LTspice tutorial - SMPS EMI and electrical noise and filtration simulations

#42 #ltspice
In this tutorial video I look at various ways to simulate most electrical noise generated when a switch mode power supply operates in real life. And after its all modeled it needs to be reduced so a look at filtration methods is also studied. All these using some LTspice simulations of course!
Here is a link to the simulation file if you want to play arround with it: drive.google.com/open?id=1M8e...
Other relevant data:
www.ti.com/seclit/ug/slyu036/...
If you liked this video be sure to check out my other videos and you can also subscribe to be up to date with all the new ones!
If you want to support the creation of more and better videos you can at: / feszelectronics
Tutorial series
EP1-Getting started:
• LTSpice Tutorial - EP1...
EP2-AC simulation and the Baxandall tone control circuit:
• LTspice tutorial - EP2...
EP3-.param and .step directives:
• LTspice tutorial - EP3...
EP4-How to import libraries and component models: • LTspice tutorial - EP4...
Ep5- Stock example simulations to play with: • LTspice tutorial - Ep5...
Ep6- Basics of FFT analysis and .four statment: • LTspice tutorial - Ep6...
EP7- Dependent voltage and current sources
• LTspice tutorial - Ep7...
Ep8- Turning a datasheet into a component model
• LTspice tutorial - Ep8...
EP9- Independent voltage and current sources
• LTspice tutorial - Ep9...
Ep10- .wave statement and audio file processing
• LTspice tutorial - Ep1...
Ep11 The effect of temperature on your circuit
• LTspice tutorial - The...
EP12 Modeling a DC brushed motor
• LTSpice Tutorial - Mod...
Ep13- Worst Case, Monte Carlo and Gaussian statistics circuit analysis
• LTspice tutorial - Wor...
EP14 - Digital circuits and logic gates • LTspice tutorial - Dig...
EP15- SMPS EMI and electrical noise and filters
• LTspice tutorial - SMP...
EP16- Feedback loop simulation
• LTspice tutorial - Fee...
Ep17 - Verifying Simulation Models
• LTspice tutorial - Sim...
Ep18 - Simulating Capacitors
• LTspice tutorial - Sim...
Ep 19 - Simulating Inductors
• LTspice tutorial - Sim...
EP 20 - Noise simulations
• LTspice tutorial - Noi...
Ep21 - Models and Libraries
• LTspice Tutorial - Cre...
Music:
The Builder by Kevin MacLeod incompetech.com
Creative Commons - Attribution 3.0 Unported - CC BY 3.0
Free Download / Stream: bit.ly/_the-builder
Music promoted by Audio Library • The Builder - Kevin Ma...

[@ashwinr3469]

This video is a gem, It is so close to the practical problems I have faced programmable buck & boost converter

[@nikhilvaliveti5717]

This is wonderful. It's really helpful. Great conversion from idea to real practical circuit.

[@Analog_isnt_dead]

Great video fesz!

[@yaghiyahbrenner8902]

wow another gem! In the coming months I'm putting together parts for an class-d amplifier + matching offline SMPS I will turn to some of these simulations for help modeling switching behavior and noise reduction.

[@FesZElectronics]

Sounds like a great project! What topology of supply do you plan to use?

[@yaghiyahbrenner8902]

@@FesZElectronics For the PSU half bridge ~800Watts, using a TL494 or SG 2535.. I would like to in the future learn more practical and theory on resonant typologies but I don't have the confidenceand time, so im going for low hanging fruit. (although on semi has some nice chips for resonant typologies and they use less parts). for the amplifier x6 channels class-d ucd amplifiers half bridge self oscillating producing ~200watts each and one subwoofer version 500watts using a irs20957strpbf. i have lots of dsp chips so im intending to run them all through a dsp. I'm waiting on parts, its a big project hopefully i can document my experiences.

