Common mode chokes are a hoax for SMPS?

[cupoftea]

Are common mode chokes (and their usefulness in SMPS) a bluff?
Because common mode noise is due to the circuit strays, and exact calculation of an exact common mode choke is
not possible (not in the normal timescales of a typical project).
So why don't we prefer an inductor in both Line and Neutral?...That gives a common mode impedance.
OK, you don't have the advantage of field cancellation and thus small size that a common mode choke has...but for <200W
SMPS, the currents aren't that high anyway, so the alternative inductor solution wouldn't be that big.

We are all told that because of mains RCD currents, we must have small Y caps,
and then we need the common mode choke, with its enormous common mode inductance
despite small size.

..But common mode currents are very high frequency, and are likely to couple between the coils of a common mode choke anyway
, and couple through the coil capacitance too. This degree of coupling would be less with separate inductors in line and neutral.

Surely common mode chokes are a devil to wind for assembly staff. So this means westerners keep relying on the Chinese
to do it for us.....but aren't we led to believe that we need common mode chokes, and so must keep buying SMPS from the Chinese
(because we couldn't possibly do the fiddley common mode choke winding ourselves).
I appreciate their are torroid winding machines, but these are too expensive, and don't tend to get used for common mode chokes.

How much of the "common mode chokery" is bluff?..and how much is real?

I mean....take a look in a <100W offline SMPS...take out the common mode choke...look at the coils...they are messily wound with multiple layers
right over each other and the inter, and intra winding capacitance must be relatively enormous......which most certainly does not help
with high frequency common mode noise filtering.

 

[dick_freebird]

Common mode choke could be for conducted susceptibility
but more likely for conducted emissions. Keep your switching
crap to yourself, like.

So since it's not about the converter (other than how badly you
have to grow input filter to deal with the inductors you just
added) there's probably no real downside to crappy quality
as long as you meet the spectral mask for whatever.

 

[KlausST]

Hi,

My explanation is this.

I see the metal enclosure as a faraday cage. Let´s consider it as a somehow perfect faraday cage.
No matter how you try you can put any electronics (battery powered HF transmitter) inside a faraday cage, you will never get the faraday cage to emit any HF.

To be able to emit HD, you need "something" like an antenna (a wire) to "leave" the faraday cage which can act as a dipole ... as some counterpart to the faraday cage, in the meaning there could be HF voltage between the wire and the faraday cage. Now the combination "wire and faraday cage" is able to transmit EM waves.

Back to SMPS:
Now I see the power supply with a metal enclosure as the faraday cage. It may not be perfetc, but let´s consider it as a rather good faraday cage.
Every power supply has wires to leave the cage. And due to the internal circuitry there will be unwanted HF noise voltage between the cage and the wires.
Thus this system is able to send out noise. It acts like a transmitter antenna. Unwanted.

Now we have the opposite problem as explained in the first couple of lines. We want to stop the noise radiation.
Here comes the common mode choke into play. It makes the dipole system "enclose to wires" high impedance for HF.
The current is limited by this high impedance, thus the emitted power is limited.
The antenna is becoming worse ... which is a good thing in order to stop/reduce HF radiation.

For sure inidividual inductors will work, too, but they are not as effective (size vs impedance) as a common mode choke.

Klaus

 

[D.A.(Tony)Stewart]

CM chokes are essential to raise the attenuation ratio in both lines and both directions, ingress and egress EMI. It may be considered as important as CMRR in Op Amps. X caps provde the diff mode suppression while Y caps divert the CM noise the PE gnd. A good choke offers 50 to 60 dB CMRR with the Y cap.
The grid EMI from induced lightning transient impulses are CM noise just as SMPS can couple CM noise to the grid.

CM chokes are also called Baluns because they BALance Unbalanced wire impedance with respect to ground by raising the CM impedance 2 o 3 decades higher than the DM impedance from high mutual coupling.

It's not a hoax nor is it a trivial choice and in some cases 1 stage is not enough and only spans <= 2 decades of spectrum for effectiveness..

