Power transformer grounding

[catalin gramada]

Hello. I wonder about induced current into PT end bells and I did some test on new transformer received. First the transformer end bells and bolts was isolated with paper gaskets till 0 ohm received between end bell caps on my 10 MO Fluke DMM. still 250pF in between. The transformer was posed on chassis with firm contact between end bell legs and chassis. secondary mid tap was tied with 100k0hm resistor to chassis. No load conected. I received 1.78V over resistor with no electrostatic screen conected and 200 mV with elscreen connected.
Starting with this premises I wonder if is not more normal to isolate the PT from chassis and to return this currents with a wire to first decoupling and not through direct contact from chassis.
I.m pretty sure is not a particular situation and perhaps direct coupling between end bells and core goes worthless. Bolted PT to chassis is a common practice walking this currents through chassis to star point. Why to keep transformers bolted to chassis and not isolate it and return this currents throuh a wire to correct point,please? Or with other words Why to reference the PT to the ground through chassis and not with a seperate wire as time we have to carry some return curent induced into the shield, please ?
Thanks


edit: supposed we have a functional earth point seems more normal to drain this currents to this point....but also with a wire, not through chassis. but in case of a "floating" mains...that.s rubbish in my understanding

[nickb]

I think the real point is the current is minute. Therefore any voltage developed over the chassis resistance will be negligible. Furthermore, that voltage current shouldn't couple into the signal path unless you are using the chassis to carry signal current, which of course you should not.

[catalin gramada]

Hi, thanks for answer. I.m not pretty sure... for instance my previous experiences with some Hammond transformers which have electrostatic screen tied directly to the end bells. Disconnect and Running electrostatic shield with a separate wire to star point did a bunch of difference in term of noise. I disconected all time this type of conection and run with a separate wire to correct point...cause it matter. But I did not tried to isolate the PT from chassis and reference through separate wire to do a difference,yet, so I asked for. Thanks

[nickb]

For a power transformer the screen should be tied to the safety earth. End of story.

[catalin gramada]

through chassis, please ?

[R.G.]

Hello. I wonder about induced current into PT end bells and I did some test on new transformer received. First the transformer end bells and bolts was isolated with paper gaskets till 0 ohm received between end bell caps on my 10 MO Fluke DMM. still 250pF in between. The transformer was posed on chassis with firm contact between end bell legs and chassis. secondary mid tap was tied with 100k0hm resistor to chassis. No load conected. I received 1.78V over resistor with no electrostatic screen conected and 200 mV with elscreen connected.

Magnetic induced voltages are funny. The induced voltage depends on the **area** of the conductor loop you're measuring across, including the measurement device. Bigger loops "capture" more magnetic flux lines to make voltage. The resistance of the current loop matters. A true superconductor with zero resistance will actually force any magnetic flux to flow outside the loop. Loops with various resistances will let different amounts of current flow, and have different voltages.

All that's painted with a brush a meter wide! The actual physics get complicated. Fast.

Starting with this premises I wonder if is not more normal to isolate the PT from chassis and to return this currents with a wire to first decoupling and not through direct contact from chassis.

It's not, and for other reasons, as below.

I.m pretty sure is not a particular situation and perhaps direct coupling between end bells and core goes worthless. Bolted PT to chassis is a common practice walking this currents through chassis to star point. Why to keep transformers bolted to chassis and not isolate it and return this currents throuh a wire to correct point,please? Or with other words Why to reference the PT to the ground through chassis and not with a seperate wire as time we have to carry some return curent induced into the shield, please ?

It's important that you bolt the transformer to the chassis, for safety reasons, not audio reasons. The transformer may have an internal short, and this may let the laminations and/or end bell be connected to the mounting surface. It is important that this be electrically connected to the chassis so the fault current is carried away to the third-wire ground for safety reasons.

There is also the consideration of where the currents resulting from magnetic leakage flow. Many transformers have a solid copper "belly band" a wide strap of copper around the middle of the transformer. This is an proximation of the "superconductor loop", and has the effect of forcing leaking magnetic flux back into the transformer core. If you set up a conductor loop that includes the signal star ground point, you're actually conducting current into that ground point from the possibly ill-defined current loop that intercepts leakage flux. So isolating the transformer from the chassis and connecting it to the signal star ground point may actually make things noisier, depending on the physical path the wires take.

edit: supposed we have a functional earth point seems more normal to drain this currents to this point....but also with a wire, not through chassis. but in case of a "floating" mains...that.s rubbish in my understanding

Magnetic-field currents can only flow in the conductor loops that intercept the magnetic field flux. You can't drain current to a single point from a magnetic loop. You can induce a voltage across a loop where one section of the loop is open. In that case, the voltage appears across the ends of the conducting loop.

