Hi, "John"!

I looked up an LM386 data sheet from National.
I also consulted your advice.

It turns out that the best solution is to
strike C18 and R12 completely, and to change
C20 to 10 microfarad.

Looking at the original schematic, it makes a
lot of sense... if you're trying to amplify
100 millivolt p-p of audio into 1 V (1000 millivolt)
p-p of audio, we clearly only need a factor
of 10 amplification, a.k.a. 20 decibel.

This is best accomplished simply by leaving
the connection between pins 1 and 8 open,
which by the way programs the LM386 for
its default 20 decibel amplification.

I found that R12 between pin 5 and pin 7 simply
provided a greater chance of a ground loop, so
eliminating that provided less chance of
motorboating even at 20 decibel amplification.

Then I took your advice to simply bypass
pin 7 to ground with a 10 microfarad capacitor.
This was easily enough done by changing
C10 from .1 microfarad to 10 microfarad.

With 100 millivolt p-p coming into the phase
splitter, the scope reads a clipped sine wave
at several volts... meaning that I have plenty
of gain to spare, I should even crank down the
input a little.

Because of the LM386 is amplifying a differential
signal, our effective amplification is actually
a factor of 20... so it would seem that nominal
input signal is actually 50 millivolt.

So, I believe that I have a working version of
that section of the circuit. I think I will
work on the detector portion (Q3) next, and
try injecting a 20 millivolt rf carrier at
14 Mhz.

I'll be back in a couple days with the results.

The Eternal Squire


John Moriarity wrote:
Dear Mr. Moriarity,


"Mister Moriarity"??? ;-) How about "John"?


R12 seems to be misspecified as well. On the
pdf schematic, R12 shows as "10W". I assumed
at the time that the W was simply a misconverted
omega symbol, so I had soldered in a 10 Ohm
resistor there.


Ten ohms is the "right" value, looking at a LM386
data sheet.


U1 does not amplify at all!!!


The LM386 should have plenty of gain if wired
correctly with these values. Something is very
wrong!

I am also suspicious of the connection of pin 7
on the LM386. I would disconnect it from the
junction of R12 and Cwhateveritis, and separately
bypass it to ground with a 10 microfarad cap
(plus side to pin 7). You might even get away
with not connecting pin 7 at all!

73, John - K6QQ

With a 10K resistor as R12, I got a VERY nice
volume that filled the entire room. The only bad
news is that when I remove the input signal, I get
about 400 hz motorboating.

Squire-

I believe R12 should be ten ohms. It provides a load at higher frequencies
where the speaker or headphone impedance might be too high for stability.

The motorboating is most likely feedback via the power supply line. Therefore,
the ten ohm series resistor should be higher (perhaps 100 ohms?), and the
filter capacitor at the IC's power terminal could be increased.

73, Fred, K4DII

Fred,

Maybe R12 needs to shunt the speaker to ground, rather than the
output of U1. If R12 is needed just for AC response, then it doesn't
make sense for it to conduct the DC between pins 5 and 7. By
shunting the speaker it should force the speaker impedance closer
to 10 ohms. I'll try that.

The Eternal Squire


Fred McKenzie wrote:
With a 10K resistor as R12, I got a VERY nice
volume that filled the entire room. The only bad
news is that when I remove the input signal, I get
about 400 hz motorboating.

Squire-

I believe R12 should be ten ohms. It provides a load at higher frequencies
where the speaker or headphone impedance might be too high for stability.

The motorboating is most likely feedback via the power supply line. Therefore,
the ten ohm series resistor should be higher (perhaps 100 ohms?), and the
filter capacitor at the IC's power terminal could be increased.

73, Fred, K4DII