Understanding the Common Mode Choke using LTspice

[@okidoki3515]

Dear FesZ, I don't find a word to tell you how u simplify the problem and explain them in simple term. Excellent , tons of youtubers can not come near you. Go ahead.

[@FesZElectronics]

Thank you for the kind words! I'm happy you are enjoying my videos!

[@wtfucrazy]

@@FesZElectronics I agree. Your doing a great job.

[@Sheevlord]

Thank you for explaining this in such an easy to understand way. I made a noise filter similar to the one shown at 11:02, and now I can use cheapo SMPSs with an audio amp without hearing hissing sounds from the speaker

[@muratacar155]

"Hello, and welcome back" favorite phrase 😄😄

[@pao_MSTATOR]

Great review. Please note that for common mode chokes it is better to use not ferrite, but nanocrystalline material! There, the benefit is really higher for common-mode chokes.

[@kaztros2595]

Thank you! Especially for the very helpful knowledge @ 9:52 for how common mode configuration avoids core saturation.

[@ilyassargul1080]

One of the best comprehensible video about EMC Filter. Thanks.

[@caeserv]

Good Video, it is also required to understand the leakage Inductance from the CMC to the circuit. Your presentation with the comparison of various LTSPICE circuits are wonderful.

[@ilterkoksal8386]

Hi. The first circuit gives different results with different versions of LTSpice. LTSpice XVII gives the same result as your simulation whereas at Version 24, vp swings around 9V and vn swings around 0V. By the way, great channel...

[@Chetankarkera]

Just brilliant !! I always search for different materials to understand electronics fundamentals more deeply ..now I can easily visualise by seeing these videos ! Thanks a lot and keep up your good work 👍

[@FesZElectronics]

Great to hear! Let me know if there is any topic in particular you might be interested in.

[@Chetankarkera]

Thanks for asking ....m just learning ltspice for simulation..! Could you please tell me is there any difference between ltspice inbuilt noise sources and the model you developed? Secondly AC analysis only used in the circuit where pulsating or sinusoidal signals are there?

[@FesZElectronics]

I'm not sure what the "inbuilt" noise source you are referring to; but the model I used is built using the available components. On the other hand, dedicated noise simulations are a completely different thing - I tried to cover this in a different video: kzbin.info/www/bejne/epmYiaB8jdGfoKs Regarding AC analysis, this works by finding a stable operating point and then inserting a minute signal, so this will only work well for linear circuits. For switching circuits (or ones that use pulsating signals) you will not get very good results. On the other hand, there are methods to perform "ac simulations" by running transient simulations. In this video I focused on feedback loops: kzbin.info/www/bejne/jmXcknWtlrRknc0 Hope this helps!

[@Chetankarkera]

When I mentioned inbuilt noise source means I was referring to arbitrary behavioral voltage source in ltspice .! Is it possible to export data points from ltspice worst case analysis and plot histogram in other tools ?

[@pao_MSTATOR]

The fundamental difference between the operation of a nanocrystalline core and a ferrite core lies in the absence of a resonance characteristic of ferrites at frequencies of 0.5÷2 MHz, and work on the absorption of RF interference, in contrast to the reflection of interference characteristic of ferrites, which is associated with a change in the inductive nature of the ferrite choke to capacitive and can create problems in the work of the IIP. Increase in impedance and attenuation by 2 - 6 times, High resonant frequency and wide bandwidth (up to 5 times), Absence of many resonances in the frequency response of the impedance, Reducing the leakage inductance by 2.3 - 3 times, High saturation induction (1.17 T), Excellent temperature stability (in the range from -60°С to +155°С the change in permeability is less than 15%), Low active losses in the wire, Wide operating temperature range: from -60°С to +100°С, Reduction of volume and weight by 2 - 3 times.

[@ArjunAKed22d005]

Imho, if you look at the simple LC filter as shown in the circuit on the right at 3:36, the common mode noise still has a path to ground through C6 through the lower wire of the differential wiring. Hence, this noise comes through. However, the loop immediately to the right of the differential mode source is broken by the inductor at noise frequencies. Hence, you can't see the differential mode noise on that loop. When you add another inductor to the lower wire of the differential wiring, then common mode noise has no path to ground through C6, hence blocking the common mode noise.

