OTB - Replacing Burned-Out P-P Audio Transformers and other Lore

Replacing Burned-Out P-P Audio Transformers and other Lore

Edited by Ken Owens
478 Sycamore Dr.
Circleville, OH 43113
radiowd11@yahoo.com
Please send correspondence directly to the
above address, including SASE for reply.

The radio restorer often finds that half of the driver transformer secondary is open in transformer-coupled P-P output stages, causing one of the tubes to get no signal. Triode tubes require a lot of driving voltage. The rule of thumb is that maximum output occurs when the driving voltage equals the grid bias. Grid bias in the typical P-P 45 output stage runs 50-65V. For adequate volume, the driver transformer often needs a step-up ratio as high as 10:1.

Fig. 1. Circuit for temporary replacement of a push-pull output driver transformer with half of its secondary open.

Such transformers are not easy to find, nor are they easy to replace in sets like the Atwater Kent Models 55 and 60 where everything is potted in tar. Langford-Smith in [italic]Radiotron Designer's Handbook, (4th Ed.) gives a circuit for P-P pentode tubes where signal for the lower tube is taken from the unbypassed screen of the upper tube. Philco used this circuit in their Model 40-180 and Echophone used it with triodes in their 1931 Model F.

The signal on the screen (or plate) of the upper tube is 180? out of phase with the grid signal. If some of this signal is used to drive the grid of the lower tube, phase inversion is built-in. Distortion is high because the inverted signal is taken off before cancellation in the P-P output transformer.

I simulated an open secondary half in an AK Model 55 by disconnecting the transformer lead to one tube at X (see Figure 1). Then I installed the R-C network shown to obtain drive for the lower tube from the plate of the upper one. After trying different values for the resistors and capacitors, I arrived at the indicated values. It works. The 'scope shows the expected higher distortion compared to the original circuit, but it doesn't sound bad. The .001 µF capacitor is needed to suppress a tendency towards high frequency oscillation at high output levels.

This fix will get your set going until you can find a replacement transformer. The added components can easily be removed at that time and the set restored to original.

* * * *

Atwater Kent battery sets used only one carbon resistor and, until 1929, the AC sets used only 3. These consisted of a carbon element in a glass tube with metal ferrules on the ends. They were held in clips like a fuse. In 1929, A-K redesigned their chassis and went to resistors with the carbon element inside a white ceramic tube with lead wires embedded in cast metal end caps. They used many more resistors and devised a house code consisting of colored bands to identify them. Since each color combination is unique, the order in which the colors are read is not important. The code is shown below.

ATWATER KENT CARBON RESISTOR COLOR CODE (1932)

Black 65K ohms Green 2Meg ohms

Black-blue 1100 ohms Green-blue 4000 ohms

Black-green 65K ohms Green-red 3300 ohms

Black-purple 500K ohms Green-yellow 900K ohms

Black-red 20K ohms

Black-red-blue 1100 ohms Maroon 10K ohms

Black-yellow 40K ohms Maroon-blue 300 ohms

Black-yellow-red 50K ohms

Purple 6000 ohms

Blue 100K ohms Purple-red 12.5K ohms

Blue-gray 1Meg ohms Purple-yellow 12.5K ohms

Blue-red-green 500 ohms

Blue-yellow 5000 ohms Red 12.5K ohms

Red-blue 100K ohms

Gray 30K ohms Red-gray 800K ohms

Gray-green 15K ohms Red-yellow 250K ohms

Gray-yellow 15K ohms

White 40K ohms

Yellow 500 ohms

In the case of a single color, the whole resistor body may be painted or it may be just a band on the white tube.

These resistors are as unstable as any molded carbon resistor and will often show unacceptable increases in value. Information is from the Atwater Kent manual, Electrical Values of Resistors and Condensers in All Receivers, 1924-1932.

* * * *

Another interesting "fix" comes from George Capen (OTB, Vol. 22, #4, Mar 1982). He describes a transistor amplifier designed to replace audio transformers in audio amplifiers. It is shown in Fig. 2. Parts values are:

	R1, R2 - 22k		C1 - 5 uF Tantalum
	R3 - 150k			C2 - 20 uF Tantalum
	R4 - 15K			C3 - .01 uF/100V
	R5 - 18k			D1 - 12V, 1W Zener diode
	R6 - 1500 ohms		Q1 - 2N3905
	R7 - 220k

Fig. 2. Transistor circuit for replacing burned-out interstage audio transformers. See text for parts values.

Fig. 3. A completed transformer replacement using the circuit of Figure 2.

I built this circuit and tested it with the following changes. Tantalum capacitors were probably specified for their small size, but I used ordinary aluminum ones. The rating of C3 was increased to 400V because the potential across it can exceed 100V when the plate supply is over 100V. A higher gain 2N3906 was used because I didn't have a 2N3905 on hand.

The operating power comes from the voltage drop across R1. It is filtered by R2 and C2 and stabilized against fluctuations by the Zener diode. Signal is coupled through C1 to the base of the 2N3906 in a common emitter amplifier circuit and then to the grid of the following tube through C3. R7 provides the grid return path.

Capen states that the gain of the circuit equals that of a 1:3 audio transformer. I found the gain to be approximately 8.2 when used between two 01A tubes with 90V on the plates and -4.5 V bias. Experimentation showed that R4 should be reduced to 8.2K ohm for a gain of 5 and to 5.1K ohm for a gain of 3. These gains approximate those of the most commonly used transformers. Distortion occurs when the input signal exceeds 1V. The size of the completed unit can be seen in Fig. 3.

* * * *

Michael Rakochy (Allentown, PA) sent in some useful tips. He uses a little transistor radio as a signal source to troubleshoot audio amplifiers in radios. He gets the signal from the earphone jack on the transistor radio. He uses a plug with a 8-10 ohm resistor across it to provide a load and then a small blocking capacitor in series with the test leads to keep set DC off the transistor radio. It can be attached to plate or grid terminals to test the amplifier with actual program material.

Michael also points out that factory-wired Variacs have the fuse in the output lead. That's because, depending on the setting of the unit, very large output currents can be drawn without exceeding the limits of a fuse placed at the input. You could burn out a portion of the winding and never blow the fuse. Good advice for those who have Variacs incorporated into test equipment.

BACK TO OTB TABLE OF CONTENTS