common mode choke to pass emc test

Go To Last Post
35 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When doing an EMC test the PCB failed the conducted emission test on AC lines. (it is the external AC/DC converter that is the problem we discovered using dummy load with AC/DC, but I can not change that AC/DC converter because of different reasons)

When we tried to put a common mode choke the design passed the test (even thou we put choke between PCB and AC/DC converter. which feels like the wrong "side" for me since the measure is on AC line... But it worked..I guess some stray capacitive coupling or something..)

but here is my problem: the common mode choke we used  (732-1448-ND digikey) for the test that it passed with, did not have enough current rating (we need 5 amp). 

 

So my question: now when I need to implement this improvement how do I pick a choke that will work?

what is the important value to look at? (except for current rating?)

inductance? DC resistance? Attenuation?

Do I need a freewheel diode when I am using a common mode choke together with a fuse?

why are they not using standardized sizes for the chokes so one can make a design and then just swop and try different ones when EMC-testing?!..that they have all different footprint makes it quite hard..

 

I feel like it a bit of a chicken and the egg situation: I need a PCB to test in EMC lab to see if it is ok. But to make the PCB I need to know what choke to use..and for that I need to test to see what choke to use...

sorry for maybe some stupid questions, but my first time in EMC-test and first time using CM chokes..

it is just that we tried some different common mode chokes in the lab. and the result was quite different for all. some even much worse than with no choke. for me (as a beginner in this field) it feels a bit like black magic =)

 

conducted emission on AC lines:

conducted emission

 

 

old desin used in emc test

 

new design. tried to implement improvements. what you think?

Last Edited: Sun. Apr 28, 2019 - 11:32 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
When doing an EMC test the PCB failed the conducted emission test on AC lines. (it is the external AC/DC converter that is the problem we discovered using dummy load with AC/DC, but I can not change that AC/DC converter because of different reasons)

 

I'd be finding out why the psu is generating all that noise into a dummy load. I gather the dummy load is a resistor? Does the power supply generate that noise on too small a load, too great a load? You need to be very scientific when doing your tests otherwise you'll send yourself around in circles. Ask the power supply manufacturer about it - assuming the manufacturer has claimed it passes the required tests. Otherwise I'd be looking at another power supply - even though you don't want to change, you might be forced to.

 

I can't see why you'd modify your pcb when you've failed on the dummy load. Your pcb is not the dummy load, so when you come to test with your pcb, the situation has changed. 

 

Yes, it can be bit like 'black magic' - that's why you need to be rigourous. What works on one thing might not work on another. What you think works, may not be working for the reasons you expect.

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When we tried to put a common mode choke the design passed the test (even thou we put choke between PCB and AC/DC converter. which feels like the wrong "side"

You fail to mention--did you try a filter on the AC input side?  That is where you want it for AC conducted emissions.  Maybe you should have put more effort on that side before jumping to the output side.   The current  levels will be lower on the AC side (assuming step dpown)...so filters are easier to come by.

 

See the very end of this link...maybe you can get /rent a LISN  (in fact you can rig up your own) & do some relative measurements...this is somewhat easier than radiated measurements.

https://emcfastpass.com/pre-compliance-testing-guide/

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,

 

Thank you for your feedback!

 

yes, the dummy load is an effect resistor drawing a few ampere(2,4amp).

We have used the power supply already in a very simular, but more simple product(that we develop now at the same time as this product). And for that product passed all the tests. But in that case the drawn current was almost none. So this is one of the reasons why I don´t want to change (it is also extreamly cheap, wide voltage in range -so i can sell it all over the world- and enough current out, DIN rail and 24V. lots of boxes to tick..).

this is the power supply:

https://uk.rs-online.com/web/p/din-rail-panel-mount-power-supplies/7654682/

and yes they clame it passes tests:

"EN61000-6-2 (EN50082-2) industrial immunity level"

Yes I will call them and ask. 

