The voltage tolerance boundary

New Voltage-Tolerance Boundary New parameters (bullets) Old Voltage-Tolerance Boundary

Duration of Disturbance in Cycles (c) and Seconds (s)

Voltage-Tolerance Envelope

106 90

200 s

1 ms

3 ms

20 ms

0.5 s

10 s

Steady State

Transient Turn-on Overvoltage


Self-pulsation :
0

1 = L.C L R

Quality Factor : Q =

L. R

Vout
C

Vin
Time response 1.8 ( Vin = 1 )
1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 1 2 3 2

Q = 10 Q=5 Q=2

Q=1 Q = 0.7 Q = 0.5

0.t

Where to install a DM Voltage Transient Suppressor ?


Overvoltage : kV Peak rsidual voltage : kV

Power line

EMC Filter

CONVERTER

Without TVS : serious risk of destruction


Peak rsidual voltage Clipped overvoltage 500 V 800 V !

MOV (Ageing)

EMC Filter

CONVERTER

Varistor on line input : a risk remains


Overvoltage : kV Peak rsidual voltage 500 V

Power line

EMC Filter

CONVERTER
& Filter inductance prevents TVS premature triggering

OK Varistor at converter input : best results

Where to add protection components ?


Overvoltage protection Diode
No impedance on the DC side to limit reverse overvoltage on rectifier bridge

NTC

AC Line

Output

PFC Boost

No voltage output voltage doubling A Diode avoids doubling limitation no inrush current limitation ABut No inrush current limitationcurrent NTC thermistor limits inrush yet

The problem of the negative impedance of a DC/DC converter


Zout Output EMC filter Z cable R + jL Input EMC filter

Z1
SOURCE

Z2

Zin
DC / DC

Negative impedance

positive Impedance

Zin
65

0 Phase
-180 0.1 Hz 1 Hz 10 Hz 100 Hz 1 kHz 10 kHz

Risks :

Solutions :

- No Start Up - Add a large capacitor at DC / DC input - Output voltage instability - Reduce cable inductance (several pairs in //) - Destruction of DC / DC converter - Reduce the converter regulation bandwidth

Introduction Differential Mode Immunity Differential Mode Emissions Common Mode Emissions Electromagnetic Radiations

Some problems of converters harmonics


Harmonics are generated by non sinusoidal currents. For an electric network, harmonics are a low frequency problem ( < 2 kHz and in Differential Mode only). Usually, even harmonics are low (because + and half-waves look the same). Most of inverters and AC / DC converters without PFC exceed normalized levels. Odd harmonics of converters can be severe ( > 50 % @ H3 ; > 30 % @ H5 ). For most single phase converters without PFC on a 3 phase network, the 3rd harmonics (150 Hz) is an homopolar current. So, Ineutral can exceed Iphase. Anti-harmonic or active filters are useful for a low power source (electric generator). For a high power network, the problem of harmonics is not the voltage distortion but the mastering of cabling protection scheme (cables & circuits breakers).

Differential Mode interferences

IDM

Power Line

V1
IDM

Z1

I1

I2

2 Z2 V2

Z Load

V1
2

V1
2

1 V1

Z1 . I1 ( If Z LISN >> Z1 - Generally, V1 = f ( F ) )

V1 is not applied to the secondary, so it does not disturb the load.

2 V2

Z2 . I2 ( If Z Load >> Z2 - Generally, V2 = f ( t ) )

V2 may be disturbing ( typically if peak-to-peak V2 1 V )

Differential Mode Emission Spectrum Without Filtering


Converter DM Equivalent Scheme
Switching frequency = F0 Transition time = r

I
dBA

Z capacitor C ESR ESL


LISN 100

1 F

Fc =

0.35
r

Switched current simplified spectrum


1 F2

VDM
F0 Fc

0.5 to 5 MHz

log (F)

Convolution Result

Z
dB
1 C

Electrolytic capacitor impedance ESR Fd = 2 .ESL ESL. ESR

VDM
dBV

Rectifier bridge wideband noise (diode recovery for AC/DC converter)


1 F

KHz

Fd

0.5 to 5 MHz

log (F)

F0

Fc

Fd

log (F)

