What is the reason some EI transformers are welded with a strip along the height of the stack?

[SoulFetish]

Quiet simply, I was of the understanding that EI transformers are constructed using thin sheets of silicon steel, insulated from each other, so as to increase the electrical resistance of the core to current flow without inhibiting the flow of magnetic flux. All this to reduce the core losses caused by eddy currents, correct?
I find it interesting that some laminated transformers have a welded stripe across the stack on the outside of the core (Marshall transformers come to mind). While I would think this would break the layers of insulation across the stack, I wonder if it serves a similar purpose as a copper band wrapped around the outside of transformer would. So, I ask our transformer guys out there to shed some light on this (you know who you are).
The first example I remember seeing something like this is when I was servicing some Peavey amps, like the...... shit, I can't remember the name of the model right now. The one which is like a better answer to Marshall's 800. In any case, I was surprised to see that they welded the stack right to a flat mounting plate. I'm pretty sure they did this on both the mains and the output transformer.

[loudthud]

Look carefully where the weld is. Does it just weld E's to I's on one side, or does it weld E's to E's that are facing the opposite directions ?

[Helmholtz]

I've seen this only with cheaper PTs, where all the I's are on one side and all the E's on the other. The 2 packs are attached by welding.
The welding strip will increase eddy current (core) losses most notably at higher frequencies (OT).
To me it looks like a cost/time saving production mehod.

Preferably EI cores are stacked alternatingly to minimize the effective airgap. That's labor-intensive.
Smaller airgap means higher inductance, less leakage and magnetizing current and with an OT better bass response.

The welding might suggest zero airgap, but at normal operation the welding strip can be expected to saturate and disappear from the "equation".

The position/orientation of the welding strip (as well as the poor conductivity of silcon steel) does not suggest outer leakage attenuation like a copper flux band.

[trobbins]

Agreed. I've also seen examples in microwave oven transformers. It can also suppress audible noise and vibration. It's not uncommon on large industrial chokes and transformers (along with welding on angle iron for mounting), especially for chokes where the gap can be quite large and difficult to make inaudible.

[J M Fahey]

IF aligned with the "I" leg centerline, that is a minimum (almost zero) flux point, so a full weld there hurts nothing magnetically and supposedly avoids vibration/buzz.

The other nil flux points are at rectangle corners, youŽll see Peavey PTs welded to mounting plates there, same thing.

[Helmholtz]

Magnetic flux always takes the path of least reluctance.
So if the welded joint bridges even the smallest airgap, flux distribution will changed and flux density in the joint will be high.
That's why I assume it will saturate at normal operation.

Furthermore the high welding temperature is likely to degrade the magnetic properties of the laminations leading to a decrease of effective cross-section. This effect is probably not very significant as the outer legs carry only 50% of total flux each.

Last not least, a welded core doesn't allow for easy disassembly/repair/rewinding without cutting the core.

I guess the welding may help to reduce core buzz. - if that's an issue.
In my job I've seen several transformer buzz issues where the buzz clearly was caused by the windings and not by the (ferrite) core.
Solution was vacuum impregnation of the complete transformer.