[@thezodiace7399]

These guy's videos makes Orcad Pspice look like a complete waste of money, LT is indeed powerful. Great video btw.

[@FesZElectronics]

I think the best tool, in any situation not just circuit simulation, is the tool you are most familiar with. Once you learn to properly use it, all the tools are really good.

[@andresgongoraYT]

I really appreciate your efforts. Great video!

[@andyparadis342]

Very good explanation, removes the mystery for a novice electronics nerd.

[@mcintoda9]

I would also add that the differential mode inductance is due to leakage flux that doesn’t couple in the ferrite. Some CMC employ purposeful leaky core winding to have both CM and DM inductance. Compact power supplies make use of this.

[@26manish26]

In practical case the inductance varies (due to change in permeability) with change in frequency. Can you let us know how to simulate such case in LTspice.

[@sheetalmekhale]

Nice Video, one can reduce the coupling factor from 1 to 0.99 or so, a common mode filter also filters difference mode noise due to the leakage inductance of the 2 coils, the leakage inductance works as the difference mode inductor.

[@FesZElectronics]

You are right SM! The real life CMC does not have a coupling factor of 1. I used that just for simulation purposes to highlight the Common Mode effect. The real CMC has a smaller coupling factor, and this can be seen in the datasheet since there is a differential component also to its impedance.

[@mumbaiverve2307]

Hi Fesz, Do you have any video on selection of chokes ? Thanks and Cheers !

[@williamogilvie6909]

A common mode choke does provide a lot of filtering and also keeps noise that is generated inside the circuit, eg: A DC-DC converter, from going back down the power leads and creating interference in other modules. A common mode choke is also a balun, so, to some extent the differential noise in the circuit is converted to common mode noise that goes out on the power leads. Your Spice simulator probably doesn't show you that. One case where you shouldn't use a common mode choke is when there is I2C signalling between modules or analog levels passed between them. This is especially true if some modules have motors or other high current intermittent loads. With a common mode choke there will be ground bounce. The ground reference level required to read I2C reliably will not be there. I saw an example of this at a medical robotics company. I was not able to convince a manager that The ground inductor had to go. He spent many weeks trying kludge fixes, all to no avail. Eventually a big investor pulled out and life there became very unpleasant. I was long gone by then. The only solutions to this problem are to either use optical coupling, or get rid of the ground inductor. If you have any non-differential signalling between modules, don't use a common mode choke.

[@ryanharriss7950]

Good explanation

[@hichamlhachimi9594]

Great video and the use of LTspice has made it more perfect keep going

[@cyberphox1]

Awesome videos, your channel is a gem

[@FesZElectronics]

Thank you for the kind words!

[@eryksokolowski8706]

Hi can you share a link to the spice model please?

[@scollyer.tuition]

This is very helpful for those of us who know nothing about EMI problems (e.g. me...) However, I have a question: given that there are no lossy components in the LC filter that you show about 3 minutes in, where is the energy in the differential mode noise going? I am somewhat confused by this, and am wondering if I'm missing some obvious point.

[@Impedancenetwork]

You know what you need. You need to got to engineering school. It is impossible to teach you everything you need in a video. Jesus you friggin amatuers just piss me off. Go READ! or Go to college. you would be so better off than being an Uber driver asking dumb questions on the internet.

[@p_mouse8676]

Nice video, as always! Btw, when using FFT windows, it's REALLY important to choice the right FFT windowing function. Very often I see a lot of people (including colleagues) using the wrong type of window for the wrong type of comparison. In this case for example we are mostly interested in the amplitude difference, not so much in frequency accuracy. So a good window function would be something like a flattop. Using the wrong type of window function can quite often really lead to wrong conclusions.

[@FesZElectronics]

Hello Piet! I wasn't really aware that you could select the way in which LTspice analyzes the FFT data, and just how many options there are. Thanks for pointing this out!

