Switch mode power supplies

[lyledunn]

In very simple terms, why do SMPS cause currents in the earth conductor?

 

[john1]

Hi lyledunn,

They shouldn't really.
How is it arranged?
If its fed from a battery, sometimes the low voltage
return is the same connection as earth.
If thats the case then the earth return may include
part of the battery current, have a look at the layout
if thats the prob, then you could re-arrange the
connections so that the current path doesn't go along
the earthing. Sometimes that is not possible, such as
where the battery supply has only one feed wire, and
returns via local metalwork (like on a car).
Even on cars high current items like starter motors
sometimes have extra cabling to avoid excess current
through local metalwork.

The output from a battery fed power supply shouldn't
involve or cause earth currents in its output circuits,
and you should be able to use normal type safety trips.
(i think). Using the output will mean that around ten
times the output current of the unit will be drawn from
the battery so it could be the return path that is
showing up.

John

 

[electricuwe]

Look at my answer to mrdee's question on the same topic.

 

[lyledunn]

Thank you electricuwe. I understand that the current should be only a few mA. However, take a large installation of say IT equipment in an office building all using smps, that small current will now become a large current when it sums in the distribution circuits. This then becomes a potential danger. My original question was really to help me clarify just why smps and any associated filtering equipment really need a third conductor, ie the circuit protective conductor in order to function properly. It has obviously been accepted practice for a long time so I assume that there is no other way?

 

[lyledunn]

Thank you John,
Just to clarify, I refer to units in IT equipment served by 230volts in the uk. I am an inspection engineer and I have been able to measure considerable currents in protective earthing conductors at the origin of installations.
Please see response also to Electricuwe

 

[Dragonclaw]

In SMPS input filtering, there are Safety Class Y capacitors installed from both the live and neurtal terminals. There is a leakage current flow from the live-to-earth (also from the neutral-to-earth). The magnitude of which can be calculated very accurately by integrating the AC sinewave input voltage etc. but can be estimated as Ileakage = Vac(max) x 2 x PI x frequency x Cmax
Where Cmax is the total capacitance from live-to-earth (or neutral-to-earth).
Example: SMPS with 4700pf (+/- 20% tolerance) safety class Y capacitor fitted from live-to-earth (and 4700pf from neutral to earth) Operating from UK 240VAC 50Hz (note: 264VAC max)
Ileakage = 264 x 50 x 2 x PI x 5.64E-09 = 0.468mA

If you had say 20 pcs of equipment like this then the total leakage current on the single 240VAC utility supply would be additive, so = 4.68mA
(for 100 psc equipment then = 46.8mA etc.)

 

[Dragonclaw]

Oops - spotted the typo 20pcs should read 10 pcs

In SMPS input filtering, there are Safety Class Y capacitors installed from both the live and neurtal terminals. There is a leakage current flow from the live-to-earth (also from the neutral-to-earth). The magnitude of which can be calculated very accurately by integrating the AC sinewave input voltage etc. but can be estimated as Ileakage = Vac(max) x 2 x PI x frequency x Cmax
Where Cmax is the total capacitance from live-to-earth (or neutral-to-earth).
Example: SMPS with 4700pf (+/- 20% tolerance) safety class Y capacitor fitted from live-to-earth (and 4700pf from neutral to earth) Operating from UK 240VAC 50Hz (note: 264VAC max)
Ileakage = 264 x 50 x 2 x PI x 5.64E-09 = 0.468mA

If you had say 10 pcs of equipment like this then the total leakage current on the single 240VAC utility supply would be additive, so = 4.68mA
(for 100 psc equipment then = 46.8mA etc.)


By the way - not all SMPS utilize the earth conductor as part of their overall filtering and EMC solutions. Example: Some notebook computer AC/DC adapters.

 

[tjdiamond]

Dragonclaw, thank you for the clear explaination on earth currents due the caps from live and neutral to earth. Is there a reason that the 4700pF value is used in most supplies?

Also is there an equivalent explanation earth current due to the capacitor used to couple the DC output an isolated supply to earth? I assume it is a factor of the switching frequency and the shape and magnitude of the output ripple.

 

[Dragonclaw]

There is no precice reason why 4,700pf can be seen used on a lot os SMPS. This value is certainly used in volume, on a worldwide basis, and as such good low cost part price can be realized. 4,700pf is a good compromise between keeping the leakage current down whilst having a reasonable capacitance available to shunt series mode noise.

Decopuling capacitors on the DC side, equivalent earth current flow? Don't believe there is due to the nature of the DC. However, maybe there is someone out there can answer this question.

 

[Heydave]

More modern SMPS run at ever higher frequencies and therefor require lower value filter capacitors. The reason ground is used for filtering involves the stray capacitance between the primary and secondary in the high frequency transformer and other parts of the circuit. This coupling produces common mode EMI back through the supply mains. Series filter inductors alone are not sufficient for high power supplies but quite are good enough for lap top supplies. Often, this kind of inductor is attached to the power cable making lump.