Ever tried the type 0 toroids?

[tinitus]

Nice info
But could be hard to get
I doubt they sell to private customer
And not in small quantities either

Around here such things usually works the way that IF they have e stock of a certain item fore a big customer, it may be possible to buy a few
Or they sometimes forward you to one their big customers
If lucky you can beg to buy a few out the backdoor

I once bought a huge 3foot heatsink that way
On a request on custom trafos, the same thing, but really way too expencive

[stoc005]

Class D audio amplifier

I am developing a full bridge Class D amp and will be using the gapped toroids for the filters. That is why I know a bit about these parts. I do have a friend that can get samples from Ferroxcube. I have also developed an XL (2007) spreadsheet that automates that Ferroxcube app. note for class-D filter component calculations. It even tells you how many turns to use on that core of the gapped toroid and helps with finding the wire size to use. Helps with wire losses too. I'd be willing to send you the spreadsheet but it is at work and I won't be there until monday. It is not documented well and somewhat of a work in progress since it was for my own consumption. Send me an email at: stoc005<at>att.net with your address and I will send it. I will convert it to Excel 1998 since 2007 is so bizarre. I'll send a schematic of the amp so far as well. It's kind of rough. It is similar to the UcD type amps with analog feedback taken after the L-C filter plus it is self-oscillating at 300-500Khz.

On a related note about component values for output filters. With the LTSpice simulations I've done on these filters and the amps that drive them, I've pretty much decided that the correct impedance to design towards is 6 ohms. An 8 ohm filter is just wrong for a 4 ohm load and a 4 ohm filter is wrong for an 8 ohm load. 6 ohms seems like a reasonable compromise since most single speakers start out at around 8 ohms and end up a bit less, after the crossover and zobel(s) networks are added in the speaker. I use zobels for crossover design since it levels the impedance you have to work with for the crossover filters. With that taken care of, a Class D amp should only be connected to one speaker at a time, IMHO. Paralleled speaker pairs on the amplifier should not be allowed, I think.
Bill

[Pafi]

Helps with wire losses too.

Bill!

How do you calculate the wire loss caused by eddie current induced by leakage from the gap?

[sendler]

Originally Posted by stoc005

http://www.ferroxcube.com/appl/info/gaptoroids.pdf

Use a TN26/11-3C20-A113 if you are going to order some. It is the largest and will handle 100W or so. Read the app. note and find out why you want the big one.....

Ferroxcube can be bought at their distributors but I see no stock on the above PN at any of their distributors. Bummer. I am going to need (4) pcs soon.

.
We might have to settle for one of these which are in stock.
TN26/11-3C20-A201
TN23/14/7-4C65
Also, I wonder why they show their inductors wound so loosely?

[sendler]

The toroid air cores do exhibit much better containment than the cylinders but still leak a little more than the T106-2 cores. Even the type 2 cores are not as silent as you might think. Are they leaking from the leads? The original little smd shielded bobbins were dead silent. Anyone care to try their amp and report back? Maybe all of our hand wound cores leak a little. I am listening with an AM radio tuned between stations to the switching frequency. Turn your amp on and off and listen to the difference.
.
This amp is giving me the best sound I have ever had though, and it is the cheapest, running off of a single $32.00 smps.
.

[Andrew Eckhardt]

@sendler

Try some air toroids out of more reasonably sized wire and pop them into a set of hacked off pipe caps and solder them together.

No iron distortion.
No radiation.

It looks like you're off the deep end of milliohms with that wire.

Edit: You're never going to get a leak-free toroid because the windings are not perfectly uniform and everywhere.
Edit2: What is the RDSon of the output transistors?

[sendler]

Originally Posted by sendler

Anyone care to try their amp and report back? Maybe all of our hand wound cores leak a little. I am listening with an AM radio tuned between stations to the switching frequency. Turn your amp on and off and listen to the difference.

Nobody has an AM radio?

[stoc005]

I think air core inductors are going in the wrong direction. You want a very low DC resistance for these inductors. One way to achieve that is to use large diameter wire. Next, use a core that produces the desired inductance with as few turns as possible. Now, pick a magnetic material and mechanical configuration that allows you to create the inductor you desire without causing saturation of the magnetic material at the level of DC/AC (audio) current that it will see. Plus the core should confine the magnetic field such that radiated energy is tolerable. If the inductance collapses because of magnetic saturation, the filter no longer works as designed.
Air core inductors use too much wire. The DC resistance is higher than necessary. This will cause a lower damping factor on the output of the amplifier. The bass will be less controlled (dare I say "flabby" or "less tight"??) than it could be with a higher damping factor. Yes, people listen to AM, which is why you need to keep the stray magnetic fields under control. Air cores just radiate too much, unless you completely shield them magnetically.

[sendler]

Originally Posted by stoc005

I think air core inductors are going in the wrong direction. You want a very low DC resistance for these inductors. One way to achieve that is to use large diameter wire. Next, use a core that produces the desired inductance with as few turns as possible. Now, pick a magnetic material and mechanical configuration that allows you to create the inductor you desire without causing saturation of the magnetic material at the level of DC/AC (audio) current that it will see. Plus the core should confine the magnetic field such that radiated energy is tolerable. If the inductance collapses because of magnetic saturation, the filter no longer works as designed.
Air core inductors use too much wire. The DC resistance is higher than necessary. This will cause a lower damping factor on the output of the amplifier. The bass will be less controlled (dare I say "flabby" or "less tight"??) than it could be with a higher damping factor. Yes, people listen to AM, which is why you need to keep the stray magnetic fields under control. Air cores just radiate too much, unless you completely shield them magnetically.

.
That's the problem. You can't think what something will sound like. You have to build the circuits and listen to them. That's what the whole DIYaudio forum is about and why modified gear can sound way better than stock. Whether it's a DAC, an amp, a cable, passive component, whatever. If the shortest wire was the best then every loudspeaker would have a massive core with a few turns of wire on it for all of their inductors. But they don't. They all use air core unless the value is too large. 10,000 times the values we are dealing with in class D.
I have three identical amp boards. One with the 20ga. type 2 toroids and one with the 20ga. cylindrical air cores and the air cores sound much more transparent even though I can't use them because of EMI. The big 14ga. air core toroids destroy the sonics of either one and have emissions on par with the type 2. It's not just me. The forums are full of class D people who have listened various (and unknowingly noisy) air cores and I have yet to read a single post where they preferred the type 2 core which is otherwise the de facto. Too bad the forum search doesn't consider "air" to be an important enough word to find as it would be easier to read about.

[stinius]

Bob Cordell about classD

here

and

here

Cheers
stinius