Earthing the chassis has nothing to do with EMC improvement?

[Faringdon]

Hi,
We all know that the earth connection to an offline isolated SMPS fed system has absolutely nothing to do with EMC improvement.

However, this video says that it does, and appears to proove it with examples..


At 29:09, he shows how his system was once accidentally powered through a 2 core mains cable with no earth. ..The non-earthed system was plagued with noise issues. These were solved by connecting the system with a 3 core mains cable with earth connection. So it does appear that Earth is needed to prevent noise issues after all….it is not just for safety reasons.

However, he refers to a blue “filter” at 14:50. This is obviously a Y capacitor.
At 16:10 he relocates the Y capacitor across the transformer, which is where you expect to see it. But what is the “functional earth” connection from the secondary side of the SMPS to the other board? Surely that shouldn’t be there?

I actually suspect that the shot at 19:10 is  telling us what is really happening. At 19:10 it shows  comms cables connecting to the board. These cables probably contain an earthed connection. Earthing the secondary of an isolated SMPS always worsens common mode emissions...and this is what I think is the problem(?) I suspect that the offline isolated SMPS does not contain enough common mode filtration to handle the situation of the secondary getting earthed by the comms cables. He then needs the extra Y capacitor based “functional earthing” in order to solve the lack of proper common mode filtration in the SMPS.

But what do you think?

[f4eru]

We all know that the earth connection to an offline isolated SMPS fed system has absolutely nothing to do with EMC improvement.

Ah?

PE is earthing for safety.
FE is earthing for EMC.

[Faringdon]

PE is earthing for safety.
FE is earthing for EMC.

Thanks..
Though if we listen to 4:50 to 5:10 of this video below, it says that there is no thing such as "EMC improvement due to Functional earthing", and that earth never plays any role in EMC improvement...

[f4eru]

Never say never.

[Faringdon]

I think the video in the top  post shows that earth does play a role in EMC improvement.
In an aeroplane, the same applies, its just that the plane's fuselage and wings is the new "earth".

You need to have an earth wire coming in for EMC purposes, so that your common mode filter can work with it.

If you dont have an earth wire coming in with the mains cable, then you end up with a large common mode noise problem.
The conclusion, which is prooven in the top post, is that Earth is not just for safety, earth is for EMC improvement.
Would you agree?

I mean, Once I built a crystal set radio. It did not work. I then soldered a wire to the crystal set’s circuit ground, and connected this wire to the earth in a mains socket. The crystal radio immediately worked. It would not work when this earth connection was removed. This shows that earth has some influence on removing noise from circuitry.

[penfold]

At 29:09, he shows how his system was once accidentally powered through a 2 core mains cable with no earth. ..The non-earthed system was plagued with noise issues. These were solved by connecting the system with a 3 core mains cable with earth connection. So it does appear that Earth is needed to prevent noise issues after all….it is not just for safety reasons.

So he successfully proved that EMC treatments for class I and class II equipment are not always equivalent.

In terms of the mains earth in your 3-pin plug, it was only ever designed as a "safety earth". At LF it has a low(ish) impedance to mains neutral (and thus mains earth) and similarly to most other mains earthed equipment nearby.

With regard to noise, the mains earth has some attributes that make it "not-terrible" for helping or preventing noise, it's always run nice and close to the mains conductors, but it would be a massive stretch to say that at several MHz it was anywhere near *optimal* for noise suppression. So, you could draw the conclusion that "Mains earth is primarily intended for safety, but provides some benefits to EMC".

If you consider the scenario where you have a flyback, for example, where there is some amount of capacitive coupling from the primary switching node to the secondary and some amount of coupling to some chassis metalwork. Without a low impedance return path from the secondary back to the switching node, a large common-mode voltage results on that secondary. With regard to the earthed chassis metalwork... is a simple connection to the incoming mains earth sufficient (bearing in mind to return to the switching node, the currents must travel a very long way), will the impedance be low enough to keep common-mode voltages low and make it less appealing for them to travel through any attached equipment? And for noise in that case, what is the point in even connecting to mains earth?

Maybe you could revise the conclusion to "Mains earth is only for safety. Good EMC is the result of good design, planning and analysis"...

[Gyro]

This has been explained to you before. Maybe check your threads library for your present and previous incarnations occasionally.

https://www.eevblog.com/forum/renewable-energy/earthing-the-secondary-of-an-offline-isolated-smps/msg3617090/

[Faringdon]

Thanks, yes,  its interesting, Paragraph 8  of this (from the link you kindly supply)
https://www.edn.com/safety-concerns-for-practical-emi-line-filters/
..says that a chassis needs to be connected specifically to earth in order to improve EMC.

So he successfully proved that EMC treatments for class I and class II equipment are not always equivalent.

Thats a good point, though the chap in the video of the top post is having it that without the earth connection, it was plagued with noise...then when he made an earth connection, the noise problem went away....i would have to admit, that he's right.

I agree that the earth connection is terrible for several MHz, but then again, its better than nothing. And Earth is pretty well the only thing you can use to connect chassis to....the result is certainly that earth wasnt just for safety...but as we can see, it made noise problems less....same as with my crystal set.