[@yaghiyahbrenner8902]

On the UCD amplifier topic its rather interesting so many modulator topics, the math is hard it boils down to compensation networks and pole/zero placement.. I have no experience in this but I find it very intriguing :D

[@yaghiyahbrenner8902]

On the PSU topic Magnetics is key, and for resonant converters it will mean success or failure it less forgiving than a half bridge design.

[@FesZElectronics]

Sounds like a massive undertaking, but also a lot of fun! I wish you the best of luck and hope all the circuits work just as planned!

[@cyberphox1]

Fantastic video!

[@MJ-vt1kp]

This is fantastic! Thank you.

[@FesZElectronics]

I'm happy you found it interesting!

[@narasimha4660]

Good Information 👍

[@KidCe.]

This is soo damn helpful! Thank you so much for this awesome content!

[@malinbay]

great work

[@my_home_interiors_hyderabad]

Very helpful

[@minazulkhan8287]

Your new subscriber.. great videos

[@dimuthudissanayake4006]

How did you design the LC low pass filter how to calculate those inductor capacitor values

[@vickramreddy64]

Can you model IC like (TL494) , because it is difficult to find a spice model for some Ic's.

[@jakubladman5113]

Hi. Actually there is much better way to speedup the start than with the diode and additional voltage supply (which can add noise or other problems by the way). Make the value field of the output capacitor say "10u ic=5" then the simulation will start with the 10uF capacitor charged to 5V, observe the polarity of the capacitor in that case even if the capacitor is non-polarized type (see the current arrow when the mouse cursor is over the component). This technique is also useful to starting with current already flowing through an inductor (ic=1 then means one amp)

[@FesZElectronics]

In the mean time I also discovered this statement; you can use it stand alone (not linked to a specific component) like the form ".ic I(L1)=1" to set a 1A current trough a component or ".ic V(n001)=1" to set a voltage in a net.

[@CARlosDAN783]

Thanks for sharing the simulation, I have seen many of your videos and you almost never share the simulations. Thanks again.

[@FesZElectronics]

In general all the simulations are available for Patrons, above a certain tier.

[@CARlosDAN783]

Ok and how is the way to go to the other level. Can you explain me.@@FesZElectronics

[@thezodiace7399]

Hello Fesz, Great video. One thing that I noticed on your videos when using LTspice and working with multiple signals is that you plot a lot of signals in the same time in the same plot and you speed up your explantion not giving us the time to process what we see and what you say.

[@FesZElectronics]

Thank you for this observation. This is quite an old video though and in the mean time, I do try to use multiple plot planes to highlight the waveforms more clearly and I did change the background to white, to make things a bit more visible. Let me know if you think the recent improvements make visualizing the waveforms better.

[@edgarbaron3797]

Could you share the schematics of all the simulattion with ideal components and when you started adding real components and all parasitics

[@echohu5126]

Great thanks for the video. Only a small comment, the resistance in the LISN should be 50 Ohm or 50Ohm//1kOhm, instead of 1kOhm.

[@FesZElectronics]

Indeed, you are right. I forgot to add the 50R termination... Thank you for pointing that out!

[@thezodiace7399]

Hello Fesz, I have a couple of questions regarding the LISN circuit: 1) Why did you add he LISN to the circuit? is it supposed to be on the final circuit or it is just added to simulate the behavior of the circuit when conducting CE testing? 2) You probed the noise on the input after you added the LISN, why is that, I thought it should be the other way around because the LISN will add some filtering on the input so we won't be able to see the real noise.

[@FesZElectronics]

Hello! Well, I wanted to measure the noise the circuit is generating into the supply line; if I would have added an ideal voltage source, there would be no noise, since the impedance would be zero. The main reason for adding an LISN (other than as part of the CE test setup) is that it provides a known non-zero impedance to the line, so any noise can actually be measured. Regarding filtration, the LISN is filtering noise coming from the initial battery (to prevent it going into the circuit) and its preventing noise from the circuit from passing trough towards the battery; but it does not however prevent the noise coming from the supply from appearing in the supply side of the lines.

