Interesting feedback circuit

[KE6DF]

I was reading the section on modulation in Terman's Radio Engineering book (1937 edition) last night and noticed a couple interesting points in a broadcast AM transmitter he discussed.

The transmitter was a 50KW rig.

It used low level class A modulation of a 50W stage followed by three stages of linear amplification to get to the 50KW output level.

A couple interesting poiints:

The true Heising class A modulator tube was connected to the B+ through a choke as usual. But the modulated RF amplifier was connected to B+ through a separate choke.

A blocking capacitor transmitted the audio off the plate of the class A modulator to the RF plate circuit.

Separating the B+ in this way allowed the designer to use higher B+ on the modulator than on the RF stage enabling them to achieve 100% modulation.

Another even more interesting design point (to me anyway) was the way they handled feedback.

They coupled an RF link on the plate tank if the final (50KW) RF stage.

The RF picked up by the link was rectified by a full wave rectifier, the RF filtered out, and the resulting audio was fed back as negative feedback to the first low level audio stage.

So in effect, they had negative feedback around all the audio stages, the modulated RF stage, and the three linear amplifier stages.

I'm wonder if such a scheme would have advantages if applied to a ham tx in the case where you have a low powered AM transmitter (probably with a class B transformer coupled modulator) feeding into a linear to boost the signal.

It doesn't seem any more complex than the resistor ladders you see in 1K level broadcast transmitters, and it might compensate for distortion in the linear amplifier also.

Has anyone tried something like this?

Obviously it would be easier on a HB tx and linear that to retrofit the link to a store bought linear.

[WD5JKO]

David,

   The idea of staggering the voltage to achieve 100% positive modulation is not new with a modified Heising system. We most recently saw this with a post by John, K1DEU with his post near the end of the string on the topic, "Sweep Tube Modulator":

http://amfone.net/Amforum/index.php?topic=19743.0

   Notice he uses a push pull mod transformer, and also notice he does not use the secondary as is customary with plate modulation. Therefore I call this arrangement an interesting variation on the Heising concept.

   Another variation on the Heising concept was done by me with my Gonset G50 where dual parallel 6L6's Heising modulated a single 6146. I also used a push pull transformer in a way that I no longer needed a separate reactor:

http://amfone.net/Amforum/index.php?topic=18961.0

   With the G50 I used a common supply for the RF and Modulator, so in order to achieve 100% modulation, I had to drop the modulated B+ with a resistor (10% or so) and bypass the resistor with a AC bypass capacitor. A lossless way to accomplish the same thing would be to return the modulator cathode to a minus voltage (say -100v) so that the modulator B+ was effectively 100v more than the RF final tube.


   You mentioned the feedback loop mentioned in Terman's 1937 book. I use this on one of my Central electronics 20A QRO projects (the one with a EL-34 RF final tube). This method was tricky in my application, and a NF of any more than 6DB created more problems than I was trying to correct. That said, 3-6 db of gain reduction done this way can overcome a fair amount of non linearity in the system. Keep in mind that the detected RF voltage will vary with the antenna impedance given a constant power level, so you really need both a means to vary the feedback amplitude, and maybe a switch to get the phasing correct. Any distortion introduced in your AM detector is a problem with this idea since the NF sample is not accurately following the RF envelope.

Regards,
Jim
WD5JKO

[W7TFO]

Western Electric used that scheme on many of their broadcast transmitters to keep efficiency up when using that mod scheme.  

They also used that global feedback circuit on all linear and Doherty final models to keep noise and distortion down.  

When they stopped production in the '50's, Continental Electronics took up the line and used it for many more years, most notably for XERA @ 500kW.

http://www.oldradio.com/archives/hardware/WE320A.htm

Today the station runs 100kW, but the treaty allocation still shows 250 kW allowable.  Tough to sell enough time to pay for the power to run it up there.