Conducted Emissions Tester
The goal of this project is to make a small LISN (Line Impedance Stabilization Network) to "see" with spectrum analyser or oscilloscope with FFT the noise on power line.Why ?
The goal of this project is to make a small LISN (Line Impedance Stabilization Network) to "see" with spectrum analyser or oscilloscope with FFT the noise on power line.
Why ?
Only a few people like EMC compatibility issues ... In most common case, i (and, i think a lot of electronic hobbyist) don't think about emc questions when prototyping. Last week, i've got a lot of noise problems on a (quite) low frequency oscillator. For three days, i've observed reasonable performance the mornings and very bad stability in the evenings. The offender was an energy saving lamp ..!
Maybe other people here have lamps who loves jokes ... :)
Discussie (2 opmerking(en))
Hooligan0 11 jaar geleden
This first step, will be a prototype, only for the proof and component selection.
Design and schematics
The design is directly based on CISPR22 and CISPR16 recommendations. From the input, there are two capacitors (C1, C2) and two inductors (L1, L2) used to filter the AC source. With them, it is possible to differenciate interferences from the AC and from the device under test. This components must be choosen with appropriate security properties (see X2, MKP, ...)
/* Note : The zip file below, contains the complete kicad project for this prototype */
On figure 1, you can see the prototype. For this test, i've used a hi-voltage 100x probe to connect oscilloscope. I haven't see any hi-voltage spikes, but for the next revision, i will add a protection (maybe diode suppressor).
I have a differencial circuit breaker on the AC line. At the input and the output, capacitors are connected to ground and create leakage currents that are not allowed by circuit breaker. So, it is not possible to connect the filter directly. On figure 2, you can see a 1:1 transformer used to isolate the AC line. It can be very dangerous (maybe security device are here for a good reason :) ), at the begining, a very low power transformer have been used (Triad 25VA)
Power on
Firstly, a resistive load have been connected. The result can be seen on figure 3. /* Later, i'll write an analysis */
As second load, a small CPL adapter have been connected. The result can be seen on figure 5.
On this two tests, only low frequency are displayed.
Now :
emi_tester.zip (4kb)
emi-proto1.jpg (76kb)
emi-proto1-table.jpg (124kb)
emi-proto1-res-load.jpg (168kb)
emi-proto1-load-noise.jpg (182kb)
emi-proto1-load-cpl.jpg (90kb)
emi-proto1-sch.png (40kb)
Hooligan0 11 jaar geleden
Introduction
In the beginning were resistive loads. In this case, the current used by the load is always in phase with the voltage. The apparent power (VA) is equal to the real power (W). One day, loads have become complex (mathematical sense), in other words, when electro-technicians have added capacitors and inductors. Current and voltage are no more in phase and a new power has been introduced : reactive power. (See Fig 2) Now, the power can go forward to the load, and come back ! The final act of complexity starts with active components. They use power when they want (in phase, out of phase or at any frequency).
In a perfect world it would not really be a problem, but in our world : wires have resistivity, more than one equipment is connected to a generator, and more generally, the actual operation is only approximate version of what was expected.
Power Network
When an arbitrary number of equipments are connected to a power line, resistive, capacitive and inductives effects are present into wires. If a big noise is produced by one of them, it is propagated to all others. This kind of noise is called conducted emissions (see Fig3).
To avoid (or reduce) the problems, each equipment connected must limit emissions made to a reasonable value, as defined by national and international standards.
Test and mesure
To identify and measure the interference of a device on the network it must be isolated by a filter. At the filter (just after) it is possible to "see" the noise from the device and the noise from the network (just before). See Fig 4
The limits values are given for industrial (class A) and residencial (class B) target environments. The figure 5 give in a graphic the class B limits, and a gnuplot configuration file is also available (see below)
usage : gnuplot class_b.cfg -
power-4equip.png (2kb)
powerrsil.png (1kb)
class_b_gnuplot.zip (1kb)
power-b-limits.png (21kb)