Earthing in SMPS for EMI noise reduction?

Hello,
The following 3kw electric car charger is totally encased in a metal enclosure which is connected to earth ground, -I assume multiple parts of the internal PCB are electrically connected at various points to this earthed enclosure. The reason for earthing is safety , but also , it is good for EMI noise reduction. This is because the noise radiated out of the switching power supply components, has an affinity to couple to earth, and as such, it readily couples to the earthed enclosure (instead of coupling to earth outside the enclosure and causing interference). Then, when the noise has coupled into the earthed enclosure, it runs along it, eventually finding a Y capacitor, through which it flows, before going back down the neutral wire, and in the process, getting attenuated by the differential mode filter.
It doesn’t run down the earth connection wire because the common mode choke makes that a high impedance path to the noise.

So anyway, we see how essential an earth connection is for EMI noise reduction. So how do European plugs manage to avoid this noise problem?, they have no earth connection. The power supply would not be earthed and the noise will run wild.?

EV charger ;3kw
https://www.google.co.uk/url?sa=t&r...-mu3UYfLLlcKKhARQ&sig2=97vBkIqtM4HfatAm9Dxkcw

European wall sockets have a side contact earth. The power supply you linked to has a screened mains input lead and what looks like a three pin connector.
Frank

Do European rectangular 2-pin plugs have Earth tied to Neutral, or is that not possible? The round 2-pin plugs have those two slivers of contact material perpendicular to both pins that are the Earth connection.

Only ask as it perplexes me re safety, and some buildings have Earth tied to Neutral at the entrance from what I read.

I thought only the "schuko" 2pin plugs had earth in them, most euro 2 pin plugs have no earth whatsoever?

treez said:

I thought only the "schuko" 2pin plugs had earth in them, most euro 2 pin plugs have no earth whatsoever?

Click to expand...

Thanks, cue cheap suspense film music "Tan, tan , taaaaaannnnn!" - that is worrying and pretty unsafe for users if correct. Will have to look one day for how it's done using two-pin.

A "Europlug" is flat, has 2 pins and no earth connection. They are only rated up to 2.5A and are only allowed for double-insulated equipment.

The 3 kW charger must use a Schuko plug with earth connection.

thanks, so anyway, what is the concensus regarding the statement about earth ground and noise management in the top post.?..ie using an earthed enclosure around an smps in order to reduce the EMI problem.
You agree that Earth connection to smps is not just about safety?...its about EMI reduction.

The trick (if you can call it a trick) is the use of electrostatic screens between any "hot" in RFI terms part, and the grounded outer metal case.

So you might have something like the drains of your mosfets frantically switching up and down, then an electrostatic screen which is connected back to mains neutral, then the grounded exposed heat sink connected to the outer metal case, and safety ground. Creepage and clearances here for the screen need to follow statutory requirements.

The idea is to keep as much of the noise as possible differential between mains active and mains neutral, where it can be more effectively filtered, plus adding some fairly minor common mode rejection as well.

As little noise as possible should actually be able to reach the outer metal case and safety earth by interposing an inner electrostatic screen at any likely coupling points.

So you might have for example, mosfet drain, broad mica insulator, copper foil connected to mains neutral, another broad mica insulator, grounded heatsink.

The mosfet drains are probably the worst offenders for capacitively coupling high dv/dt into the outer case, but there may be other coupling paths such as the magnetics, where electrostatic screens, sometimes multiple layered screens can be quite useful.

The other issue is minimising the size of high di/dt loops, by careful placement of wires and tracks within the enclosure. Air gaps in magnetics may also sometimes require attention.

Its mainly thinking the whole thing right through in detail right from the very start.
Once its built, the multiple problems can otherwise become intractable.

Thanks, the point is, from an EMI point of view, why is the metal enclosure connected to earth ground?, why not just have it "floating"? And why are so many points on the PCB inside the enclosure connected to the earthed enclosure, and would they still need connection to the enclosure if it were floating?
A 'floating' metal enclosure around an SMPS would mean less chance of passing EMC?

treez said:

A 'floating' metal enclosure around an SMPS would mean less chance of passing EMC?

Click to expand...

It might if the enclosure was really "hot" with high dv/dt with respect to the surrounding ground.
Receiving RF burns from a high power radio transmitter not properly grounded can and does happen ....

Ground is relative, you can ground something within an airborne aircraft.
Its a big flying Faraday cage.
Similar with a vehicle on rubber tires.
Neither may radiate EMI as long as the isolation from ground is total.
If it has a trailing mains cable, it changes everything.

But rules is rules, and any exposed metal containing mains voltages within, must be connected to ground for very obvious safety reasons.

Provided the SMPS injects high frequent currents between case and other parts of the circuit, connected either to the mains or isolated output circuit, the primary EMI reduction means is having input and output filters, not returning the current to the earth ground. Any EMI current flowing through the earth line will generate a complementary current in either input- or output lines and appear at least in conducted EMI measurements.