[catalin gramada]

Thanks all of You. Starts to have sense for me. My problem is I don.t beneffit by a earth point to my domestic mains so dig for a resonable solution. Thanks.

[catalin gramada]

allright. I did again some test.
In absence of mains supply earthing point it was again confirmed ihe best noise rejection found tied the electrostatic shield together with the center tap into secondary side of transformer. No to chassis , no to transformer end bells. It works better and is very effective not something I feel. I don.t know why it works better.Maybe someone have a explanation for this.... Thanks.

I want to remind: I have not earthing on my mains so in respect with its destination the electrostatic shield it is not functional and normal there is not reason to use it, but surprizenly tied to secondary center tap it reduced the noise and also the induced current between transformer end bells and center tap is reduced.This working against I understood. Why is happened that , please ?

[nickb]

allright. I did again some test.
In absence of mains supply earthing point it was again confirmed ihe best noise rejection found tied the electrostatic shield together with the center tap into secondary side of transformer. No to chassis , no to transformer end bells. It works better and is very effective not something I feel. I don.t know why it works better.Maybe someone have a explanation for this.... Thanks.

I want to remind: I have not earthing on my mains so in respect with its destination the electrostatic shield it is not functional and normal there is not reason to use it, but surprizenly tied to secondary center tap it reduced the noise and also the induced current between transformer end bells and center tap is reduced.This working against I understood. Why is happened that , please ?

Any noise current flowing through the screen to primary capacitance gets safely conducted to ground instead of coupling to the secondaries.

For the other bit, I'm guessing the primary is wound closest to the core and so the screen prevents coupling to the end bells too.

[catalin gramada]

HI. This is the way I did it. Normaly should not matter where el screen conection was done,I think ,(the traces between star-ground conection-end bells bolts are short, thick and large),so I don-t see any isolation condition, but seems manage better the currents in this way in my case

[trobbins]

I suggest that your amplifier should be connected to the AC mains supply protective earth. It should not be an option to disconnect it unless your equipment is specifically approved for such a connection (usually compliant to double insulated standards) scheme, or is allowable in your country. The thread should be closed if you are purposefully defeating a normally applied AC mains protective earth connection.

[catalin gramada]

Hi. It was wrong understanding. I didn.t propose such of ... I just try to found the best compromise as time my domestic mains supply did not benefit from a earting point. It was a standard in my country many years ago to fit 230v mains branchaments without earthing. And was legal and still is.



The beneffit using earth connection is evident and discussed as below. We discussed here how to best manage the induced couplings from a safe and noise pespective in absence of a functional earthing point. Thanks

edit : but don.t think was a standard just in my country... My Philips scope also have power cord branchament without earthed plugand have metalic case. But have an earth screw nut in the back to can conect a earth wire separately if is available. That means realy dangerous pick up measurement as time can put its chassis to high potentialby accident ...And is a Philips.
I don.t say it is normal , cause it is not and don.t recomand to anyone
Old normative,think, put working null and protection null together into main distribution pannel and run active and neutral phase for individual consumers....even without a standard unisense plug connector , so can be reversed without problem ...that.s rubbish

[trobbins]

Thanks for clarification - safety is the first aspect to settle.

A transformer would typically isolate any internal clamping bolts such that the internal bolts did not form a closed turn within the cores flux path - that is typically done with insulating washers or tubing on the bolts so that there is no loop electrical path. Not all transformers use that clamping method, but it is common.

The bell ends themselves would be electrically connecting to the core, and the bell ends and core would be reliably bolted and electrically connected to the equipment chassis. Are you wanting to somehow galvanically separate the bell ends from the core, and those parts from the chassis?

Ciao, Tim

[catalin gramada]

Thanks. Yes,the core is dc isolated as mounted thin paper gaskets between end bells and core till got 0 ohm resistance between end bells. The measured capacitance between end bells was 250 pF with my Fluke dmm (I think at 1 khz,but could be wrong) Of course the bolts was completly isolated with fiber washers and tube inside the core. The feet of end bells was cleaned by paint for a firm tight contact to chassis

[trobbins]

Is your experiment setup with the core isolated from bell-ends only being used for 'research', and not as an aim for normal operation? My concern is that the core is floating, which is not appropriate from a safety perspective.