 

[hiphei]

I've confirmed Klaus's explanation also in my own home.. I have multiple HF antennas in my backyard for DX hobby from lf to vhf.. i have 'shortcircuited loops' for the reception of plain magnetic vector-part of the em wave..and i have few capacitive e-field probes too.. and the secret to true DX lies in the house electrical network and smps's found in your own house..exactly how low noise they have.. If the new smps design i bring into my house has conducting noise i immediately see it's spectrum in my sdr.. If i minimize the length between external common mode filter (believe me..i've built many..) between the product and the filter i typically am able to silence them down to below the atmospheric and surrounding noise levels at my site.. So i personally really like the common mode chokes and filters.. ..it has turned out that my whole antenna system is full of common mode chockes..it's also imperative that the transmission lines have common mode chokes since the signal received with antenna is only differential signal and all signal carried by common mode are either leakage (messing antenna directional pattern) or conducted noise.. So..not hoax.. they work well for the purpose.. I've sometimes wished all power supply designer had a hobby of DX listening..

 

[cupoftea]

It's not a hoax nor is it a trivial choice and in some cases 1 stage is not enough and only spans <= 2 decades of spectrum for effectiveness..

Thanks...yes common mode chokes are very good...but awkward to manufacture...needing ususally a torroid to be wound and the coils to be separated.
Unlike say two dogbone inductors which would be easy to wind.

As you say, often two common mode chokes are needed......this is often the case with common mode chokes, since they are wound with so many turns and on top of each other and the HF goes straight through the interwinding capacitance.
So in fact, the extra HF common mode impedance needed (aswell as the lower frequency common mode choke) would be best provided by just using two inductors , one in line, one in neutral...certainly for offline SMPS <200W.

 

[mtwieg]

OK, you don't have the advantage of field cancellation and thus small size that a common mode choke has...

This is basically the reason CM chokes are used in power electronics. The core size can be reduced by orders of magnitude with coupled windings.

but for <200W
SMPS, the currents aren't that high anyway, so the alternative inductor solution wouldn't be that big.

Please show some example, with actual components. I think you'll have a hard time coming up with an implementation where individual inductors aren't much larger or more expensive than a CM choke.

 

[D.A.(Tony)Stewart]

Two separate L's are not the same as a transformer or a CM choke in s-parameters. There are several criteria for choosing the right X caps and Y caps as well as L including the Middlebrook criteria for resonance.

 

[cupoftea]

Please show some example, with actual components. I think you'll have a hard time coming up with an implementation where individual inductors aren't much larger or more expensive than a CM choke.

Thanks, yes, i was given a 150W "sequence switched linear regulator" LED streetlight (it also had a 3W HV Buck bias supply based on LNK304)... this was a total fail of EMC.....was mostly common mode....from 150khz to 7MHz or so it was fail by 20dB or so.
No offtheshelf common mode chokes would fit in the low profile enclosure, so i just used SMD power inductors in power and neutral...they werent that big. It got it through conducted EMC. The inductors were each about 12mmx12mmx6mm(ht).
I also put a 470pF y cap from HVDC- to earth via screw rest ring.

 

[mtwieg]

There are certainly common mode power chokes in such sizes, but you don't give sufficient info to select a component. Need maximum current, dielectric strength, impedance/inductance, etc.

 

[cupoftea]

Thanks, but as you know, for a 240W PFC'd Offline SMPS, the input current in UK/EU is only ~1Arms. So the currents arent large anyway. So the "common mode choke effect" isnt so needed.
As you know, there is quite literally no way of calculating what common mode choke inductance and Y caps (etc etc) will end you up with a certain plot on the EMC graph. So as you know, one cannot say exactly what common mode inductance is needed.

I was not able to find 240vac rated common mode chokes in surface mount that were less than 8mm in height....i still doubt that such exists.
Current rating to 1Arms. The choice was either a custom torroid wind common mode choke for <8mm height....very expensive, ....or just a couple offtheshelf inductors , line and neutral...the latter won the day and worked fine.