[@Marinho0034]

Nice comment! Do you have any material to point out which window should be used at what occasion? As you are essentially multiplying the acquired signal with the function in the time domain (a convolution in the frequency domain) I suppose that your windowing choice should be influenced by it's frequency counterpart...

[@alexlo7708]

@@Marinho0034 Seems window choice complicated with filter band.

[@Marinho0034]

@@alexlo7708 Yes. Tho only 1 year has passed by I've gained some more insight into this problem. As of my current understanding, every ideal filter can be described with at least 6 design metrics: pass band/rejection band attenuation/ripple/cutoff frequency. Each type of filter tweaks each parameter a bit. If you go to Matlab/Octave and plot the frequency spectrum of each filter type and compare it to the "ideal" filter such parameters become more obvious. Now for the difficult part. The better you know the signal that you want to filter out, the more obvious your filter choice becomes. So when someone asks which filter should he/she use, instead they should be asking about the morphology of their "noise"/disturbance and the morphology of their signal of interest. Hope I gave some good insights!

[@richardneifeld7797]

Insightful, thanks.

[@ashokchakravarthi87]

Nice explanation 👍

[@baharehadokht]

Thank you for the great explanation. I am wondering why you did not compare the common noise for the voltage between (l2& load2 ) and (L4 & load 3)? Why only comparing Vgnd_llc & Vgnd_cmcc? For the common noise, you measured V+ & V- in the 1st circuit, so I was expecting you to compare these two values after filtering with the common choke. I measured these in my circuit and it is smaller when inductors are not used as a transformer! Please advise. Thank you.

[@FesZElectronics]

Hello Bahareh. When I measured V(V+,V-), I was measuring differential noise - the noise that one line shows in regards to the other. With CM noise, I look at what noise is in both lines in regards to the earth ground (the simulator ground in this circuit). In principle the CM noise should be identical on both lines, in regards to the earth ground. When you use inductors on their own (not as transformers) you get a better differential noise filter (V(v=,v-) is lower). When the inductors are used as a transformer, the filter works on CM noise (V(v+,v-) is not really affected only noise in reagards to earth ground is lower)

[@baharehadokht]

@@FesZElectronics thank you so much. much appreciated.

[@gtcollection6933]

Wonderful video, concise explanation. Thank you

[@anoimo9013]

6:40 Not quite sure why does coupling the two inductors give better attenuation for common mode.

[@FesZElectronics]

The coupling is usefull to cancel the differential inductance; in general, the differential current is very large compared to the common mode one, so any inductance that interacts with the differential current needs to be able to not saturate - this limits the realistic values that can be achieved - usually 10-100uH; with a CMC, since it only interacts with the common mode current, which is very small, inductance values can go in the 100uH-22mH; such large values would not normally be feasible as uncoupled inductors which do not saturate.

[@danganh870]

Brilliant explanation! Nice work!

[@FesZElectronics]

Glad you liked it!

[@Alex-el6cu]

Great video. I am just starting to learn how to use LTspice and how to model a common mode choke. Just a quick and maybe silly question. Why do you include a 1Ohms resistor in series with the noise sources?

[@FesZElectronics]

Honestly, I don't remember. I guess it was to simulate wire resistance. Anyway, the simulation should give the same type of results regardless of those resistors.

[@damiendorleac3446]

Great video ! Very educational. However I did not understand what is the usefulness of capacities close to the CM generator and why did you choose these values (100 pF) ?

[@FesZElectronics]

Hello Damien! Well, the point of using the capacitors is first of all to create galvanic isolation - there should be no DC common mode current in most real life applications, only AC. There are 2 capacitors so that the CM current is inject uniformly to both + and - terminals - both of these terminals should normally have the same CM noise. Regarding the value, there is no fixed value you should use, this depends on the application. A real life example would be an AC-DC supply, and the capacitors represent the transformer primary - secondary capacitance and any other capacitor that is placed between input and output. I recommend kzbin.info/www/bejne/jJzHpXiJfpt2bZY in which I try to filter some real life CM noise.

