toroid for car amp smps

[wrenchone]

The power handling will depend on the amount of wire you can get on the toroid. You will need to scale the number of primary turns to get an acceptable flux swing. Np = 10^8*Vin/4*Bp*Ae*f , where Np is the number of primary turns (1/2 of a center-tapped winding), Vin is the input voltage, Bp is the peak flux swing in one direction, Ae is the core effective area, and f is the switching frequency. This assumes you are using a standard push-pull PWM or square wave drive circuit. A good starting value for Bp is about 2000-2500 Gauss.

A long while back I actually used a saturating core inverter to power my car stereo.The key to making it work right was to place an inductor in series with the transformer primary center tap to limit the shoot-through current in the switches when the transformer core saturated. Since the transformer was designed to saturate, it was very small. I don't know what the actual power rating of the supply was, but it generated +/- 28V to drive a discrete power amp of my own design, anfd it could bottom any speaker I put in the rear deck of my car.

[tennisballg]

Now is that Np=(10E8*Vin)/(4*Bp*Ae*f ) ? Otherwise the number seems way too large. Ae= effective area, is that a cross section of the toroid, simply the surface area of the entire toroid, in centimeters, meters, millimeters, or some other area? Now that gives me the number of turns to use, I was planning on using 4, since that seems to be the common practice for car SMPS, but it's nice to have the actual calculations! However, how does that relate to power?

[sss]

ok lets say my primary coil is 4 turns each (center tapped) and
i'm using a toroid

now when winding the primary i need to cover all the surface of the core ,right?
so how should it be ?
4 turns covering the core and then another 4 turns above it ,
or 4 turns covering half the core and another 4 turns covering the other half
thanks

[Eva]

Use 'bifilar' winding : Make both primeries identical and spread them both around the entire core

This example shows two 6 turn primeries center tapped [4 wires each] :

[sss]

thanks Eva looks nice

[tennisballg]

(not that my power Q is resolved, but...) Ok, now I am a little confused. What I am seeing is this, correct ? (primaries are different colors, these examples are 4 turns, center tap, 4 turns)



For some reason, I thought that the primaries had to be as identical as possible, example:



It could be that they're really the same, but the first example covers the area of the toroid much more nicely if you have few windings, by interlacing the primaries rather than sandwiching them into one bundle. I would prefer to do it the first way, especially with thick wires or multiple wires as Eva has shown, I just didn't know that it was close enough to 'identical' windings. Or perhaps since each of your primaries is four wires, you are winding two from one primary with two of the other, then interlacing the other two from each? (It doesn't look like it though, just grasping at straws)

[Pabo]

I suggest that you go for a bobin based transformer from Ferroxcube or Siemens. One good example is E32.

If you run at about 70kHz you can do with only two primary turns made out of copper foil (20x0,2mm). Wind the first primary closest to the bobin. Then wind the two secondaries interlieved on one layer. Last wind the second primary.

You will end up with a primary resistance of about 3mohm and really good coupling between the primary and secondaries.

With such few turns you can never design a good toroid transformer because you simply cannot get the coupling that you get with foils. The absolute optimum would be to wind as I described above but with only one primary winding. That way you would have the entire primary voltage "seeing" the secondary winding. This is possible if you increase the frequency to 150kHz or if you increase the core area. If it is absolutely necessary to obtain low building height you can allways remove the pins of the bobin and only use the cylinder in the middle.

[lumanauw]

Hi Wrenchone,

The power handling will depend on the amount of wire you can get on the toroid.

Thanks for answering that. This also have been in my head for many years. Looking at any magnetic equation, Faraday equation, including the one in mag-inc handbook, I drew conclusion that the limit of power handling of toroidal ferrite is how much wire you can put in them. Nothing to do with Ae or window area. But I dont believe myself for my conclusion.

Your short answer has clarify it. Thank you

[sss]

Originally posted by lumanauw
But I dont believe myself for my conclusion.

i dont believe it either

it cant be true because its cant be 100% efficient , if thats so then the only losses will be in the copper wires .
i never checked it but i think the core gets hot while under heavy load , right ?

[tennisballg]

But that's still too vague, there's no " this toroid will handle this many VA". Or perhaps they meant that one is supposed to tweak the equation they gave to get the highest number of Np, at which you will acheive the highest power possible for the toroid. Could it really be that most DIY people just guess until they find one that doesn't saturate under the conditions they need?