 

[betwixt]

The bottom line is that both methods will work but the cancellation effect of a common mode choke makes it a little more efficient but also a bit bigger,

Brian.

 

[KlausST]

choke makes it a little more efficient but also a bit bigger,

I don´t think that "bigger" is true if one considers identical common mode inductance .. and identical current rating.

I just looked for a standard 1.5A common mode choke and came on this: B82722A/J
The 1.5A type comes with 10mH and 240mOhms DC resistance: diameter 22.3mm, hieght 13.3mm
Price: about 2€

At farnell I only find a single inductor specified at 10mH and >1A. It is 38mm in diameter and 26mm in height. Mind: one needs two of them.
Price: 2 x 11€ = 22€

I did not investigate whether the 10mH of the coupled choke is the "combined value". If so one needs 2 x 20mH single inductors to get a (combined) 10mH common mode.


Klaus

 

[cupoftea]

Common mode chokes offtheshelf for offline PSU are very poor. They feed on the fact that engineers just think "More inductance is better", and so they wind loads of enamelled wire all clumped up on top of each other and interwinding capacitance is very high. It would be better if they used 7/0.2mm wire with thick plastic insulation because then the interwinding capacitance would be less, but this is not done.

 

[KlausST]

Sorry .. but I can´t agree with your statements.

There are many different common mode chokes, with different values coming with datasheets, coming with calculation examples

Every manufacturer provides a wide variety of inductance, current.

I can´t detect:
* the more inductance is better
* loads of windings
* on top of each other
* ...

Such statements I expect from ignorant hobbyists.
But you are experienced. Can you give sources of your "knowledge"? so we can validate them .. by their context.

Klaus

 

[cupoftea]

Thanks, i saw one PCB where it was a control PCB for a 3kW 3 phase equipment.
The control board needed to take in all three phases and neutral so that it could do the zero cross detection on all three phases (as well as phase drop detection).
Obviously the zero cross circuitry was very low power...but in order to do common mode filtering, they then took all 3 phases into the pcb, then put all three phases and neutral onto a 4 coil common mode choke. This was a very , very expensive part as you can well imagine. All that was needed was a filter inductor in all three phases and neutral, and that would have given easily enough common mode filtering for the zero cross circuitry. (as well as the y caps)

There was also a 130W offline SMPS on this control PCB...but that just used one of the phases and neutral, and put them through a FWB. That did use a common mode choke.....and in that case , yes, a common mode choke was warranted........but for the zero cross detection circuitry...??.....i dont think so...i am sure all would agree(?)

 

[KlausST]

Hi,

I´m not sure .. are you doing it as trial and error .. or using math and physics?

Klaus

 

[scopeprobe]

Thanks, i saw one PCB where it was a control PCB for a 3kW 3 phase equipment.
The control board needed to take in all three phases and neutral so that it could do the zero cross detection on all three phases (as well as phase drop detection).
Obviously the zero cross circuitry was very low power...but in order to do common mode filtering, they then took all 3 phases into the pcb, then put all three phases and neutral onto a 4 coil common mode choke. This was a very , very expensive part as you can well imagine. All that was needed was a filter inductor in all three phases and neutral, and that would have given easily enough common mode filtering for the zero cross circuitry. (as well as the y caps)

There was also a 130W offline SMPS on this control PCB...but that just used one of the phases and neutral, and put them through a FWB. That did use a common mode choke.....and in that case , yes, a common mode choke was warranted........but for the zero cross detection circuitry...??.....i dont think so...i am sure all would agree(?)

FYI Putting inductors in all three phases would give you additional differential mode filtering and NOT common mode filtering, thats not to say that radiating cables arn't a souce to common mode noise but sounds like your confusing the two.

 

[cupoftea]

YI Putting inductors in all three phases would give you additional differential mode filtering and NOT common mode filtering, t

Thanks yes, we have coupled inductors in all 3 phases as well as neutral...all on the same torroid...a 4 coil comm mode choke.