Pattern for power xmfr magnetic leaking ?

[mahleriana]

Hello,
Has anyone has information about the directions an unshielded EI core power transformer leaks, magnetically speaking ? The solder lugs on mine make it can't be easily shielded...and I have some noise I think it is related to this.
Thanks in advance
best regards
eric

[tomchr]

Eric,

The magnetic field of an EI core follows the same rules as any other magnetic field. Wikipedia has some good illustrations of this: Magnet - Wikipedia, the free encyclopedia and here: Leakage inductance - Wikipedia, the free encyclopedia

Basically, you can figure the transformer will leak field lines that originate perpendicular to the windings. I would also expect some leakage field at the corners of the core. Shielding the solder lugs will do absolutely nothing for shielding. And even a metal shield needs to be very thick (>10 mm) to be worthwhile for magnetic shielding. Unless, you have access to mu-metal...

~Tom

[trobbins]

Below link is to a good simulation schematic of the stray magnetic field emanating from an E-I configuration transformer with an outer leg winding. Your PT would most likely have windings on middle leg, so you would need to do some mental gymnastics to move the stray field emanate from the centre leg. Steel chassis or shield plate would redirect the field - although a thin chassis/plate won't significantly redirect field lines hitting it at right angles.

http://dalmura.com.au/projects/Strayfield.pdf

Ciao, Tim

[BudP]

If you make a mag field search coil you will find that the long side of the core emits the strongest field, in plane with the individual laminations. The short side will emit a field at 1/8 of the long side's strength. This will show up on the trailing slope of the AC waveform. Directly perpendicular to the face of the laminations, the coil will emit a field 1/8 the strength of the short side of the core. This will show up on the leading slope of the AC waveform. This is also the most invasive of the various emitted fields. There is measurable field strength from the angles between core and coil, but they are off the scale from the rest.

To contain fields you must have perfect boxes, no gaps anywhere. It is easier to shove them away from a sensitive area with ferrous materials with poor permeability, like heat treated cold rolled steel. ).06 thick CRS heated to Osteonite temperature will get you -100 db of field suppression. Typical 22 gauge will provide about -40 db. Another tack is to suck the field off in another direction with high permeability materials, like the conetic alloys. Yet another is to form a near chassis width triangle out of high perm material and suspend it between emitter and receiver. You will need about 1/4 inch clearance on either end. To use a core suppression band around the core it should extend beyond the core stack by 0.050 inch on either side of the core faces.

Keep in mind that stray flux will enter and propagate down the bends in a CRS chassis, only to re-emit at some point you don't want it to. Again, heat treating the chassis is the solution.


Bud

[trobbins]

Bud, do you bring the AC mains cord in to your amp in fixed or flexible steel conduit?

[BudP]

To date I haven't done anything to quell fields from the AudoPrism Debut amp I use. I will have to sooner or later, as my next speaker system will jump from 93 db efficient to 96 db. The current set up does have audible hum directly adjacent to the main full range driver, though none from the plate amp driven TL subs, so I will be investigating all manner of power supply remedies for the upper frequency range drivers.

Bud

[danFrank]

Originally Posted by BudP

If you make a mag field search coil you will find that the long side of the core emits the strongest field, in plane with the individual laminations. The short side will emit a field at 1/8 of the long side's strength. This will show up on the trailing slope of the AC waveform. Directly perpendicular to the face of the laminations, the coil will emit a field 1/8 the strength of the short side of the core. This will show up on the leading slope of the AC waveform. This is also the most invasive of the various emitted fields.
Bud

Hi everybody,
I have a question that is along the lines of this post; I just can't seem to wrap my head around the geometry that you all are describing above...
I want to build a stereo amp and the space on the chassis is rather tight. I want to know the best way to situate the output transformers so their magnetic flux will LEAST affect the pre tubes that are situated very close by to the output transformers. The choice is picture "A" or picture "B".
The power transformer is at the left, the middle of the chassis is for the output tubes and the 2 transformers at the right will be the outputs with the pre tubes (pre drivers) in between the 2 output transformers.
The pre driver tubes will be 6DJ8 or 6CG7 tubes, the output tubes will be 811A's in push-pull. This will be a stereo version of the Gotham Audio's cutter amp that they made back in the 50-60's. I have built a mono version of this amp and love the way it sounds, now I want to do a stereo version.
Thanks in advance...
Daniel

PS. One more question.... If the long side of the transformer core emits the strongest field, then why does the sides perpendicular to the core emit the most invasive of the magnetic fields? I don't follow. Wouldn't the side that emits the strongest field be the side that is most invasive?

[BudP]

I have no idea why the coil generated flux is more invasive. That's just an empirical finding from 35 years of helping other folks track down their noise issues. Were I to build the amp you show in your pictures and wanted the least interference and could not mount the tubes on the other side of the main board or just sticking up through pretty tight holes in the board I would place them as shown in the picture.

I would strongly suggest a bread board first and then a perfboard with sockets and OPT's mounted on the desired metal plate just so you can move the tubes around.

Bud

[mahleriana]

If I understand well, a good way to reduce noise from magnetic field would be to have the "showing" sides of an unshielded xmfr up/down, and if possible the longest side parallel to electronic components ? (see drawing)

[jackinnj]

Originally Posted by tomchr

And even a metal shield needs to be very thick (>10 mm) to be worthwhile for magnetic shielding. Unless, you have access to mu-metal...

~Tom

You can get sample packs from mU-shield -- although you have to pay for them.