Offline SMPS Earthing & EMC article begs for clarification

[ocset]

Hi
Please could anyone elaborate on the meanings of various things in this (attached) document on Earthing & EMI Filters?..

The document is called “Safety concerns for practical EMI line filters”…
Code:
   https://www.edn.com/Pdf/ViewPdf?contentItemId=4169156
Some of the statements of this doc appear to be begging for clarification…As follows below....
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ** * * * * * * * * ** *

Pg 2 top

Without a metal enclosure, whether Earthed or not, there is little chance that an
off-line switching power supply can ever comply with the radiated emission limits.

…can this be right?, I have seen many plastic enclosed offline flybacks which pass radiated regulations.

Pg 2

But the metal enclosure (of a power supply) is naturally eyed as an excellent fortuitous heatsink by engineers. So power semiconductors are often going to be mounted on it (with insulation). However by doing this, we also create leakage paths (resistive/capacitive) from the internal subsystems/circuitry to the metal chassis. Though these leakage currents are small enough not to constitute a safety hazard, they can present a major EMI problem. If these tiny leakage currents are not 'drained out', the enclosure will charge up to some unpredictable/indeterminate voltage……..So we need to connect the enclosure to Earth

…..So an unearthed metal enclosure around a power supply can “charge up to some unpredictable/indeterminate voltage”…..The thing is, in many parts of the world, the earth connection to the households etc does not exist, due to the terrain…this is prevalent in many parts of Switzerland, and even in a third of USA homes…..so for them, metal cased SMPS’s are going to be highly dangerous?

Pg 2

Generally, if the equipment is designed not to have any Earth connection (i.e. a two wire AC cord), there will usually be no metal enclosure present either. Keeping the issue of radiation limits aside for now, the good news here is that no significant common mode (CM) noise can be created, simply because CM noise needs an Earth connection by its very definition. Therefore a CM filter is superfluous and can be omitted.

…Surely this is not correct?, offline SMPS’s which have no earth connection, usually always have a common mode choke at their input. This is because "planet" earth is all around everything, and noise inside the SMPS will couple to it, even through the air.

Pg2

One of the simplest ways of suppressing any noise is to provide decoupling between the nodes involved. For CM noise this means connecting high frequency ceramic capacitors between the L and E wires and also between the N and E wires, possibly at several points along the filter. But each of these CM line filter capacitors also unintentionally pass some of the AC line current into the chassis (besides the CM noise). To reduce the chances of a fatal electrical shock, safety laws restrict the total amount of current that can be injected into the Earth/enclosure. This in turn means that we have to limit the total CM filter capacitance.

…Surely if the chassis is solidly connected to earth ground, then its impossible to get an electric shock by touching it?, no matter how much Y capacitance connects to it? Surely the reason to restrict Y capacitor size is so that the current from the Y capacitor doesn’t trip the RCD’s .?

Pg 3
The following concerns the use of balanced (symmetrical) filters in offline SMPS EMI filters…

In general we try to maintain balanced impedances because any imbalance basically causes some of the CM noise to get converted to DM noise. When this happens the resulting EMI spectrum may be rather confusing to analyze and fix,

…..Now I agree with this, but debate whether the problem of CM noise getting converted to DM noise is really a problem?, because one simply needs to filter it out with a DM filter.
Another point is that a balanced filter potentially needs more inductors (one in line and one in neutral), and so is inconvenient and more costly.

Pg 4

As per safety standards, we cannot for example, just wind the two windings of an off-line CM choke carelessly overlapping on each other. We need to maintain a certain specified separation. Nor can we just use any bare toroid to wind them on. We need an approved coating, or a suitable designed bobbin.

…..I would have thought as long as the CM choke was wound with wire coated in acceptable level of insulation, then the torroid would not have to have an “approved coating”? Also, how would an approvals agency test whether a torroid coated in an “approved coating” had been used?......there is no safety test involving connection of any voltage higher than mains peak to the live and neutral connections…and torroids wound with ordinary enamelled copper wire would pass this….even with non coated torroids.