 

Well the testcenter is a bit far away and I will change some other things on the PCB anyway, so I am attempting to filter this now when I am at it anyway. If i could solve this situation with making some changes to the PCB that would be the cheapest and easiest solution in all aspects. But of couse not so easy when I can not try and see what happens since I don´t have a lab. and the error cause is outside my PCB..

 

I have given thought to put put an extern filter between AC line and power supply. my line of thought is like this:

Since this testcenter is far away and expensive I thought I bring this three diffrent posible solution with me:

1.   updated PCB filtering away external noise.

2.   try a few diffrent external AC filters

3.   try a few diffrent alternative power supplys

 

number 1 is my prefaired solution and 3 is least prefaired I guess (or maybe 2).

I try them all and choose the cheapest easiest solution that works. 

that is my gameplan anyway. 

 

 

this is how it looked with just dummy load (2,4 amp in effect resistor):

 

Last Edited: Mon. Apr 29, 2019 - 08:11 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

p.s.

btw how do you guys do if you are intruducing common mode chokes in the design? you make a combination footprint? (so you can switch between chokes to test) or do you have any other solution?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why do you think you need a common mode choke?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
Why do you think you need a common mode choke?

 

well because i did not pass the test. and when i put a choke between pcb and power supply it lowerd noise on the power lines enough so I could pass.

but the choke I happended to have availible did not have enough current rating. So I can lower the noise with a common mode choke. but I can not use that one I used last time because of current rating.

 

Last Edited: Mon. Apr 29, 2019 - 01:19 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
(it is the external AC/DC converter that is the problem we discovered using dummy load with AC/DC, but I can not change that AC/DC converter because of different reasons)
Unfortunate as there are power supplies that will pass EMC.

mannen wrote:
... I guess some stray capacitive coupling or something..)
like a ferrite bead.

A ferrite bead has controlled LRC such that its response is relatively broadband.

Could replace L1 in "old desin used in emc test" with a ferrite bead; can greatly reduce the set of ferrite beads by selecting on maximum DC current, termination style (SMD), and impedance.

A ferrite bead's impedance versus frequency will be in its datasheet.

Because the EMC failure is near to 1MHz, "some" correct power inductor is likely a better fit than a ferrite bead (an order of magnitude greater impedance at 1MHz, 10micro-henry versus a ferrite bead, 5A ratings)

A power supply's response is dependent on the load's impedance (R, L, C)

mannen wrote:
I need a PCB to ...
no as frequencies are low enough; therefore, patch the PCB.

"conducted emission on AC lines:" has the issue at approximately 1MHz therefore inexpensive test equipment.

Do pre-EMC evaluation on the bench before sending the corrected PCBA to the EMC lab.

mannen wrote:
for me (as a beginner in this field) it feels a bit like black magic =)
it's not (recall your electromagnetics class at the engineering college, likely the most difficult class for electrical engineers though the return on the knowledge is worth one's effort)

 


SimSurfing

Using the Spectrum Analyzer [Reference.Digilentinc]

Using the Network Analyzer [Reference.Digilentinc]

Technical Tidbit - March-April 2014

Troubleshooting Radiated and Conducted Immunity Problems in the Development Lab

Switching Power Supply Noise - Magnetic fields - YouTube (4m44s) via Doug Smith - YouTube

 

edit : Analog Discovery 2 [Reference.Digilentinc]

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Mon. Apr 29, 2019 - 04:01 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

First off, I think that mannen may be operating with incorrect or missing information.

 

Particularly if the supply is a switching mode power supply, the conducted emissions can be very dependent on the load current. The fact that it may pass some emissions standard at a low current does not mean that it will also pass at a high current. You really have to look at the test conditions. Thus, it might be fully sufficient with a product that has a low current drain, while failing miserably when powering something that requires close to the supply's maximum current.

 

Second, industrial emissions level is not the same as office/residential emissions level. The latter is more strict.

