Earthing the isolated secondary of an offline SMPS?

[cupoftea]

If you earth the isolated secondary of an offline smps, can you confirm that an earth connection to the secondary, must come from the earth wire in the mains cable that supplies the smps?
Also, that the smps must feature y caps from live and neutral to earth, as well as a y cap across the smps isolation barrier?
Also, that the "earth loop" that runs through the y caps across the transformer (across the isolation barrier) and also through the earth connection that runs across the isolation barrier of the smps, (to secondary ground) must be as narrow in area as possible?

 

[crutschow]

If you earth the isolated secondary of an offline smps, can you confirm that an earth connection to the secondary, must come from the earth wire in the mains cable that supplies the smps?

No, you can ground the output to any convenient earth ground.

I don't really understand you other questions.

 

[cupoftea]

No, you can ground the output to any convenient earth ground.

Thanks, but if that is done, then there is a very large area earth loop formed.

My other questions also concern the size of earth loops formed.
As you know, earth loops (any loop of a circuit) causes radiated emissions and so must be reduced in area.

 

[cupoftea]

Do you agree that you cant just earth an isolated secondary by connecting any old earth lead to it? The position of the lead, the length of it, etc, has to be thought about, otherwise EMC problems ensue.

 

[crutschow]

As you know, earth loops (any loop of a circuit) causes radiated emissions and so must be reduced in area.

But tying the output to ground at one point does not create a loop (complete circular loop with two separate connections to a point) or EMC problems.
Why do you think it does?

 

[cupoftea]

Thanks, as soon as the secondary is earthed, even if at one point, then there is a large earth loop from the earth at the power station -----> mains input connector------->through the cross-barrier isolation Y caps of the SMPS---------------> to the earth point you kindly spoke of -------> to earth at the power station.

 

[betwixt]

As you know, earth loops (any loop of a circuit) causes radiated emissions and so must be reduced in area.

Not true at all. A loop does not cause radiated emissions, it may however assist with radiating a signal already there in some circumstances.

The loop topology might increase radiation at it's resonant frequency but equally, being a tuned circuit of kinds, will also reduce it at other frequencies. When talking about power cables and links back to the power station (in reality probably to a local distribution board or transformer) there are so many variables it would be virtually impossible to calculate the frequency and it would change as the cable and things around it were moved. In any case, the 'Q' of the loop is likely to be very low so the boost given by resonance is likely to be small.

Brian.

 

[KlausST]

Hi,

I think there is a misunderstanding.
What is the aim of connecting the output to earth?

If you want to avoid the output from floating away - in terms of slowly building up DC voltage, then a big loop is no problem.
Then also you may use a series inductance to avoid a low impedance HF loop (reduce HF loop current)...or use a series R.

But if tou want to reduce HF switching noise coupled to the secondary .. then you can not / should not use a large loop.
In this case the primary input is L and N ... and the noise source is related to this L and N (not related to EARTH). Thus it should be suppressed against L and N. This is the job of the Y capacitor across the transformer.

Klaus

 

[cupoftea]

But if tou want to reduce HF switching noise coupled to the secondary .. then you can not / should not use a large loop.

...Thanks, yes, this is to what i am referring.
When you earth the secondary of an offline SMPS, then you unfortunately unavoidably create an earth loop.
Your task then is to reduce this loop...since part of this loop is the dreaded isolation barrier that the noisy transformer spans across.
The attached diagram is what i mean....
Do you agree with this?

What is the aim of connecting the output to earth?

...Yes its partly to do with stopping the output floating up high...also, many of the incoming wires to the board will be near earth potential so we dont want to damage things by connecting them to a floating board.
Also, when you earth the secondary from the incoming mains earth wire, you can do it in such a way that the overall earth loop is reduced in area, (as attached)......and this means less emc problems

 

[KlausST]

Hi,

I agree that using an auxilary EARTH to reduce HF is counter-productive.

Klaus

 

[crutschow]

You can break the high frequency loop by adding an inductor in series with the ground.

 

[cupoftea]

You can break the high frequency loop by adding an inductor in series with the ground.

I agree that certainly improves things, but as you know, radiated emissions can "jump around" a ferrite bead, and can go through the interwinding capacitance of an inductor.....but your point is of course a good one, and well worth doing.

I agree that using an auxilary EARTH to reduce HF is counter-productive.

Thanks yes, when an offline SMPS's isolated secondary is earthed, then the earth connection (or at least one of them), must come from the actual mains lead in the mains input cable itself......doing it this way allows you to minimise the overall earth loop that goes through the Y capacitors and through some component parasitics.