Pg 4

A bare ferrite can be a very good electrical conductor, especially if it is the more commonly used manganese-zinc ferrite, as opposed to nickel-zinc formulations. This can be confirmed by simply pressing the tips of an ohmmeter at two points on the surface of a typical ferrite core lying around in the lab.

…..Surely this is not right?, its very difficult to get any kind of resistance reading by probing ferrite surface, even with Mn-Zn ferrite.



Pg5

Line to Line capacitors are called 'X-capacitors'. X-caps when used before the input bridge (in offline applications) must be safety approved, but after the bridge it's basically a 'don't care' situation.

I agree that this is how the standards appear to tell it…..but why should an interposing bridge rectifier give creedance to the use of non X2 rated capacitors?
Are we saying that if a big mains transient happens in an offline SMPS, then a non X2 rated capacitor is going to be protected from it just because its fitted downstream of the bridge rectifier?...

Also, are the safety agencies going to conscientiously take the schematic and bill of materials and check that all DM mode capacitors upstream of the bridge rectifier are X2 rated?


Pg 6

We should be conscious however that large input (X2 rated) capacitances can cause undesirably high inrush surge currents at power-up. This may also cause eventual failure of the X-cap, especially if it is the very first component after the AC input inlet. Film caps may self-heal from such an event each time, but after every event, the capacitance gets degraded just a little. Finally, after many such events, there is a cumulative effect, and we would be left with a capacitor that is barely one. Therefore, despite EMI concerns, we should rather place X-caps after any input surge protection element (e.g. NTC thermistor or wirewound resistor), and perhaps even after a front-end choke.

…Have to agree with this one, but it fails to state the real problem concerning cumulative damage to metal-film X2 capacitors due to repeated exposure to high inrush currents….and that is that the X2 capacitor can become highly conductive, and heat up and eventually explode in service…I don’t know why the article failed to mention this?


Pg7

Safety regulations for Nordic regions (and Switzerland) may require each Y-cap shown in Figure 1 to be actually two Y2 capacitors in series (or a single Y1 capacitor). Historically, this has been necessitated by the fact that Earthing is poor in those geographical regions. In fact, it used to be pointed out that even the central meeting room of the Norwegian safety agency NEMKO (literally Norwegian Electric Material Control) did not have any Earth point in the wall outlets. Therefore, practically speaking, a lack of Earth is not considered a fault condition in many parts of the world, but is just a normal condition (this actually also includes about 1/3rd of homes in the US!). Therefore, very often, whether the equipment is supposed to be Earthed or not, it is expected to have reinforced insulation anyway. The Earthing if present, is just for helping out with EMI.

….“The Earthing if present, is just for helping out with EMI”……
…surely this is wide of the mark?.....surely it doesn’t matter whether or not the building is actually connected to "planet" earth ground…but surely an earth connection is useful (even if it doesn’t ultimately connect to “planet earth”) because it can be used to trip an RCD and save someone’s life?

Pg 9

Engineers are often mystified to find that making the DM choke out of (low permeability) powdered iron or lossy ferrite helps too, when all else has failed, despite all the talk about DM noise being essentially a low-frequency emission. The reason seems to be as follows. The CM noise in a power supply is actually a nonsymmetric mode at its point of creation, though ultimately, by cross coupling it does tend to spread into both the lines equally. We can show that nonsymmetric noise can be considered to consist of part CM and part DM components. Therefore in practice, we do get a fair amount of high frequency DM noise too from nonsymmetric CM noise. That is where high bandwidth/low permeability/lossy materials help in DM noise suppression.

….this suggests that its always adviseable to add some small value inductors (using nickel-zinc or powdered iron core) to the line and neutral inputs of offline SMPS…to handle the 10-30MHz noise….because they will give a joint CM and DM filtration…….and this is prefereable to using a small CM choke…because a small CM choke won’t offer enough filtration to the high frequency DM components which always accompany high frequency CM components. Would you agree?

