EMC tutorials - [2/3] Testing the LISN

[@placiduzzu]

There’s always the option to buy these things and some may say that you end up with a higher cost if you do it yourself because of the calibration process involved. But during the LISN design you really get an insight of what are the pitfalls for the measurement and set the solid premises to interpret the data correctly. Good job!

[@FesZElectronics]

I guess the same could be said about any electronic project; once you factor in all the parts, and spare parts and hours spent, the "home made" device is usually more expensive than the commercial product. But of course, this way you fully understand the device, something that could not be done just by purchasing and using it. But now, talking specifically about an LISN, the parts involved are really cheap, I mean the connectors and metal case are the most expensive parts, and the commercial units are produced in very small quantities so its much easier to build something and spend less then just buying it ready made.

[@davehyypio4261]

I’m also thinking of building my own 15A, 240V LISN. Your tutorial really opened my eyes to the performance parameters of such a device, particularly crosstalk. Thanks for addressing this rather rare topic!

[@dimitarzhekov9550]

how did the design came out to be?

[@JoyeMusiccom]

Nice, thorough explanation! I'm considering making my own LISN for lower-powered electronics since the commercial units are outrageously expensive.

[@FesZElectronics]

If you are looking for experimentation and not proper compliance testing then it should not be a problem to use home-made LISN's. I mean its quite a simple circuit, and even if its not perfect, you can still get valid results. If you build it, let me know how it turns out!

[@WR3slo]

I built my first LISN 2 years ago, just a single channel with 5x 1uH inductors and a capacitor. It help me a lot with precompliance testing for a few drivers. Now that you gave me all the information regarding measurement I designed a new one with dampened air coil and copied 10dB transient protection. It is the same one box construction as yours, but I will split it with a solid wall in two chambers. First measurement gave me nice shape but a 20 OHM impedance. To simplify the formula I removed the attenuator and got 300OHM... It will take some time to finish xD

[@FesZElectronics]

The main part of the 50ohm's is the external termination; so for the basic LISN, without an attenuator, you will get the right impedance only when you add a fixed 50R termination or you connect it to a device that has 50R input; on the other hand if you want to add an attenuator, that has to ensure both the voltage division rate, but also to maintain 50R on the LISN side when the load (a spectrum analyzer or a 50R termination) is connected after the limiter. I guess the issue is somewhere related to this...

[@AbdulBasitMirza-h2v]

Comprehensive and informative video!!. Just I have question on insertion loss. Normally insertion loss is defined as the ratio of voltage at output without the component (LISN) to the voltage at output with the component (LISN) present. But in the video the insertion loss is measured without measuring the voltage at output 50 ohms terminal without LISN. Instead, the ratio of input terminal to output 50 ohms terminal voltage is measured.

[@R2AUK]

As always, very educational. Thanks!

[@Ghostpalace]

Amazing! Thanks for sharing your knowledge

[@ОлегМуравьев-з7е]

Now we are also building a 5 µH LISN for a voltage of 800 V and a current of 100 A. The most difficult issue is the inductor. When we achieve acceptable results, we will share the implementation with you. In your opinion, is a flat coil on a printed circuit board or the same flat coil made of a thick conductor suitable?

[@FesZElectronics]

At 100A you will have quite a lot of trouble with PCB inductors. I think the thick conductor will be better; in the end other than the rated inductance - which is dependent on the number of turns and physical dimension, you care about inductor ESR and self heating. The pcb coil will have quite a bit more esr than the thick wire so it will heat up more.

[@ОлегМуравьев-з7е]

@@FesZElectronics Yes, we decided to start with a coil made of a thick conductor. I think the biggest problem is resonant frequency and parasitic capacitance. In any case, we will share our experience. Thanks for your comments and helpful video!

[@FesZElectronics]

I will have to look into this in a bit more detail, but it would make sense that the more compact the coil is the higher the parallel capacitance and thus lower the resonance. So to get minimum capacitance, you would need a single layer multi turn coil. What sort of target SRF do you have?

[@samruthkrishna2947]

Can you please post the reference for derivation of Impedance formula from measured results

[@textorijum]

Vi/(Z||50)=Iout=Vo/(50+Z||50) Where Z||50 is (Z*50)/(Z+50). You get (50)/(Vo/Vi-2) or as he wrote (50*Vi/Vo)/(1-2*Vi/Vo) So I used this formula: 50/(V(out)/V(in)-2)

[@fabrizionelganes1540]

Very good!

[@李博翔-j2w]

@FesZElectronics I am chrious that Why your spectrum analyzer can measure impedance. Could you tell me about it? Thanks!

[@FesZElectronics]

It cannot directly measure impedance, but you can derive the impedance based on the signal value - with a 50R port, running into a 50r load you get -6dB attenuation (incoming signal is at half value). In a similar fashion you can calculate the exact unknown impedance knowing the port impedance and the achieved signal level.

[@dimitarzhekov9550]

Can someone explain how he got the equation for the Input impendence ? Thanks :)

[@canaltiago]

Muito bom! Parabéns!