 

Third, immunity is not the same as emissions. Immunity measures how well the device resists signals from other devices. Emissions measures the amount of signal generated by the device. The OP includes the statement: "EN61000-6-2 (EN50082-2) industrial immunity level" yet shows an emissions spectrum. If that is all the spec sheet says, then it may not have ANY emissions spec!

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
this is how it looked with just dummy load (2,4 amp in effect resistor):
approx 2MHz - an artifact of the SMPS's control frequency versus switching frequency.

That SMPS's datasheet may state the minimum capacitance for the load (though I couldn't quickly locate that)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 the common mode choke we used  (732-1448-ND)

Did you test it with this choke on the AC side?  If no, why not? The current is lower on that side

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

 the common mode choke we used  (732-1448-ND)

Did you test it with this choke on the AC side?  If no, why not? The current is lower on that side

 

no, because even if it works that solution would be to hard to implement in production anyway. ( a coil that need to be solderd on some wires could not be assembled by anyone). 

if would need a ready solution like something with screw terminal and a box in a DIN rail or something if it is to be on the AC side>>I would be a of the shelf bought. (And I did not have accces at the time to any readymade sutable filter. so I could not test that either.. I did not expect this problem before going there).

While if the coil is on the PCB side I could just add it to the PCB and problem fixed. 

Also I did not really have time to check because I was just there 1 day. 

Last Edited: Mon. Apr 29, 2019 - 04:50 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

Jim:

I am trying to get it approved to industrial EMC demands EN61000-6-2 and EN61000-6-3.

And right as you are in point 1 and 3 it still does not solve my problem =)

ay..what a headache. nothing is easy it sthis world it seems. 

Last Edited: Mon. Apr 29, 2019 - 04:57 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:

like a ferrite bead.

A ferrite bead has controlled LRC such that its response is relatively broadband.

Could replace L1 in "old desin used in emc test" with a ferrite bead;

 

but L1 in original design is aferrite bead already. And I have come to understand that ferrite beads have their "god spot" very high (like 100 Mhz). that is to high for this noise that is around 1 Mhz. 

I think a common mode choke is in the correct frequency for this as you say.. 

 

But I don´t know.. all this is new to me. how to chose a choke? how does the inductance for example play a role in the behavior of the choke? higher indutance chokes of better? get rid of more noise?

 

Last Edited: Mon. Apr 29, 2019 - 05:03 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have to ask ... how much would a "better" supply cost you compared to the testing costs associated with the cheap one? You are having to redesign a board, have it constructed, then retest, all this maybe several times. Add in your time and the delay to market cost. So, if a supply that works, out of the box costs you 1 Euro more, you could build a lot of these things and still be ahead.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
I think a common mode choke is in the correct frequency for this as you say..
concur on frequency, disagree on implementation (edit: appropriate power inductor to replace L1 ferrite bead versus choke)

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Mon. Apr 29, 2019 - 05:31 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I would happily pay my way out of this (to a certain extent) and I totally agree. but its not 1 euro. more like try twice as much or more. I have looked at one from PULS for example (CT5.241) as an alternative. 

so maybe 50 euro for original one vs. 100-120 euro per unit for the PULS one.

and also I would get much less voltage in range. from now 180-550 VAC to 380 to 480VAC for the PULS one. 

 

PULS

 

if anyone have any other alternative one that have something like this spec I am all ears:

DIN rail, 24VDC, atleast 3 amp out, "not to expensive" -not hundreds of euros- and have a voltage range that is "good" (so it can be sold to as many places as posible. motor is used with 3phase so that voltage can differ in diffrent countries I guess. ).

Thou the voltage range maybe is something I have to give up on. 400 for Europe and 480 for USA maybe is good enough(?). (any american have any input to this? 480VAC is mainly used in USA, right?)

