Improving the fidelity of AM reception audio

[wa2ise]

This is a set of mods to improve the fidelity of AM detectors commonly used in tube AM radios and the AM section of tube tuners.

We split off the AVC circuit from the audio detector. Because the filter cap in the AVC circuit will reverse bias the detector diode some. This will cause the high amplitude negative peaks of the audio (which causes very low RF valleys) to be clipped off. This rarely happened back in the old days, but with today's heavy audio processing by radio stations, it can be a significant source of distortion. Most AM radios use tubes like the 12AV6, which provide a pair of detector diodes. The second is usually not used and is tied to ground in most radios. Here we unground it and use it for the AVC circuit. A 100pF cap coupled to the plate of the last IF tube feeds the new diode. This also gives a benefit of keeping the gain down as we tune across a station, so we don't get blasted as much when slightly off tuned.

A second mod is to reduce the usually 200pF audio RF filter cap. Here we used 50pF. This avoids "tangent" distortion; the clipping of negative going peaks of higher frequency audio. Contrary to popular belief, most AM stations broadcast a full 10KHz of audio, for an RF bandwidth of 20KHz. Sure, station frequencies are spaced every 10KHz, but not in the same town.

Another mod is to reduce the heater voltage applied to the detector diode tube. Here in a series string AA5, we can do this by paralleling a 330 ohm resistor with the 12AV6 heater. This will cause the 12AV6 to see a heater voltage of 10V. This lower heat will produce less "contact potential" seen by the diodes. Less stray electrons boiled off the cathode. Less bias for the RF signal to overcome. Sure this also impacts the triode, but its self bias only drops from 1V to about 0.6V. Which is not serious. If this becomes a problem, one can "borrow" some negative bias off the 12BE6 local oscillator circuit by connecting a set of resistors totaling around 30 megs from grid 1 of the 12BE6 to the grid of the 12AV6. You probably won't need this.

If, in a tuner, the detector is a solid state diode, replace it with a 6AL5 or equivalent sub-mini tube. And use both diodes, one for audio and the other for the AVC. And run the heater on reduced voltage of around 4 to 5 volts.

[asynchronousman]

That's about it, Bob, but this is a trick some of us want to do to our vintage SS receivers, if we are truly lucky enough to have ANY music stations left. Take the instance of my current GM Theftlock AM/FM cassette unit, which has a moderate response (yes, plain ugly FLAT with a little bass but better than some). Assuming I manage to deal with the code, I would love to figure out how to widen it. 1994 Pontiac Grand Prix if I knew the model after driving it only 8 days I'd run for office.

[wa2ise]

Originally Posted by asynchronousman

That's about it, Bob, but this is a trick some of us want to do to our vintage SS receivers, if we are truly lucky enough to have ANY music stations left. I would love to figure out how to widen it.

There's probably a ceramic IF filter buried in there. Most of those have about 10KHz passband, which yields 5KHz of audio at best. I've modified home stereo receivers to widen them from 5KHz to 10KHz. See my page
http://pw2.netcom.com/~wa2ise/radios...tml#solidstate
about halfway down. Only thing is that car radio sets are likely pretty crampt inside so fitting that new filter might be tough. A new wider band ceramic
filter might work (*IF* you can find one that is the same frequency as the old one in your digitally tuned radio, less critical if the tuning is analog) if you can find some. You might be able to find a usable one from a junked FRS radio or
cordless phone.

[asynchronousman]

I will do some research to find which model it is and that might help both of us.

See that 4700 pF cap across the 470K grid resistor? If this was not a module with those resistors and caps, one could in essence remove the 4700 and replace it with a smaller version to get more high end audio through. Unless you feel daring enough to buy the components and remove the module altogether.

Tha 4700 pF cap is there to filter out whines and whistles of RF interference. Experiment with the value until you get the best balance. This is likely true also with the solid state variety.

I thought that the 200 pF was a neutralizer to eliminate oscillation due to phase shift????

The separate diode idea is always an excellent choice.

I have replaced the tube rectifiers with solid state (germanium) diodes and got excellent results in both fidelity and sensitivity. Why do you suggest changing a SS one with a tube?

One can also detune the IF stages to get a wider audio response, since too high a Q narrows bandwidth if you so dare. But first mark the original setting with magic marker, pencil or nail polish, and don't go nuts with the screwdriver or you can break the lugs, and go overboard with turning them and forget where they were originally. One loses some sensitivity and selectivity, but those are the trade offs for sound quality. This is true for both technologies.

BTW, 5KHz is the highest note that is considered musical anyway ("The Science of Music" & "The Psychology of Music". Good reading).

Gabe

[asynchronousman]

Yes, and I could probably synthesize a bland, monotonous 5 kHz tone myself...I'm reminded of the political catch phrase when I point out, "It's the harmonics, stupid!" (No disrespect meant at all)...Just as color needs complementary hues, sound needs harmonics.

Before we get into bandwidth hoohah, here it is...

535 LSB|----10 kHZ----||545 carrier||----10kHz----|555 USB

(an example).

It's been capable of that since at least the 1930's, and in the early 90s, with AM stereo, some were pushing it to around this:

529 LSB|----16 kHz----||545 carrier||----16 kHz----|561 USB

Digital looks something like this (oversimplified):

~530 LSB|10 kHz IBOC-AM|*Analog 545-530||545 carrier||*Analog 545-550|10 kHz IBOC-AM|560 USB

(an image, phase modulated or similar to FM, will appear at every 30 khz or three channels out upper AND lower sidebands, due to the nature of the IBOC signal, wiping out any weaker signals at that frequency).

If you owned a GM car radio from the late 50s or early sixties built with tubes and especially that plate reverb, please do not tell me how low-fi AM is. You would know better.