How to build a transformer and find core

[Cliff_J]

In my car the OEM amplifier runs dual-voice coil subs. Its a SMT PCB with all ICs on even the outputs, no SMPS so its just 12V for the rails and the outputs are already bridged. Easy to see why many junk it and use an aftermarket amplifier come upgrade time. But I want to try to use it as is on a single voice coil sub of higher impedance and don't know the formulas to build a transformer (and would like to do it cheap by hand winding).

So I need two primaries to feed a single secondary with a 2:1 turns ratio. I'm figuring 20W RMS (maybe 30W) but have been told the low frequencies of a sub make the transformer large (and I thought it was more a function of VA capacity/saturation level than freq).

I'd like to think I could ballpark this thing and buy a cheap torrid doughnut and some 22ga enameled wire and wrap a couple hundred turns to start testing and then scope it to see how bad it performs. So can anyone give me a idea where to find a cheap doughnut (or U-shaped laminate core) and an idea what gauge/turns to start with?

I'm guessing that if we start with say 60Hz like mains power will the performance suffer greatly at 30Hz, and if we start with 30Hz then the inductance would need to be managed to keep the performance in-check at 60Hz? Ahh, I have a ton of questions but will try to keep this short so any help is appreciated.

 

[Antiphon]

I don't know where you could buy a core alone, but if you go to this company
http://www.amveco.com/
they will make you one for a reasonable cost (probably ~$100.) Also, they
may tell you how to get a core or sell you one.

 

[Cliff_J]

Antiphon - thanks, that was a little more than I wanted to spend but I always did appreciate the beauty of a nice torroid.

Ok found a couple sources of cores through the Thomas Registry - well a couple that sell single units anyways.

Also found a resource that helps with some of the requirements for an audio transformer in an article written by Bill Whitlock of the Jensen Transformer company (who makes $300 signal transformers with excellent performance). Its focused on low-signal performance and gives no formulas for determining windings though.

In the article he details that the lower frequency performance is a product of inductance but to keep phasing correct the -3db point should be kept very far beyond the intended range, so far below he referenced primary windings of 1000 henries! And that high frequency performance is a product of capacitance in the windings.

Little discouraged thus far, and when he mentioned how poorly regular silicon steel is with hysteresis distortion I was more discouraged, but I'll explain why I'd like to continue exploring. At this rate, the idea is dwindling but curiosity is persistent and the challenge of making it work keeps a neat goal. Especially in a day when car audio amplifiers with peak wattage ratings that are outrageous can be found at just about any retail store.

Our hearing is poor at low frequencies for both amplitude (Fletcher-Munson) and distorition with even 10% THD representing the threshold for audibility. Frequency response below even 40Hz is difficult even in a car with the advantages it offers in terms of bass reproduction. So within the envelope of frequency response of 40-60Hz with 10% THD this narrow bandwidth and high distortion I'd hope (really like to know but still hope it works) I could make it work.

In short, this is more a hobby experiment than a precision exercise. Can anyone even give me a basic transformer windings calculation formula like would be used for a regular old mains transformer? Or what would be good search terms on google. Thanks in advance.

 

[Cliff_J]

Nevermind, I finally found some good sites after the internet search terms were sorted out. Seems the inclusion of the word "tube" helps tremendously and was much easier than trying to find an old article on winding a eletrostatic transformer.

Ok, now off to try to find surplus suplpies for cheap...