[@damiendorleac3446]

@@FesZElectronics Thank you for the explanation and your reactivity ! I did not know your video but they are very interesting

[@TheCalvinSkinner]

what about ferrite chokes?

[@rafaelvisintin5981]

Very nice video! Thank you very much!

[@placiduzzu]

Căutam ceva despre leakage in cazul common mode chokes, cum anume se calculează in cazul custom made ... Felicitări pentru canal din București!

[@FesZElectronics]

Salut! Mersi mult! Legat de leakage inductance, nu sunt singur cum se poate calcula; conteaza mult tipul de miez si efectiv cum este realizat bobinajul. Cei de la Kemet au o serie de produse dedicate special pentru a avea rol dublu - "dual mode chokes" dar nu dau prea multe detalii despre cum anume sunt proiectate; cel putin nu am gasit eu nimic.

[@hedleyfurio]

after a few hours I cant seem to get the generic noise displayed with the 2 frequencies - suspect some tricks in the setup ?

[@FesZElectronics]

If you want, you can give a few more details about which circuit you simulated and where exactly the measurement point was so we can figure this out.

[@hedleyfurio]

@@FesZElectronics many thanks for response . All working , it was my lack of knowledge of LTspice especially setting scales on the fft that caused confusion . I watched all your lt spice tutorials which were a big help . My final step is to figure out how this chap did the s-parameters in ltspice. kzbin.info/www/bejne/iZiagnVmeZqUqbM. Many thanks for your excellent channel as I work toward understanding why rotating a poe lan transformer eliminated all the problems experienced when the cable length got longer than 50 meters . Basically common mode choke was moved to phy side from lan side - the investigation continues .

[@BassFenderJazz]

Thanks for this video. What is the purpose of existing these two capacitors: C1 and C2 in generator section? why 100pF?

[@FesZElectronics]

@BassFenderJazz - The CM noise soruce needs to be coupled uniformly into the 2 lines - that is why there are 2 capacitor. The reason they are capacitors is that the CM noise needs to be AC coupled (there is no DC component in CM noise). The value of 100pf is arbitrary - it can be any value and it depends on the real life noise source. I went into more detail on CM and DM noise - how to generate it in real life and then separate here: kzbin.info/www/bejne/qZe3p2N6iLmqhpY

[@bhavnatalesara5555]

Appreciate your efforts. I have a query. On what basis frequency for CM and DM is noise is selected. 1MHz for CM and 500k for DM.

[@FesZElectronics]

In this video I just chose some random frequencies, something to highlight the effects of the filters; but in real life you need to measure the exact noise of your circuit; there is no standard value, it will depend on the noise source.

[@VarunPandithurai]

Awesome brother Thanks for the Clarification

[@PrinceWesterburg]

Superbly presented! I'm interested in using these to filter ultrasonic noise off audio signals.

[@Impedancenetwork]

You could have learned this in engineering school. The information contained in his videos is well known in the industry. You think this is amazing when it has been known for 40 years. Technology always imporoved which improves the techniques but still . He is talking about techniques that have been around since before you were born.

[@joemulhern]

Thanks for the video. One question: You measure the "stray" inductance of 500uH. If you were to implement this in a schematic, would you put 250uH in series to the entrance and exit of one side of the Choke?

[@FesZElectronics]

It should not make a difference if the 500uH is a single inductor on one of the terminals, or if its split up on the 2 wires going into one of the inductors. Its important though to mention that each of the windings has stray inductance; we can assume the same percentage of each inductor as being stray. so each inductor in a trafo needs to have its own stray inductance.

[@markobrelak2582]

Very well explained, good job!

