Mullard-"Deacy" amp redrawn for you! (schem)

[col]

The transformers are Maplins miniature Audio transformers. Maplins numbers are HX82D (LT44 Driver)and LB14Q (LT700 output). The input transformer is 20kohm primary and 1KohmCT secondary, the output transformer is 1.2KohmCT primary and 3.2ohm secondary. If you follow the deacy amp schematic you can see where to put the centre tap. For the actual output I put one of the wires to ground (I think they were both brown on these transformers), I don't think it matters which one and a 100k log pot between ground and the other output wire, the centre of the pot being output. I don't have the latest catalogue but in the last one they were £1.73 each.

[col]

It was a 1 meg pot not 100k, and here is a link to my drawing;

http://s159.photobucket.com/albums/t159/col07768/?action=view&current=BM.jpg

Hope it helps

[brett]

Col
according to the Maplins site, those transformers have high impedance e.g. 12k:32ohms.  Is it the DC *resistance* that is 1.2k to 3.2 ohms?

In any case, for reasonable power output, you don't want to step the voltage down too much in the output transformer.  That 12k:32 or 1.2k:3.2 transformer has a voltage step-down of 19 times (the square root of the impedance ratio).  So even if the transistors are managing square rail-to-rail swings in the signal going into the transformer (e.g. 9 V), only 0.5 V would be coming out the other side.  0.5 V and 3 ohms is only 80 mW of potential output.

The output transformer that I used in this amp was 48 ohms : 8 ohms.  As you can see, the impedance ratio is much lower, so the the voltage step-down is about right (2.5 times).  It might be difficult to find a cheap, good transformer.  Mouser has this one, with a step-down of 6, and nice, low DC resistance (the lower the DC resistance the lower the losses): 42TM004 (see http://www.mouser.com/catalog/specsheets/XC-600134.pdf for details).  Using a reasonable supply voltage will help (e.g. 18 or 24V).  Most PNP germanium output devices will tolerate this amount of voltage.  With small heatsinks, two AC128s should produce 1W without melting, though one will probably do more work and get hotter than the other..
cheers

[Steben]

With small heatsinks, two AC128s should produce 1W without melting, though one will probably do more work and get hotter than the other..
cheers

That's because the gains are different. This results in assymetric amplification and clipping a la Fuzz Face.
With a matched pair it is more standard distortion.

[brett]

Hi

With a matched pair

and equal loads and equal bias and equal heatsinks, etc

Only likely with modern Si transistors and circuitry. In my experience, the DC resistance of the halves of the secondary of the coupling amps can vary quite a bit. But maybe that's coz I'm only paying $2 to $5.

cheers

[Steben]

The 42TM series have 200mW max power...
To achieve 1W I think you need Hammond 149's...
It is indeed very difficult to find an affordable transformer I guess.