Re: I7SWX PTT Pop Mute

Jerry Gaffke

Lots of hits on the web for "LM386 pin 7 mute".  Most pull pin 7 low, though pulling high works too.  Allard's now routing PTT through the Raduino and driving a transistor to turn on the relays.  So we could have a second 2n3904 with base and emitter in parallel with Allards 2n3904 transistor,  the collector of this new 2n3904 pulls down on LM386 pin 7.  Add a cap from pin 7 to ground, discharges almost instantly but would take a few miliseconds to charge back up through the LM386's internal 15k resistor when PTT is released, long enough to keep the LM386 quiet till after this exit noise is gone.  So a 2n3904 plus a cap replaces Gian's 2n3906+resistors+diode plus Joe's relay, assuming you are going to implement Allard's "TX-RX Line Wiring" mod anyway.  Beware, I have not tried this.

Pin 7 is meant to be bypassed to ground through an optional 10uF cap, preventing power supply noise from disturbing the diff-amp input..  But the Bitx40 has some serious external 12v filtering going into the LM386, so no large cap on pin 7 is probably fine.

Not being an analog kind of guy, fully figuring out the LM386 from the datasheet's internal schematic is quite the puzzle.  I think one key is to assume the IN- pin is grounded (as it is in the Bitx40), so the collector of the NPN at the top of that side of the diff-amp is always two Vbe above ground, or about 1.2v.  The current mirror below the diff amp draws an equal amount of current from the two sides of the diff amp, any excess current on the right side goes off to drive the output stage.   This guy:  says the gain of the LM386 is equal to  2*Z(1-5)/(150+Z(1-8)),  where Z(1-5) is the impedance between pin 1 and pin 5 (typically just the 15k internal resistor) and Z(1-8) is the impedance between 1 and 8 (typically just the 1.35k internal resistor, unless you have an external cap across those pins to increase the gain as the Bitx40 does).   The 150 comes from the internal resistor between pin 8 and the top of the left side of the diff-amp, the factor of two comes in because of the diff-amp.  That all sort of makes sense, though I'm not convinced my understanding is complete.  When muting, I assume bringing pin 7 to gnd or vcc upsets the diff-amp to where it's no longer functioning as a diff amp.  Could be that going in and out of mute like this creates its own pops, I'll have to try driving pin 7 both high and low, see if either works better.

This guy does a dc analysis without just too much hand waving.
He's trying to get the LM386 behave like an op-amp, no internal feedback, so he can add an output stage to the LM386 and have the feedback loop go around everything.   He says that at 125mw driving headphones, the LM386 is clean enough.  Ask for more power from it to drive a speaker, and it distorts badly.  The uBitx has moved to a better amp.

Here's an interesting webpage on cheap and simple and easy to understand alternatives to the LM386.  He gives up on the LM386, drives an output stage with an op-amp with good results, written long before the above article:

I like how those two articles show the process of figuring these things out.  Both seem quite competent, and they're not afraid to describe stuff they try that didn't quite work.  Even the simplest circuits can keep a good engineer scratching his head for days.

Kind of amazing the LM386 is still such a popular chip at a hoary 34 years of age.  Hobbyist interest on the web seems equally divided between qrp rigs and guitar amps.  In 20 years the kids will be messing with their DSP algorithms to get a guitar amp that sounds grungy like those badly overloaded LM386's, having no clue how an overloaded vacuum tube amp would sound. 

Jerry, KE7ER 

On Sun, Jul 30, 2017 at 05:45 pm, Joe wrote:
I also installed the LM386 PIN 7 mod and found it works very well,  the best noise  reduction so far.
I still found the TX exit noise a bit louder ,so I added a relay to open the speaker, this further reduced the noise when exiting Tx.

Join to automatically receive all group messages.