Amelia Earhart Search Forum => Radio Reflections => Topic started by: Chuck Varney on July 11, 2011, 11:11:37 AM

Title: Western Electric 13C transmitter harmonics
Post by: Chuck Varney on July 11, 2011, 11:11:37 AM
This is for Bob Brandenburg.


In a forum exchange with Hue Miller on 27 Oct 2000, 10:55:27 EDT (http://tighar.org/Projects/Earhart/Archives/Forum/Forum_Archives/200010.txt), you  wrote:

". . .When all the work is done, I'll give the details to Ric for posting on the TIGHAR web site where they can be reviewed and critiqued by forum members.  In a very fundamental way, my work is only the beginning of the process of finding the truth.  The informed insights and opinions of forum members will be vital additions which will help illuminate the way ahead."

The context was work you were doing on transmitter harmonics and propagation.

If you are still open to insights from others, I have a number of them relating to your  WE-13C Transmitter Harmonic Power Output (http://tighar.org/Projects/Earhart/Archives/Research/ResearchPapers/Brandenburg/WE-13.htm) paper.

Title: Re: Western Electric 13C transmitter harmonics
Post by: Chuck Varney on October 29, 2011, 09:55:07 AM

The referring post for this one is here (http://tighar.org/smf/index.php/topic,473.msg6343.html#msg6343).

Comments to your   WE-13C Transmitter Harmonic Power Output (http://tighar.org/Projects/Earhart/Archives/Research/ResearchPapers/Brandenburg/WE-13.htm) paper follow.

1.) From the Background section, relating to the 13C transmitter class of operation

The final power amplifier of the WE-13C transmitter was a Class-C amplifier using 2 Western Electric type 282A tetrode (screen grid) vacuum tubes in parallel.

Comment: I’m puzzled by your Class C assessment. For Class C operation the control grid bias voltage must be more negative than the plate current cutoff voltage. The table on p150 of Morgan's Aircraft Radio and Electrical Equipment gives the approximate control grid bias voltage for a properly adjusted 13C transmitter as  -27 V.  The schematic shows that this was a fixed bias developed by the voltage drop across 65-ohm resistor R7, and that there was no self bias provision to make for a more negative bias.

The grid voltage-to-plate current transfer characteristics for the 282-A tube at a plate voltage of 1000 V show that the plate current cutoff voltage ranged from approximately -60 V to -120 V for screen grid voltages between 100 and 200 volts. In each case the -27 V bias voltage is more positive than the cutoff voltage, giving a class of operation lying between B and A (Class AB)--not Class C.

One may also judge the 13C class of operation without resorting to numbers. The 13C schematic shows that the audio amplifier modulated the screen grid voltage of both the 1st and 2nd RF amplifiers (the 2nd RF amplifier being your final power amplifier). For a voice transmission, the 2nd RF amplifier was required to amplify a modulated control grid voltage provided to it by the 1st RF amplifier. To preserve the modulation, the 2nd RF amplifier had to operate as a linear amplifier, and that required a plate current conduction angle not less than 180 degrees; that is, not Class C, and not the 140 degrees you used in your paper.

2.) From the Background section, relating to harmonic attenuation in the 13C output circuit

The WE-13C schematic diagram (Morgan, 1941) shows that the transmitter did not have harmonic suppression circuitry at its output. Hence harmonic current components present in the output would flow in the antenna, causing radiation at harmonics of the fundamental.

Comment: The quoted sentences suggest that the 13C output circuit design provided no attenuation to harmonic radiation.

The 13C 2nd RF amplifier output circuit, like the output circuits of all tuned RF power amplifiers, performed several functions, one of which was attenuation of harmonic radiation. An analysis of the 13C output circuit will reveal how effective it was in that regard.

3.) From the Methodology section, relating to method

Comment: Rather than basing the analysis on an arbitrary space current pulse shape unrelated to the 13C operation, an appropriate method would perform a Fourier analysis on a plate current pulse shape derived from the 282-A tube characteristics and the 13C operating conditions.

Suggestion: Set aside your Terman and Roake and Terman and Ferns papers. Find the plate current pulse shape, sample it over one cycle, and run a Discrete Fourier Transform on the result.

4.) From the Methodolgy section, relating to calculation of power

Since the output power corresponding to a given a-c current component is directly proportional to the amplitude of the current component, the ratio of the output power at a given harmonic to the output power at the fundamental is given by In/I1 = (In/Im) / (I1/Im).

Comment: Power is proportional to the square of current, not directly proportional to it.  As a result, what you have called power ratios are not actually power ratios.

The output power at a given harmonic is the product of its power ratio and the WE-13C output power at the fundamental, which was 50 watts (Morgan, 1941).

Comment:  Assume for the moment that your harmonic current ratios had actually been harmonic power ratios. Multiplying all of them by 50 W presumes that the resistive portion of the amplifier's load impedance is constant for the fundamental current and all its harmonics. This is equivalent to saying that the 13C's 2nd RF amplifier output circuit provided no harmonic attenuation--and that was not the case.

The net effect of all this is that your paper reports a greatly exaggerated estimate of radiated harmonic power. As the estimates are carried forward to your 2006 Harmony and Power (http://tighar.org/Projects/Earhart/Research/ResearchPapers/Brandenburg/HarmonyandPower.pdf) paper, all comments apply to it as well.