[@researchandbuild1751]

Wouldnt a common mode choke work better if the phase was opposite on each side? It looks like your simulation has the same phase on each side of the transformer, unless i am seeing it wrong? That way the signal would cancel itself out. I thought a common mode choke was opposite..where we cancel out the currents of the common mode...ah i think i was wrong? Im still confused lol

[@FesZElectronics]

Well, the currents are of opposite phase for differential noise or signal - the whole point of the CMC is that it does not affect differential currents since they cancel out in the windings. By "canceling out" the effect is that the CMC does not affect the flow, differential currents are not affected - this is important when there is useful differential signal - like in communication lines. The CMC presents inductance, and thus opposes the flow only of current that passes the same way trough both lines.

[@satish044]

Common mode choke is used for cancelling come mode noise or differential noise? Am bit confused,pls explain.

[@tayegebrewold9896]

Job well done and thank you!

[@abhilashpatel1809]

thank you very much, extremely helpful video

[@sergnova528]

WHAT CHOKE GOOD FOR Ld1 & Ld2 IN AC 220V FILTER ? KoolMu OK??

[@robinwang6399]

That explains why my huge low pass filter didn’t help in creating smooth DC😂.

[@peterpanhelllojhon]

Such a great explanation. Thanks man. Wow.

[@FesZElectronics]

I'm happy you enjoyed it! Cheers!

[@qzorn4440]

great video... just in time...Hmmm...:/ for a HLK-pm12 datasheet power supply project... thanks a lot...:)

[@FesZElectronics]

Interesting! I guess that product already is (at least partially) compliant on the AC side with the legal EMC requirements, according to the datasheet. Are you working on filtering the DC side?

[@qzorn4440]

@@FesZElectronics Hi, the data sheet states: Products designed to meet UL, CE safety certification requirements. Hi-Link gives a 12DCV output example using a 220uF cap. So, i went with that, at least for now. any suggestions would be very helpful. thank you.

[@FesZElectronics]

@@qzorn4440 Well, it all depends on if you actually have a noise issue, there is no point in fixing something that isn't broken. If you do have an issue, I guess you could add an extra LC filter at the end, it would only help with differential noise, not CM though. See L2 and C11 on fig 14 of www.power.com/sites/default/files/product-docs/tny274-280.pdf

[@vikrantsharma9599]

How Can I reduce Noise 18 dB less at 440 kHz

[@MrFarenheit07]

Great video!

[@zariazara2093]

Exilent......

[@technicaltransistor2658]

nice

[@alexinal8514]

Thank you👍👍👍

[@patjohn775]

Thanks!

[@Arijit_VU3ICT]

Wow sir

[@Marinho0034]

Just like to comment that the mic you made at the "beginning" of your channel is so low noise that i can clearly hear you swallow your saliva, hahhahhah. Both weird and awesome because at least I know that the circuit works like a charm! Keep up with the GREAT videos!!

[@FesZElectronics]

I'm not sure I wanted that to be heard, but I guess that was supposed to be a compliment. I might do a follow up on the microphone video at some point - I'm still using a "home made" one its just that its a bit different from the one in the old video; I cut off all the extra components and its working somehow better. Thanks for the support!

[@lanchanoinguyen2914]

noise is so annoying and analog electronic is much more difficult than digital electronic.Sometimes it's out of my understanding to prevent the circuit out of types of noises and it's so terrifying.

[@supul32]

Great video (Y)

[@SamCPower]

Please note, modelling a common mode choke as an inductor is a huge over simplification. Common mode chokes have parasitics that are very significant to their insertion loss characteristics

[@zoozolplexOne]

cool !!!

[@irishguy200007]

Most electronic concepts are not well understood especially by myself.

[@electroumit]

Don't show yourself please. I constantly have to move forward. We need circuit not you. :)

[@dolbodolb6173]

why do you keep interrupting your ltspice presentation to insert the picture of yourself every few seconds like some kind of a narcissistic model girl? that is so distracting.

[@GuitarNewby]

You show way too many commercials. Not worth watching this channel 👎.

[@JohannY3]

It is extremely distracting to try and study a circuit when you constantly switch between your face and the circuit. I watched for a while and then my irritation overcame my curiosity and I stopped watching.

[@jiajunguan8675]

could you please not see yourself? we need to see the circuit and think about it.