Ever tried the type 0 toroids?

[tinitus]

What I dont understand, why is such a thing placed on the board in the middle of sensible curcuits
Wouldnt it be better placed between board and output terminals
Well, maybe its obvious I have no idea what its about

But I am interested because my next woofer amps might be classD
And I have always thought the used output inductors looked pretty awful

[phase_accurate]

The leakage comes from imperfections, for example when you wind the turns in a way that it goes around along the core, then you make a single turn choke in excess the toroid. Not equal distances are the other source of leakage.

Even if a toroidal coil is wound perfectly it still has one single loop that is causing a stray field. And this is unavidable. Don't forget about that !

Regards

Charles

[Pafi]

I didn't forget, I wrote about it! (Maybe I wasn't totally clear.) But actually it is avoidable! With two layers. Or it can be minimized with only one layer, if you split the coil into two symmetric halves, then connect them paralell (very small values) or series (higher values). You can decrease it further, use your imagination!

[Lostcause]

Originally Posted by tinitus

What I dont understand, why is such a thing placed on the board in the middle of sensible curcuits

I thought the same thing and started placing them under the PCB into one of those small shielded boxes you can get and earthing it to the case... which is in turn earthed to the UK third pin.... no idea if it makes any difference but I feel better for doing it

[Pafi]

Charles!

Another simple way to eliminate the flux leakage caused by the loop: before winding the coil make a loop in the other direction!

[Saturnus]

Originally Posted by Pafi

Another simple way to eliminate the flux leakage caused by the loop: before winding the coil make a loop in the other direction!

You're still under the misconception that you can eliminate the flux leakage entirely. Let me assure you cannot. And while that may not make a difference in low frequency applications, it certainly does in high frequency ones like this.

The traditional advice for output trafos in class-D is to use heavily cored and heavily shielded trafos. Stick to it.

[Pafi]

You're still under the misconception that you can eliminate the flux leakage entirely. Let me assure you cannot

No, obviously I haven't said (or thought) this. Please try to understand what I wrote! Simply denying is pointless.

What is your problem exactly? Does the knowledge about leakage mechanism hurt you? Do you sell magnetic shields? Or what?

The traditional advice for output trafos in class-D is to use heavily cored and heavily shielded trafos. Stick to it.

I don't know why have you told this, and what does it mean.

[Saturnus]

Originally Posted by Pafi

I don't know why have you told this, and what does it mean.

It means that high frequency operating cores like in the output of a class-D will always function best the less parasitic resistance and capacitance it has, both are ensured by correctly coring the trafo. Air cores of any type should be avoided at all cost to achieve the best sound.

[Pafi]

ClassD what? ClassD amplifier? Then why did you write trafo? ClassD power supply (it contains trafo indeed)? How does it come here?

Parasitic resistance and capacitance of a core? I think you are lost somewhere!

always function best the less parasitic resistance and capacitance it has

Doubtfully. You forgot about many important things.

both are ensured by correctly coring the trafo

And by correctly wind it, with the correct wire, with the correct insulation.

Air cores of any type should be avoided at all cost to achieve the best sound.

Tell this to Sendler, I'd ended up the idea of using air core many years ago! (Unless it's for an induction heater.)

[Saturnus]

Originally Posted by Pafi

ClassD what?

Output coil! If you don't understand that this is the class-D forum, and as such any discussion about cores and coils herein would be relevant only to the output filter coil(s), then I can't help you further.