Pg 10

CM noise suppression is usually said to require a very 'good' connection to Earth.

..Surely this is not always the case?…many offline SMPS’s do not have an earth connection, and still get round it by using common mode chokes.

[ejeffrey]

One of the simplest ways of suppressing any noise is to provide decoupling between the nodes involved. For CM noise this means connecting high frequency ceramic capacitors between the L and E wires and also between the N and E wires, possibly at several points along the filter. But each of these CM line filter capacitors also unintentionally pass some of the AC line current into the chassis (besides the CM noise). To reduce the chances of a fatal electrical shock, safety laws restrict the total amount of current that can be injected into the Earth/enclosure. This in turn means that we have to limit the total CM filter capacitance.

…Surely if the chassis is solidly connected to earth ground, then its impossible to get an electric shock by touching it?, no matter how much Y capacitance connects to it? Surely the reason to restrict Y capacitor size is so that the current from the Y capacitor doesn’t trip the RCD’s .?

That is to restrict the current if the earth connection is not present or broken.  This can happen due to a broken connection, intentional use of a "cheater plug", or attachment to an ungrounded outlet where permitted.  This is all part of the "no single point of failure".  If the ground conductor is carrying significant current from the EMI filter, then a single fault (broken ground conductor) results in a dangerous configuration -- high voltage present on the case with significant current capacity behind it.

This is a much bigger problem in (parts of) Europe than the US both due to the higher voltage -- the half-rail voltage is 115 V instead of 60V, and because of the shucko style plug allowing grounded plugs to attach to ungrounded outlets.  They usually have RCDs, but those RCDs are usually sized too big to prevent dangerous currents from flowing.  I think 30 mA RCDs are common, but 5 mA is easily enough to be dangerous. 

Safety regulations for Nordic regions (and Switzerland) may require each Y-cap shown in Figure 1 to be actually two Y2 capacitors in series (or a single Y1 capacitor). Historically, this has been necessitated by the fact that Earthing is poor in those geographical regions. In fact, it used to be pointed out that even the central meeting room of the Norwegian safety agency NEMKO (literally Norwegian Electric Material Control) did not have any Earth point in the wall outlets. Therefore, practically speaking, a lack of Earth is not considered a fault condition in many parts of the world, but is just a normal condition (this actually also includes about 1/3rd of homes in the US!). Therefore, very often, whether the equipment is supposed to be Earthed or not, it is expected to have reinforced insulation anyway. The Earthing if present, is just for helping out with EMI.

….“The Earthing if present, is just for helping out with EMI”……
…surely this is wide of the mark?.....surely it doesn’t matter whether or not the building is actually connected to "planet" earth ground…but surely an earth connection is useful (even if it doesn’t ultimately connect to “planet earth”) because it can be used to trip an RCD and save someone’s life?

I would read that differently.  The safety ground can't be assumed to be present for safety purposes because it is often not.  Therefore devices must be designed to be safe (including against "single" failures) with or without an earth connection.  Agree that connection to planet earth is irrelevant -- it is the missing ground wire to the outlets that is the problem.

[ocset]

The thing is, i am sure you would agree, that if a metal chassis'd offline SMPS has a broken earth connection, then it is very dangerous, because that metal chassis may get induced up to some arbitrarily high voltage?...then somebody could touch it and get a shock.

[rstofer]

In general, all exposed non-current-carrying metal surfaces are to be grounded to earth, the days of the 'hot chassis' radio and TV are long gone.  The exception being 'double insulated' items.  The device itself may have a tag stating that it can only be used on grounded circuits.  If a buyer doesn't have a grounded system they shouldn't use the device.  Inevitably, there will be several pages of safety information in several common languages.

