Common mode noise reduction by connecting ferrite SMPS core to ground?

[treez]

Hello,
How much better improvement would you expect to get in common mode emissions in a 30W offline flyback by connecting the ferrite core to primary circuit ground?
Here it is specified on the transformer winding spec diagram…
https://imgur.com/rdqzATJ

 

[dick_freebird]

None, since ferrite is hardly conductive.

 

[treez]

Thanks, do you know why it is done?
The diagram in the top post is of a transformer which exists inside a totally sealed plastic enclosure, so it wouldnt have been done for safety reasons.
(The pin2 in the diagram, gets connected to circuit ground....and this gets connected to the ferrite core)

The picture in the top post is of a offline transformer which is today used in 100's of thousands of consumer products...

 

[FvM]

Nothing shown about the transformer. Why do you assume it's a ferrite core?

 

[treez]

Thanks, i used to work at the company and i know this is the ferrite core of an offline 70w flyback smps.

 

[FvM]

I know this is the ferrite core of an offline 70w flyback smps

How is it connected?

 

[treez]

thanks, i will find out when i get back to where it is kept on monday...as i remember there is a kind of metal plate stamped to the ferrite, and this is wired to pin2 of the transformner bobbin...but i will check monday

 

[c_mitra]

i used to work at the company and i know this is the ferrite core of an offline 70w flyback smps. ...

Ferrites are iron and other metal oxides and you cannot solder or weld to this.

They are mostly spinel but that is immaterial. You cannot ground a non-conductor. They are very hard (you cannot drill but you can use a grinder) but also very brittle.

Because they are non-conductor, they do not have eddy losses. They have some hysteresis loss but that too is small compared to iron (or other magnetic materials).

Sometimes mistakes do get propagated (because no one questions them) - like genetic errors. You can attach a wire using a silver paste (to the core) but that serves no practical purpose.

Sometimes you may want to screen the whole transformer by enclosing it in a copper mesh. But that is a different story.

Or, am I missing some other point???

 

[c_mitra]

this is wired to pin2 of the transformner bobbin...but i will check monday

Sometimes a screen is placed between the primary and secondary windings; a copper foil or a mesh is commonly used. They must not close a loop and if the mesh (or foil) is grounded, it prevents electrostatic induction (capacitance in other words) between the primary and secondary. The foil or mesh, if closed (forms a loop), will prevent the magnetic induction to work (it will form a virtual shorted secondary winding). The screen is quite effective if grounded. But it does not connect to the magnetic core and you need to check carefully the connections.

 

[FvM]

treez has been discussing a lot about about interwinding shields, I presume he knows the difference. The question is definitely not about interwinding shield.

 

[treez]

How is it connected?

Primary side ground is actually connected to the ferrite core by way of a 3cm bare copper wire which is taped to the underside of the ferrite core. the end of the bare copper wire is soldered to pin 2 of the bobboin, which itself is soldered into primary ground.
Do you know why the ferrite is connected to primary side ground?

 

[dick_freebird]

Wonder whether this is a "legacy" from days when a
stamped steel core was used, and nobody thought to
delete the now-useless bit? I have seen laminated-
core xfmrs with a pigtail.

 

[c_mitra]

Do you know why the ferrite is connected to primary side ground?

It is not correct to describe this as a connection; the ferrite body is an insulator and one cannot connect electrically to an insulator. It serves no electrical purpose.

Some mistakes can be hereditary...

 

[treez]

The thing is (on a related subject) people often talk about winding torroids so that there is minimal capacitive coupling of the wire to the core...........and so what this "pigtail" does, i believe, is capacitively couple the core to primary ground...for the purpose of common mode noise reduction. Do you agree?
It cant be an accident, this transformer was wound by an enormous, and sucessful Chinese SMPS company.
Also, the small 1W transformer also has the core similaraly connnected.

 

[c_mitra]

capacitive coupling of the wire to the core...

Yes, there will always be some capacitive coupling between the winding and the core (even if it is air-cored).

Assume for the sake of this discussion that the core is metal. Then the insulation between the core and the winding makes the dielectric. The core and the winding can be considered the two foils of the capacitor.

Do you change the capacitance if you ground one of the two ends (of a capacitor)? Say if you ground the core of a iron cored transformer (it will still be tricky because the laminations are insulated), what will be the effect?

The thinner the dielectric, the greater is the electric field and the capacitance. We cannot increase this thickness arbitrarily for practical reasons.

Assume now that the core is an insulator. The core now is the dielectric. And the bare (it does not matter whether it is bare or not) copper wire is now the other electrode of the capacitor.

Because the solid core now acts as the dielectric, the capacitance will now be less. But how much effect you will see if you ground the other end?

In particular, as the copper wire does not go around the core, I cannot guess the overall effect.

But anyway, the core is not grounded simply because it is non-conducting.

 

[mtwieg]

If the core is a high resistivity (MN-Zn or powdered iron) material, then I doubt the connection does anything. The impedance of the core should be much higher than the impedance of the parasitic capacitance to the core.

 

[Kiriakos-GR]

Something is missing from this long conversation, and this is the appropriate test and measurement instrument which will measure -> Common mode noise.

 

[treez]

i am wondering if the core grounding is an extra manufacturing step to make the transformer more expensive to manufacture, and thus more pre-disposed to being manufactured in the Far East as oppposed to in the west?

- - - Updated - - -

Thanks incidentally, the 70W flyback transformer spec diagram in the top post is from a Chinese PSU company....who did the transformer for a UK based company……the spec as you see it there has a serious fault in it, (this fault didn’t occur in the actual unit, so the unit actually worked fine) which of course, we would put down to “human error”.
(Note...the actual power supply, as designed by the Chinese, worked absolutely fine.)

Its interesting though….how often does this happen?....because without a transformer spec, it means a company/customer cant take the transformer spec elsewhere to get it made……this could be construed as being “convenient” for the supplier.

As the UK gets more into outsourcing its PSU engineering overseas…these "spec errors" could become more prevalent…….its very convenient for a supplier if the customer doesn’t have a proper spec, and doesn’t know how to do one themselves…because the customer is then bound in to the original supplier, who can then lever the price up, safe in the knowledge that the customer cant go elsewhere…or if they do, it will go wrong for them, (because of the “mistake” in the transformer spec) so they will come running back…..many will say the customer just needs an in-house PSU designer to check the specs over….but with PSU design work being increasingly sent overseas…..UK kids aren’t coming into PSU design as a career in such numbers as before.
Slowly but surely, the UK for one, is backing itself into a very deep, sticky corner.

 

[c_mitra]

If the core is a high resistivity (MN-Zn or powdered iron) material, then I doubt the connection does anything...

Even if it is a traditional iron core, they are laminated and individual laminations are insulated. It will be difficult to connect to all the individual laminations. But some of them have a metal frame and for safety reasons the metal frame is often grounded. I have seen several linear regulators in which the bulky power transformer is not on the main PCB and the transformer is connected to earth ground. But many modern SMPS have finally got rid of the earth ground altogether (two wire connections).