[@siddhikadam3092]

Can you make a video on the common-mode and differential-mode equivalent circuit modelling of a simple DC-DC converter like Buck?

[@huseyinkahraman2374]

hello, we are trying to design an EMI filter in common mode and differential mode to prevent 50kz noise. LT spice is a SMPS circuit that we try to filter. but we cannot simulate the circuit noise in ltspice. what way should we follow

[@FesZElectronics]

Hello! I think that you should not simulate the SMPS at all, and just simulate an AC source and see the effect your filter has on a broad spectrum. So for example put the AC source from 1KHz to 1MHz and fine tune the filter to filter not just the 50KHz but also harmonics. To simulate both CM and DM you can use 2 AC sources, one between the terminals (if its an AC line then N and L) for DM and for CM use a source from PE (protected earth ground) to both L and N trough DC isolating capacitors.

[@shubhamtuli5732]

Hey. Can you please help me out how can i find common mode and differential mode insertion loss simulation in LTSpice? I am designing a three phase AC DC bidirectional converter with EMI filter at input (commom mode and differential mode choke) and want to see the noise at input with LISN netwrok in LTSpice. Please guide me how can i simulate it.

[@FesZElectronics]

I would suggest building the LISN in LTspice and measuring the output using the FFt method (one LISN for each supply pin). Also you can add stray capacitance between each circuit node and ground using .options cshunt=stray_capacity. This will help to generate the common mode current and have it uniformly distributed troughout the circuit.

[@incxxxx]

I have a question.May be you know how to simulate in LTspice aku to be charged. For example I have charging circuit in LTspice and I want to use it for charging some aku. How to simulate it in LTspice?

[@FesZElectronics]

Depends on the battery parameters you are interested in. Since I don't think you want to simulate 30 min or more of simulation time to charge the battery you wish only to see how the charger limits current and turns off. If this is the case, the most simple way would be to just to put a capacitor with a few farads (or more, LT doesn't care how big it is) of capacitance and a small series resistor. Some more complex articles I found are: www.slideshare.net/TsuyoshiHorigome/nimh-battery-ltspice-simplified-model and www.semanticscholar.org/paper/Accurate-electrical-battery-model-capable-of-and-Chen-Rincon-Mora/38febe6b89738ca05095f50f68709e985e14829a

[@incxxxx]

@@FesZElectronics Thanks a lot.

[@vmercado5290]

Does LT spice can simulate Hybrid Power Filter to Reduce Harmonics of Non-Linear Loads ? Hoping for the answer. For my thesis thank you very much!

[@FesZElectronics]

You should be able to simulate any circuit as long as you have the patience to wait for the simulation to conclude.

[@Tommybotham]

Hello? The voltage Pulse which drives the MOSFET, why is it not connected to ground? Instead, it feeds back into the circuit? Why is this? I spent 2 hours creating a Buck Converter from Texas Instruments/Power Topologies with that gate voltage pulse going to ground at the negative side, always used up 50% of my output voltage from Vin (which I don't want). As soon as I changed to your style, it works. But I don't understand why. Thank you for the video.

[@FesZElectronics]

Hello @Da Bx , In this circuit I am using a Pchannel mosfet, so to drive it I need some sort of circuit that gets the gate pin at a lower voltage than the source pin by at least 5V (I use 10 to be sure I am driving the transistor properly). The exact voltage is transistor dependent. The transistor does not care where the voltage source is connected as long as the gate-drain difference is ensured. In real life for a simple circuit I would drive the SW transistor with a low side NPN and drive the NPN with a signal source (the control IC) - that way I rely on the input power supply to ensure the gate drive voltage; I used the single signal source approach in the simulation to simplify the circuit and make sure I have fast switching characteristics. I used the NPN way in this video: kzbin.info/www/bejne/bmquf517o66qja8 The same approach with a single signal source can be done to simulate driving N-channel mosfet, this of course helps to simplify the circuit to not need the charge pump or other method to obtain the higher voltage needed for the gate. I am not sure why your approach didn't work - if you use the signal source that drives the channel transistor with voltage pulse from 0V to Vin it should work. In your case, when it didn't work, what was Vin and what pulse did you use?