[bjbb]

"In general, all exposed non-current-carrying metal surfaces are to be grounded to earth, the days of the 'hot chassis' radio and TV are long gone.  The exception being 'double insulated' items.  The device itself may have a tag stating that it can only be used on grounded circuits.  If a buyer doesn't have a grounded system they shouldn't use the device.  Inevitably, there will be several pages of safety information in several common languages."

No, there is not such a requirement per IEC60950-1, nor IEC62368-1, where it is dead metal with proper spacing. Where the metal is connected to a primary-side circuit, it cannot be exposed to the end-use environment, so the issue is mute. IEC60355-x (appliances) does have additional requirements where a single-fault condition of Basic Insulation could cause an exposed dead metal surface to a hazardous voltage.

That said there are some specific mechanical and electrical requirements for exposed conductive surface in the EU (LVD) and Japan, and elsewhere.

[T3sl4co1l]

I'm admittedly a bit impressed.  EDN articles are notoriously bad, but when even treez is finding them sorely lacking?  That's something!

Tim

[ocset]

On the third page of the document linked at the top of the top post (last but one paragraph on the third page) it says that offline EMC filters should always be balanced because otherwise CM noise will get converted to DM noise, and this “converted noise” will be very hard to fix.

However, page 2 of the following document…..

https://www.ipen.br/biblioteca/cd/ieee/1999/Proceed/00309.pdf

….states that  generally CM noise occurs mostly above 2MHz, and DM noise mostly below 2MHz. It also states that any  DM noise above 2MHz will generally get conveniently filtered out by PCB  layout stray inductances and capacitances.

….So this tends to kind of blow the first document’s statement  out of the water……because it kind of poo-poos the idea that DM noise above 2MHz will be difficult to filter out…..i mean, the first doc states that CM noise converted to DM noise will be hard to filter out….but the second doc pretty much refutes this, suggesting  that above 2MHz, DM noise will be generally easy to filter out. (suggesting that layout strays will mostly filter it out)

What do you make of this?

This idea of CM noise getting converted to DM noise also sounds like something that surely is not that common?

[dmills]

  EDN article being up to their usual high standard basically, nothing to see here!

Most of those statements make sense UNDER PARTICULAR CONDITIONS, they just seem to have left out the bits that say when particular things apply or don't!

Like all publications, take with a pinch of salt and do not use the word from on high as a substitute for knowing the theory and how to apply it.

Most power supply EMC stuff comes down to fairly basic electromagnetism (Usually you don't need to go as far as Maxwell), I have never understood why this stuff is treated as a big mystery.

Also, a bit of leakage to a metal case in a class II device thru a class Y cap (Metal cased class II do exist, see most modern HIFI amps), while it might read half mains on a high Z meter is NOT realistically any kind of danger, the source impedance is too high.

Regards, Dan.

[T3sl4co1l]

CM-DM conversion is normal. It just means unbalanced (asymmetrical) impedances.

Tim

[ocset]

CM-DM conversion is normal. It just means unbalanced (asymmetrical) impedances.

Thanks, and since CM noise is high frequency, the DM noise that it converts to is equally high, and possibly above the SRF of the DM filter inductors, thereby meaning a few low value (high SRF)  filter inductors are likely to be needed at  the inputs in "go" and "return".... Would you agree?

[T3sl4co1l]

Maybe.

Note that it's not the parallel self resonant frequency (SRF), in fact that gives peak attenuation.  It's somewhere above, at the first series resonant frequency.

That usually happens at a high enough frequency that a ferrite bead is all that's needed (if anything at all) to at least dampen the resonance.

The filter impedance is also usually low enough (i.e., the X cap, the impedance of which will be dominated by its ESL at this frequency) compared to the series resonant impedance that attenuation is still fair, even at that resonance.

For the same reason (ESL), smaller X caps may be necessary, in combination with the smaller inductors (whether actual inductors, or FBs or CMC leakage).  There usually isn't so much emission at such frequencies to require this treatment, and also this range is better treated with layout, transformer windup, and control of switching edges.

Tim