Insertion Loss of a Differential Mode EMC Filter


L = L1 + L2 L1

Switched current

C
ESR

There wire EMC Filter Cx L2 Back wire

LISN 100

V: V without filter V11 : Without filter With EMC filter V2:DM Impedance

DM equivalent scheme of aaconverter with an EMCfilter DM equivalent scheme of converter without any filter
+20

dB
+10 0 -10 -20

V2 V1

Resonance before cut-off

F0 =
-30 -40 0.1

1 L.Cx
0.3 0.5 0.7 1 2 3 5 7 10

F F0

0.2

Traps of a Differential Mode EMI filter


4 3 L C

H
LISN 7

L1 Cx L2

Electrolytic capacitors

Converter

1 - Choose the proper structure (to mismatch the impedances) 2 - Choose (L1+L2) x Cx value so that Fresonance < lowest frequency to filter 3 - Verify that no inductance saturates at max current (Max P & Min V) 4 - Limit H field coupling to leakage inductance (in air) of L1 & L2 5 - Safety margin necessary to compensate electrolytic caps ESR dispersion 6 - Add C as needed to reduce wideband recovery noise of rectifier bridge 7 - Limit H field-to-loop coupling to avoid parasitic voltage pick-up.

Take care of Differential Mode cabling

Those cabling inductances reduce filtering effectiveness

NO !

Take care of Differential Mode cabling

NO !

Those areas radiate if they BETTER carry high I / t

Take care of Differential Mode cabling

NO !

BETTER

BEST !

How to reduce cabling parasitic impedances


ZG ZL ZG ZL

V
Serial Impedances

V
Parallel Capacitance

Improper Routings
Z < 10

Z>1K

To reduce the cabling areas is necessary, but insufficient

V
To move away

Minimal length

Correct Routings

How to measure Output Ripple


C = 1 F to measure HF ripple only C = 100 F to measure 100 Hz ripple

Input

Converter

Nominal Current

Output terminal

Very short connection (Max length = 2 cm)


Oscilloscope

50

Coaxial Cable

50

input

How to analyse Output Ripple


C = 1 F to measure HF ripple only C = 100 F to measure 100 Hz ripple

Input

Converter

Output terminal

Nominal Current

Very short connexion (Max length = 2 cm) Oscilloscope

50

Coaxial Cable

50

input

Voltage 60 mV
Usually CM - to - DM conversion

60 mV
HF NOISE ( @ > 3 MHz ) RIPPLE ( @ Switching F ) RIPPLE + NOISE

10 mV : Excellent

100 mV : Average

1 V : Excessive

Introduction Differential Mode Immunity Differential Mode Emissions Common Mode Emissions Electromagnetic Radiations

Common Mode interferences ICM


2
C2 ZDM

Load

Power Line

ICM
2

V C1 ZCM

Safety wire

ICM

1 I1

2 I2

1 I1 = C1 . V / t
I1 doesnt circulate through the load, so it is little disturbing.

2 I2

C2 . V / t (but possibly modified by ZCM)

I2 can circulate through the load, so it may be very disturbing.

Measured total CM current : ICM = I1 + I2

Common Mode Emission Spectrum Without Filtering


Converter CM Equivalent Scheme
Switching Frequency = F0 Transition Time = r

V
dBV

cable inductance

1 F

Fc = 0.35
r

V
C

1 H 30 pF to 3 nF LISN 25

Switched Voltage Simplified Spectrum


1 F2

VCM
F0 Fc

C : Parasitic cap between hot conductors & ground

1 to 10 MHz

log (F)

Convolution Result
1 Z dB Loop admittance -1
F Fr = 1 2 LC log (F) Fr 3 to 30 MHz F0 Fc Fr log (F) Resonance

VCM
dBV
1 F flat Resonance 1 F 1 F3

Insertion Loss of an EMC Common Mode Filter


CCM = Cp + C'p + 2 x Cy

M 2 wires EMC Filter LISN 25

Switched Voltage
Hot Cap

V
Cp

C'p

2 x Cy
Ground

V1: : Without filter V1 V without With the filter V2:CM impedance

CM equivalent scheme ofof an isolated converter with a a filter CM equivalent scheme an isolated converter without filter
+20

dB
+10 0 -10 -20

V2 V1

Resonance before cut-off

F0 =
-30 -40 0.1

1 M.CMC
0.3 0.5 0.7 1 2 3 5 7 10

F F0

0.2

The 3 cases of Primary-to-Secondary Common Mode


Metallic chassis Grounded output Filter Converter Electronics

No disturbance outside of the chassis No CM noise in electronic circuits EMC filter easy to optimize

IMC
Not filtered output Filter Converter Electronics

No disturbance outside of the chassis CM Noise through electronic circuits EMC filter more difficult to optimize

IMC

3
Filter Converter

EM radiations outside of the chassis Input filter impossible to optimize


IMC

IMC

Not Filtered output

The output cable must be shielded or filtered


Load

To float or not to float the output, thats the question


CM inductance can saturate
Primary circuits Connexion to ground

I+ I Ig

Ground

Ig I+ I
I+

Primary circuits

M
C 100 nF

I I+ = I

A (nearly) universal solution

How to measure Primary - to - Secondary C. M. current ?