[@Tommybotham]

@@FesZElectronics I'm using an N-channel MOSFET. It worked only when I increased the driving voltage of the gate (Vg) to more than my input voltage (Vin). In this case my Vin was 24 V (standard 24v power) and my Vgate was 30 V. But I'm not sure that I need or want such a high voltage on my MOSFET. So if Vg was < Vin, it did not work and I always lost ~ 50% of my output voltage. It only worked once I: A) Increase Vg to > 24 V Or B) Disconnected the voltage from ground and fed it back into the circuit from the negative side like you did. The pulse had duty cycles from 0.05 to 0.95 and has a 50ns rise and fall time. Again, its very unintuitive to me but I'm still dissecting your explanation. Thank you.

[@FesZElectronics]

@Da Bx Well, there is the problem - you where using an N-chanel mosfet. To drive it, the grid has to have a voltage potential higher than the source by 5-10V. Or in your case 30-24=6V. The transistor never sees more than this 6V accross the gate-drain and the maximum voltage across the drain source is Vin, so its perfectly safe. The only circuit that needs to be able to handle the 30V is the circuit that drives the gate. The real-life implementation for such a circuit is by using a boot-strap capacitor. On a quick online search I found this document: engineering.purdue.edu/Courses/ECE433/exp5_5th~6thweek_.pdf See figure 3 for the circuit and fig 4-5 for explanations. Basically you get more than Vin by charging a capacitor when the switch node is at a low voltage. When the switch node is at a high voltage the voltage on your boot strap is added to Vin to get the higher voltage. This works like a charge pump circuit. And you use this higher voltage as the supply for the transistor driver.

[@Tommybotham]

@@FesZElectronics Thank you for teaching me!

[@MasoomaAli-eh3qo]

Hi, I have a few questions, 1. If I am designing an SMPS with PFC boost, do I need to use EMI filter again for the next stage of flyback? 2. If I only wish to simulate the EMI, do I still need to add all the rest of the components? 3. How to calculate the values for pi filter?

[@udohuhn-rohrbacher1406]

How would you simulate conducted EMI on isolated Power Supplies, for which the secondary Ground is different from the Earth terminal ? Thank you! Regards Udo

[@FesZElectronics]

The secondary ground will never be completely isolated from the primary ground - you can model this connection trough various capacitors - like inside the power transformer and from load to ground again. As the analyzed frequency increases, you will also need to consider the various inductances - the ground is not a uni-potential structure after all... As long as the modeling is an accurate representation of the real device, you should get some idea of the emi that will get generated.

[@udohuhn-rohrbacher1406]

Thank you very much. I will try a FFT for a Flyback SMPS, using small caps between sec.GND and earth. Also, if a headsink with insul.Pad between MOS and Case is needed, a cap to Earth is realistic. Your presentation is very impressive and helpful. ---- Udo

[@FesZElectronics]

With the transistor a drain to gnd cap should be sufficient, but with the transformer it can be a bit more complex - the different sides of the winding are presenting different amounts of noise - on the primary the side connected to the supply capacitor is far less noisy than the side connected to the transistor drain; You can have up to 4 capacitors - from each winding end to each other and the values need to represent the way in which the transformer is built - the winding end on the primary from the first layer has the smallest capacitances; then the highest capacitance is between the primary and secondary - where one ends and the other begins; and finally a small capacitance is on the end of the secondary - the outer most layer. Based on this mode of analysis you will find that EMI noise is directly impacted by the exact way in which the transformer is connected to the circuit - which coil end is connected to which part of the circuit. Hope this helps! Fesz

[@_wave64_]

Something tells me you're Hungarian. My accent sounds the same when I hear my recorded voice.

[@tharungurana4216]

Anyone help me to this