understanding star grounding

[jneutron]

ALWAYS maintain safety ground integrity.

Configure the input ground to prevent coupling of the ground current with the input.

Remember, current flowing through an outer braid shield DOES NOT CREATE AN INTERNAL MAGNETIC FIELD.

This setup allows you to reference the input ground and prevents magnetic coupling from the input ground loop current to the input ground reference..

Cheers, John

[jneutron]

As an add on note:

I stated ""current flowing through an outer braid shield DOES NOT CREATE AN INTERNAL MAGNETIC FIELD""

I have to clarify that.

A hollow cylindrical conductor does not have any magnetic field inside the inner surface of the cylinder. If you put any current at all into a copper pipe for example, there will be external field which falls off as 1/r...and there will be some field between the inner surface and the outer surface (within the copper), this field increases linearly from the inner surface (zero field) to the outer surface (max field).

Any wire that is within that copper pipe (or shield braid) will not "know" that the pipe is carrying current. This condition is NOT just limited to the geometric center of the pipe...everywhere within that pipe has no field. So for a twisted pair inside a pipe, both will see no induced field. For a twisted pair with a shield, this is not a given however... Zerofield is ONLY for a cylindrical cross section, so for a twisted pair cable, there must be additional filler material with the pair to force the braid to maintain a cylindrical cross section. Some mike cables do indeed do that..

For my previous post showing a mono setup, I force the ground current to travel on that braid shield to the chassis star ground. I then take advantage of the LACK of magnetic field within the braid to get the green wire to connect to the star ground. Any attempt to connect to that star ground OUTSIDE of the braid will see the magnetic field around the braid..the only way to get to that node is to get to the node via a zero magnetic field path..the braid or pipe..

This does not compensate for IR drop of course.

Cheers, John

[AndrewT]

Originally posted by jneutron
............A hollow cylindrical conductor does not have any magnetic field inside the inner surface of the cylinder. ............................................ Zerofield is ONLY for a cylindrical cross section, so for a twisted pair cable, there must be additional filler material with the pair to force the braid to maintain a cylindrical cross section.

What happens to the twisted pair if it is located near the middle of an elliptical shield?

[jneutron]

Originally posted by AndrewT

What happens to the twisted pair if it is located near the middle of an elliptical shield?

While reduced, it will not zero. If the ellipse is a result of the wires themselves, then the wires are also following the twisted ellipse..not good.

That is why filler is needed within a twisted pair cable, as it has to round the cross section.

ah...to answer your question...any non cylindrical shape has the possibility of a zero net field location, but the cylinder is the only shape which guarantees zero field within the entire inner volume.. A twisted pair means that there is a possibility that one of them is in the spot of no field, but not both.

Cheers, John

[h@kan]

I am new into electronics but very anxious to learning. Yersterday I started to put my new GC together and I got a nasty 230V shock when I touched the chassis. It turned out that one of the standoffs to the softstart were in contact with 230V area on the pcb. So now, naturally, I want to make my GC safety grounded to avoid further shocks. I understand the schematics displayed in this thread but I would like to know how to separate safety ground from common ground (audio ground). What parts and what values to use?

[AndrewT]

go to ESP and use his disconnecting network.

The metal chassis must be permanently and directly connected to the Safety Earth.

The external conductive parts must be permanently connected to Safety earth. This can be direct or via a disconnecting network. If a disconnecting network is used it must survive longer than it takes for the fuse to blow and the arc to extinguish when fault current passes to Safety Earth. This fault current can approach kA.

[h@kan]

I guess this is what you mean AndrewT: image link

I have somehow missed it while reading this thread. Thank you for pointing it out.

Regards H@kan

[d1983]

Hey guys.
This is a great thread for a return current intro tutorial.

I have a bit of a question though:

I believe that it's safety standard (at least here in Canada), that the ground path cannot carry current, and that the chassis of your equipment must be earth grounded.
So, what I'm getting at is this:
I think that we should star ground the circuit to the NEUTRAL, or return, path. NOT to the ground path. I believe that if you star ground to both the neutral and ground you will create a ground loop path in your house grounding circuit to your panel where the earth ground is bonded (think of it as star grounding for your house).
I think what ideally should be done is isolate your neutral from your chassis, star ground to your neutral and bond your chassis to ground so that it's never carrying current during normal operation.

I'm certainly no expert in this stuff so I could be wrong.

Any thoughts?

[Redshift187]

Originally posted by d1983
I believe that it's safety standard (at least here in Canada), that the ground path cannot carry current, and that the chassis of your equipment must be earth grounded.

So, what I'm getting at is this:
I think that we should star ground the circuit to the NEUTRAL, or return, path. NOT to the ground path. I believe that if you star ground to both the neutral and ground you will create a ground loop path in your house grounding circuit to your panel where the earth ground is bonded (think of it as star grounding for your house).
I think what ideally should be done is isolate your neutral from your chassis, star ground to your neutral and bond your chassis to ground so that it's never carrying current during normal operation.

An interesting idea, but what if the ground path only carries DC, not AC? And it only carries it between the PSU and the amp, not all the way to the "wall"?.

[d1983]

I think that it's regardless of what type of current it is.
current in = current out, so the low side of your house transformer doesn't care where it's coming from.
I believe you can still have issues with ground potential rise even with DC since the input to the amp is differential, and needs a constant reference.

I think that if you hook up the return of your amp to both the neutral and the ground it will make it all the way back to the wall through both conductors.

I think the issue is this:

If you connect neutral to safety earth you're now connecting your chassis to a system that normally carries current. There is no perfect conductor so depending on the amount of load your house is using and how close that outlet is to the panel, there will be a voltage on that chassis compared to true earth.
It should be connected to safety earth, and that safety earth should have no current flowing on it during normal operation.

Again, I'm just a student of this stuff, but to me this would be the safest way of doing it while still reducing the dreaded hum.