If you consider the scenario where you have a flyback, for example, where there is some amount of capacitive coupling from the primary switching node to the secondary and some amount of coupling to some chassis metalwork. Without a low impedance return path from the secondary back to the switching node, a large common-mode voltage results on that secondary. With regard to the earthed chassis metalwork... is a simple connection to the incoming mains earth sufficient (bearing in mind to return to the switching node, the currents must travel a very long way), will the impedance be low enough to keep common-mode voltages low and make it less appealing for them to travel through any attached equipment? And for noise in that case, what is the point in even connecting to mains earth?

Thankyou for your inciteful words.
Also, i think if you have an earth connection, plus eg,  some  Y capacitors, connecting from earth to line and neutral...then any emissions that have unwantedly escaped to earth, can be dragged back out of earth, via the Y capacitors......and returned via the neutral, as is wanted, for EMC improvement.

[penfold]

Thats a good point, though the chap in the video of the top post is having it that without the earth connection, it was plagued with noise...then when he made an earth connection, the noise problem went away....i would have to admit, that he's right.

Oh, I do apologise, I had missed the point there. Though I would be wary about taking information from a video about a specific problem with specific pieces of equipment where the conclusion is "doing this solved it, therefore not doing that was the problem" - it may be a valid conclusion, but only strictly true for that particular alarm system. Given that the example was about some kind of alarm system, there is the whole "electrically long wires" dispersed throughout a building thing going on. In that case, the safety-earth probably does something... but it's a very different beast to what the Altium video was about.

[...]
Also, i think if you have an earth connection, plus eg,  some  Y capacitors, connecting from earth to line and neutral...then any emissions that have unwantedly escaped to earth, can be dragged back out of earth, via the Y capacitors......and returned via the neutral, as is wanted, for EMC improvement.

If you assume for a second that there isn't an actual earth connection to the chassis... the currents can still return via the y-caps and therefore it's still not important that the earth is connected from a noise perspective, obviously necessary for safety. Same with the EDN article, I'm not disputing its statement, but the noise reduction may be more of a consequence of the live and neutral's y-capacitance to the chassis than the chassis' connection to earth.

[Faringdon]

If you assume for a second that there isn't an actual earth connection to the chassis... the currents can still return via the y-caps and therefore it's still not important that the earth is connected from a noise perspective, obviously necessary for safety. Same with the EDN article, I'm not disputing its statement, but the noise reduction may be more of a consequence of the live and neutral's y-capacitance to the chassis than the chassis' connection to earth.

...Thanks, well, yes...indeed,  having the chassis connection to earth will indeed allow the Y caps to do a better job of EMC improvement….so I agree with you, but say that this is an example of a chassis earth connection improving the EMC situation.

In conducted emissions testing, as you know, we  test at frequencies up to 30MHz, and that conducts down line and neutral, so why not down the earth connection aswell?….And indeed, as we know, higher frequencies than this are actually in there.

For any mains SMPS powered circuit, the actual full electric circuit goes back to the power station. Going right back to the power station, there are multiple earth connections of the neutral. As such,  there’s always the possibility of emissions coupling out of  the SMPS, such that the current coming in through the line, does not  go back out of the neutral, and then we have common mode emissions...noise…..and its because the emissions couple to earth, and return via earth, rather than returning via neutral. So if we earth the chassis, then we “invite” the noise to couple to chassis (because it will be trying to couple to the “world outside” anyway)...then having the chassis connected to earth, means that the emissions will be channelled back down the mains earth wire….and in so doing, they will get  attenuated by the common mode filtration in the SMPS.

So yes, earthing an enclosure is all about EMC improvement, as well as safety.

Now, if the chassis wasnt connected to earth.....(and by that i specifically mean a connection via the earth wire that comes into the product along with line and neutral), then you'd get emissions  from the SMPS coupling to the chassis, and with no connection of chassis to incoming earth wire, then the emissions will couple to the "world  around the smps, ie earth"...and couple back to the power station like that....and these emissions would not get  so heavily attenuated by the common mode filter (Y Caps) in the smps, so would cause a big problem.....thats why earthing a chassis is  definitely about EMC improvement.

[penfold]

If you assume for a second that there isn't an actual earth connection to the chassis...

...Thanks, well, yes...indeed,  having the chassis connection to earth will indeed allow the Y caps to do a better job of EMC improvement….so I agree with you, but say that this is an example of a chassis earth connection improving the EMC situation.

Well, yes, you're not wrong, it would be connected to a potential that we call earth, and the EMC situation would be improved. But equally, it will be connected to live and neutral and I think you'll find that particular connection makes the EMC situation substantially worse.

So yes, earthing an enclosure is all about EMC improvement, as well as safety.

That statement is just logically wrong. It has been agreed that there's no question about the protective earth's role in the safety and the general consensus is that it is possible to build safe un-earthed SMPSs. It has been shown that one video demonstrates a system that failed when no external earth was provided and did not fail when it was earth - it did not prove that it was an EMC problem. We've demonstrated through a thought experiment that common-mode noises generated in a closed box need return paths provided by capacitors or by connection to the protective earth: further work is necessary to prove whether or not the capacitors or earth connection is a more effective return path.

The conclusion is now "Earth can be necessary for safety in some systems and may or may not be necessary for EMC improvements". There's plenty of evidence from existing products and engineers' posts on this forum that proves no earth is necessary so you can either believe that not.

So, yeah, I've got no problem if you just want to validate a series of statements picked from highly specific applications, but to even attempt to make a general rule from them is just absurd.