(I don´t know if to have between neutral and phase or between phase and phase is better. I guess it depends on the power supply I find)

Last Edited: Mon. Apr 29, 2019 - 05:52 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In residential and light industrial, U.S. has 120V and 240V. You only get higher voltages in medium industrial (or equivalent).  480V is commonly used for industrial lathes, drill presses, milling machines, welders, wave soldering, and such. Heavy industrial can be anything as they use local transformers to get what ever is needed.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
(any american have any input to this? 480VAC is mainly used in USA, right?)
Maybe yes for light industrial and residential (maybe not for feeds for industrial zones or neighborhoods); 600Vrms to 800Vrms for oilfields (relatively long feed to well head)

Delta Lightning Arrestors - Specifications

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Those things are expensive! (100v - 500V input, 24V@3A out):

 

https://www.newark.com/schneider...

 

https://www.trcelectronics.com/V...

 

But my point still stands. You quote a susceptibility standard, but that is not emissions. Does your supply even have a conducted emissions specification?

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


 

I think putting a filter on the output side is simply fooling yourself (when the the test resistor load was not a culprit) for AC line conducted emissions.  Pass now, mysteriously fail later. Now, that is not to say you can't ALSO have filtering on the output side for whatever noise-producing loads you hook up.

Note that AC line conducted emissions don't have to only pass through the power supply, those AC lines could pick up noise from your output leads (especially if all the cables are bundled together).  So it is not a simple problem.  Even moving the wires around can affect things.

 

You should open up a supply & see what kind of filtering they are applying on the AC side--might give some ideas/insights.

 

Can you rent/buy a LISN & spec analyzer(or maybe you have one)? That way you can try several arrangements on a comparative basis.

 

By the way, was your power supply earth grounded?

 

Seems like something like this could be easily wired to the input side of your supply...no extra box needed (though I don't know details of your install)

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Mon. Apr 29, 2019 - 06:55 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Disclaimer:  I'm not a TVS expert...

 

When I first looked at the schematic, I saw the Fuse and the diode, CR1.

 

I (incorrectly) assumed that that was a poor man's reverse polarity protection circuit.

If someone connected the DC power backwards, the diode conducts and the fuse blows, preventing the reverse voltage from reaching the (expensive) PCB.

 

It appeared that the Fuse has a 31 A rating, or perhaps that's a part number.

It isn't clear from the schematic. (F2 has a number, but it ends with "A" for amps).

 

So, my first thought was that the diode would likely fail open circuit long before the fuse, at 31 A, would (slow) blow.

At which point the reverse polarity circuit would have failed.

 

So I looked up the CR1 diode and it is actually a TVS, not a simple diode.

I think the schematic would be much clearer if it said TVS next to it, or if it used a Zener diode symbol, or some other TVS symbol.

 

So, the TVS data summary sheet I looked at didn't list the forward conduction current limit.

Perhaps it is in a full data sheet.

 

It is, though, a unidirectional TVS.

 

So some lightning and other power line transients will get clamped, but not others.

 

So, question to everyone, what is the "best practice" for input current conditioning these days?

 

And question to the OP, is there any reverse polarity protection for the PCB, (perhaps within the PCB's power supply)? 

 

JC

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
By the way, was your power supply earth grounded?

Installation Manual | WDR-120-MEAN WELL USA Switching Power Supply

(page 3)

...

(9)The FG () must be connected to PE (Protective Earth).

...

via WDR-120-MEAN WELL USA Switching Power Supply (others tab)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC wrote:
So, question to everyone, what is the "best practice" for input current conditioning these days?
PolyZen Devices for Overvoltage-Overcurrent Protection - Littelfuse up to 15Vdc working and 1.3A hold for 15V.

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC wrote:
So, my first thought was that the diode would likely fail open circuit long before the fuse, at 31 A, would (slow) blow.
A diode fails short except for very excessive current (some lightning, line cross due to mis-wire during installation) (over-pressure, encapsulation splits, die is obviously damaged or lead wire vaporized)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So, my first thought was that the diode would likely fail open circuit long before the fuse, at 31 A, would (slow) blow.

31 amp fusing seems rather quite excessive for a 5amp supply...in fact, supply may limit rather tightly around the 5amp max limit (with some transient allowance)...any transient allowance can be taken up by a slo-blo (like use a 6amp slo blo).