Effective common mode LC filters require a ground return to handle the capacitor leakage currents, it's hardly possible with a class 2 safety insulated SMPS. Even if you manage to have an internal metal screen with reinforced insulation towards the outside, EMI currents injected to the output side can't be well filtered.

Thanks,
The point that I am trying to make is that as you know, common mode noise can still appear in SMPS’s which have no earth wire taken to them whatsoever. This is because the noise current can come into the SMPS via the live line, then radiate out of the SMPS into the “planet earth” somewhere, and thus not go back down the neutral line, thus it is forming common mode noise….As such the noise “radiates” out to the “planet earth” external to the SMPS.
What I am now saying is, ..when you have an earthed metal enclosure surrounding an SMPS and you connect this to the earth wire, then that metal enclosure forms a “platform” into which any radiating noise can go into , which is preferable to it going into the “planet earth” external to the smps. Either way, radiating noise tends to couple to earth….so by having an earthed metal enclosure you are “capturing” noise which would otherwise escape to the “planet earth”, ….and indeed , once the noise couples into the earthed metal enclosure, it can then find its way through a Y capacitor, and end up going down the neutral line, -thus stopping the problem of it becoming common mode noise…..and indeed, when it goes down the neutral line, it ends up getting filtered by the differential mode filter.
In this way, I am saying that an earthed metal enclosure can help to reduce common mode noise. Do you agree?

What I am now saying is, ..when you have an earthed metal enclosure surrounding an SMPS and you connect this to the earth wire, then that metal enclosure forms a “platform” into which any radiating noise can go into

Click to expand...

Yes that is how I see it too.

Grounding the surrounding enclosure, as opposed to leaving it totally floating will effectively block any radiated electric E field components.
For blocking the magnetic H field component, grounding is not a requirement.

In either case, the surrounding screen needs to be pretty continuous without large slots or apertures.

Grounding the surrounding enclosure, as opposed to leaving it totally floating will effectively block any radiated electric E field components.

Click to expand...

thankyou for confirmation...do you have any reference for this?...I mean, its what I have always thought, but also I think its kind of strange that noise has an affinity to couple to the planet earth.

Some years ago I put together a radio receiver using just a coil, a crystal earpiece and a signal diode, and a variable capacitor….it didn’t work at first…just gave a hissing noise, but then I connected its ground node to the earth from a mains plug, ..and it suddenly started working, and I got radio five reception very clear…….kind of as if the noise which was stopping it from working kind of “ran away” to earth when I connected the earth.

treez said:

.do you have any reference for this?.

Click to expand...

The best text I have found on this and related subjects is:
"Electromagnetic compatibility engineering" by Henry W. Ott. (800+ pages)
Its available in hard cover at fairly significant cost, or for sure in the library of your nearest university.

Its now just very recently become available as a free download from the internet:
https://s1.downloadmienphi.net/file/downloadfile5/192/1388769.pdf

Thanks Warpspeed for referencing that book in your post #15........

On page 523, Ott basically effectively agrees with what I am saying, but he uses a y cap to provide the lower impedance circuit for the noise….whereas I am using the earthed enclosure and the Y cap together to do the same job.
There can be no doubt that the noise would more readily travel along an earthed enclosure that’s close to the power supply, than ‘jump’ out of the power supply and follow the path shown by Ott on page 523 (in his figure 13-26)

..so I am not sure why Ott is so anti-earth-connection in these other pages….

….on page 114, Ott staes that “it is a myth that connecting an equipment to an earth reduces noise and interference”.
Also, on page 119, Ott states that “As a matter of fact, an earth ground is more likely to be the cause of a noise problem. More electronics system noise problems are resolved by removing (or isolating) a circuit from earth ground than by connecting it to earth ground”.


That’s all very well, but that wasn’t the case with my radio receiver (described here in post #14)…..in that case, the connection to earth ground totally removed the noise problem and allowed the circuit to work perfectly, whereas before the earth connection was made, it didn’t work at all.

So I wonder what Ott would say about that?
Did I just happen to stumble on the one single exception to his rule, or is his rule somewhat wide of the mark?

Hi,

You know I know little more than an LED + switch + battery CAN work if correctly connected, certainly little about this thread, so I should shut up, but leaving how books are written and the inherent accidental punditry aside which can escape anyone's lips or pen when trying to express a concept, I'm wondering if Mr Ott was referring to inadvertent ground loops as Earth potential may be slightly different to equipment ground and introduce "noise" that way.

Your example is quite interesting about the radio, and it may reflect how all Earth connections are not the same when looked at and measured closely maybe?

I suppose it has to do with signal to noise ratio.

Grounding your receiver vastly increased the wanted signal with respect to the unwanted noise.

Anyhow, Ott presents a lot of very good solid time proven methods in a very practical readable way.
If it does not work quite as expected, there is always a very good reason why, probably some fundamental point that has been completely overlooked.

Its all a fascinating subject, but being in retirement myself, I am now rather out of it.
Small DIY home brew projects are very different to design for manufacture, including all the official hoops and hurdles of achieving formal EMC compliance.