Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. The sunn still shines online!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  

How to make an Output Transformer??

1/25/2006 8:57 AM
How to make an Output Transformer??
It is possible for someone to build his own OT?  
I've seen R.G article about building output transformers, but don't have a clue how it can be done. For example, does the secondary winding's should be "sandwiched" with the primary windings? How about the various layers of winding's? I already unwinded some power transformer's, but the secondary and primary windings had separate bobins. I can't reach to a conclusion.  
Someone knows some site with info and pictures about that?  
1/25/2006 9:07 AM
Dai H.

should be a couple of links in the right hand side window
1/25/2006 11:04 AM

Thanks for your reply!  
I already founded on that window two links, one was the R.G article and other was a broken link. But I've missed these two:  
These are excellent pages, although I still don't know very much about the practical level.  
If windings should be intervaled and how it is done. If there should be any varnish applied while winding it. How to make multiple taps, etc.  
I guess I'm too dumb...
1/27/2006 4:42 PM
Todd Hepler  
Search the page for "Electronic Transformers and Circuits"  
About 24 mb pdf file. This has some good design data, although it leaves out most of the practical construcion techniques.  
1/27/2006 5:16 PM
Dai H.

this has a lot of detail but... it's all in Japanese:
1/27/2006 6:23 PM
I downloaded that book yesterday, but that's only theory.  
I've also founded that Japanese site two days ago, the pictures are very cool, but I just don't understand the text so it's useless.  
But thanks anyway!
1/27/2006 7:04 PM
The hardest part of making an output transformer is determining and then finding the right core.  
In general, you want a a core that is 3-4 times as big a cross sectional area as the core for an equal-powered 60Hz transfomer and about the same core stack - that is, close to square.  
You also want thinner laminations than a 60Hz if you can get them.  
Once you have that in hand, you calculate the primary inductance you need to keep your bandwidth up at the low end. Your inductive reactance is the primary determiner of the low end rolloff. If you want your transformer no more than -3/-1/-0.1db at whatever low end you need, and you know the target impedance you want, it's easy.  
If Z = 4400 (for example) and you want no more than 3db down, then Xl in parallel with Z is 1/2.  
So 2*pi*F*L = Z, and L = Z/2*pi*F where F is the frequency.  
That's the plate to plate inductance. So the number of turns you get is going to be split in two for the CT, assuming AB pushpull.  
Now you have a problem. You need to calculate the number of turns. If you have the lamination maker's literature, they'll tell you the Al (A-sub-L, or inductance constant) for the core, in henries per 1000turns, usually. If you have random core laminations, you'll have to determine it by test. Wind 100T on the core and measure L. L is proportional to turns squared, so take the square root of the ratio of the inductance you want for Lp and the L you measured, then multiply by the number of turns you used in your test. That's the end-to-end Np.  
To get secondary turns, take the impedance ratio you're trying to get (e.g. 4400 to 8) and take the square root. For example 4400/8= 550, so the turns ratio is 23.45 primary turns to one secondary turn. Calculate Ns.  
Once that's done, all you have to do is wind it.  
Unfortunately *how* you wind it makes a huge difference in high frequency response. The leakage inductance is a property of the way it's wound, and acts like it's in series with the primary winding.  
There are several ways to wind.  
Simplest is all the primary, then all the secondary (or vice versa). That maximizes leakage inductance, the wrong way.  
Next best: wind half the primary turns, all of the secondary, then the other half primary. That cuts Ll by four, and not many guitar amps get better than this.  
Next best: wind half of one half of the primary, 1/3 of the secondary, half of the other half of the primary, 1/3 of the secondary, half or the first half of the primary, 1/3 of the secondary, then the last half of the other primary.  
That cuts leakage by 25 and is a dandy fine way to do it.  
Hifi cores get more complicated.  
You should apportion 1/2 of the available winding window area to primary, and half to secondary, and fit your wires into those areas.  
The winding window is the area inside the openings of the core laminations. Out of that you have to take the bobbin area, and the end margins (don't wind all the way to the end of the window - 85% is a great usage of the length along the coil) and the space at the top of the winding, which ought to be about 1/16 inch.  
Divide up the window to see how much area you have to fit the Np and Ns turns in, then calculate wire sizes to fit those areas. Include the insulation between layers in your build up calculations.  
Then wind it, stack it and try it out.  
You won't like your first one.  
Try again.  
You won't like your second one.  
Try again.  
You'll like your third one fine, because you're tired.  
Try again.  
Your fourth and following ones will be acceptable and get better.

  Page 1 of 5 Next> Last Page>>