Nominal V

Converter

Nominal I

Time measurement
50 mV/mA sensitivity 100 MHz bandwidth 1 mA peak-peak = Excellent 10 mA peak-peak = Average

Oscilloscope Coaxial cable

50

100 mA peak-peak = Excessive

Nominal V

Converter

Nominal I

Frequency measurement
9 or 10 kHz RBW, Peak detection Span : 0.1 to 50 MHz (100 MHz) 10 dBA = Excellent 30 dBA = Average 50 dBA = Excessive

Short wire Current clamp Spectrum Analyser

This simple CM / DM SEPARATOR reduces by 10 + the time and difficulty to optimize a single-phase EMC filter From LISN Line 1 2.N turns N turns To analyzer Differential Mode Output

50 Line 2 Common Mode Output


Ferrite tore with r 5000 & AL > 2000 nH/turn2 e.g. Philips 3E25 (orange), Diam. = 14 mm, N = 7

50

Practical realisation of a CM / DM SEPARATOR

Line 1 input AND Line 2 input

DM output OR CM output

CM / DM separator adaptation on a commercial LISN

Added BNC on the non-measured output (internal 50 suppressed)

2 coaxial cables with same length

Introduction Differential Mode Immunity Differential Mode Emissions Common Mode Emissions Electromagnetic Radiations

Sources of Electromagnetic Radiations


1 H

I1
E

I2

IMC

1 Sources of H field :
Leakage fields of windings Secondary loop areas Primary loop area

2 Sources of E field :
High V/ t conductive parts (Heat sink, ferrite core) HF insufficiently filtered cables (e.g. output cable)

HF solutions must be installed close to the sources


Noisy converter
Ni - Zn Ferrite tube

Load

Noisy converter

Output

2 x 1 nF
Chassis metal sheet

High r bead

BLM Ferrite bead

Clock C
R 22 C 47 pF

R 10 to 100

MOS

HF Diode

Even small converters (few W) can be very noisy (I/O CM & radiation)

Ground Loop : Definition & Effects

Apparatus #1

Interconnect. cable

Apparatus #2

GROUND LOOP

Z
nearest ground conductor / structure

Ground loop cannot be avoided !

1 Common impedance coupling


Earth impedance does not matter

Field - to - Loop coupling

Star Grounding : Principle & Reality 3 2

1
Voltage reference

The real world ! The Theory

Ground Grid : Definition & Effects

Apparatus #1

Apparatus #2

ground conductor / structure

GROUND GRID
Other ground wire or structure

How to improve immunity ?

Ground Grid : Definition & Effects

Reduction of Ground Loop = Apparatus #1 Better immunity against radiated fields Apparatus #2

Improvement of Ground Grid = Better immunity against conducted disturbances

Adding a Ground Strap = Further reduction of the Ground Loop Area

A GROUND GRID is highly recommended !

Where to connect the shielded cables braid ?


Any power cable : At both ends, to chassis ground, without pigtail. High frequency coax : At both ends, to chassis ground, without pigtail. Digital link (except coaxial Ethernet): At both ends, to chassis ground High impedance source (> 10 k ): At both ends, to chassis ground Any cable inside an equipment : At both ends, to chassis ground Any outer shield (not signal return): At both ends, to chassis ground Low voltage signal cable, with low frequencies to transmit, with a low impedance source, in a noisy environment, without balanced transmission (bad CMRR): At one end only

But then good immunity will be hard to achieve !


Avoid aluminium foil with a drain wire (without braid).

Please, let us remember EMC is not black magic (Just simple physics) Some measurement equipments are required Usually, only simple equipments are sufficient Its good to be experienced (& confident enough) Its important to understand how system works Its useful to methodically analyse what happens Its efficient to foresee and simplify EMC problems Its necessary to know the orders of magnitudes Its politically effective to be persuasive (& smiling) Its essential never to become discouraged !

Questions ?

You might also like