 

At really cold temperatures (like outdoors), the tranzsorb trip level can fall a bit, so you have to be careful using a 24V part to protect a 24V supply (Vtripmin, 26.7V @25C) .  Toss in, say 2 to 5%, for the power supply variance riding high, and poof.  I usually get at least the next Voltage level up, if I'm just trying to keep the zapps out (no need to trip  right at the edge).

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A diode fails short except for very excessive current (some lightning, line cross due to mis-wire during installation)

Exactly, if one connects the power supply backwards, then the reverse polarity protection circuit would kick in, or fail.

I assumed that give a direct short with the forward biased fuse across the power supply, that would be sufficient to cause the diode to fail as an open, perhaps with the physical release of smoke and a newly formed physical gap between the electrodes...

 

As AVRCandies pointed out, for a 5A circuit 30+ amps seems a tad bit on the high side.

 

Sorry for the off topic, as that has nothing to do with EMC testing, but it caught my eye.

 

JC  

Last Edited: Tue. Apr 30, 2019 - 02:09 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC: it is a fuseholder for glas fuse, not a 31 amp fuse. I was thinking put something like a 5 amp fuse or simular. 

fuseholder

 

my thought was that the combination fuse and TVS would 

-overcurrent =fuse

- high overvoltage event= TVS starts to leading and fuse trips

-reverse polarity= fuse leads in the forward direction and fuse trips

 

But I am not sure if the fuse burns first or the TVS.

Peak Forward Surge Current, 8.3ms =300 Amp for TVS. 

But I don´t know continues forward current. 

Last Edited: Tue. Apr 30, 2019 - 08:35 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:

DocJC wrote:
So, question to everyone, what is the "best practice" for input current conditioning these days?
PolyZen Devices for Overvoltage-Overcurrent Protection - Littelfuse up to 15Vdc working and 1.3A hold for 15V.

 

 

but i am using a 24 volt system..

Last Edited: Tue. Apr 30, 2019 - 08:45 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:

But my point still stands. You quote a susceptibility standard, but that is not emissions. Does your supply even have a conducted emissions specification?

 

Jim

 

in the datasheet it says: 

EMC EMISSION:  Compliance to EN55011 (CISPR11), EN55032 (CISPR32), EN61204-3 Class B

 

now what that means I would know if I knew these standards. But I don´t.. I will look into it.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

By the way, was your power supply earth grounded?

 

Seems like something like this could be easily wired to the input side of your supply...no extra box needed (though I don't know details of your install)

 

the power supply was grounded on the AC side, yes. 

 

yes, I am considering putting somethink like this on the AC side.

filter

should solve the problem.. and not to expensive. would lower the max inputvolt to 480 thou. 

I don´t know...but In the same time maybe that is just giving "life support" for something that maybe I should change entierly. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think that filter would be a sensible place to start. It is not uncommon to add these.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:

ka7ehk wrote:

But my point still stands. You quote a susceptibility standard, but that is not emissions. Does your supply even have a conducted emissions specification?

 

Jim

 

in the datasheet it says: 

EMC EMISSION:  Compliance to EN55011 (CISPR11), EN55032 (CISPR32), EN61204-3 Class B

 

now what that means I would know if I knew these standards. But I don´t.. I will look into it.

 

I looked into it. Apparently it is simular demands to what I am trying to approve my product for. 

So it should have worked to use this powersupply.

I tried to contact manufacturer to hear what they have to say. byt they did not answer my email yet thou

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I tried to contact manufacturer to hear what they have to say. byt they did not answer my email yet thou

I think you have heard what they have to say...that's what these Chinese companies often say (unless you order 100000).

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Tue. Apr 30, 2019 - 03:35 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

I tried to contact manufacturer to hear what they have to say. byt they did not answer my email yet thou

I think you have heard what they have to say...that's what these Chinese companies often say (unless you order 100000).

 

hehe.. yeah..you are probably right.. =)

but one can hope..

they say that hope is what dies last.