Jan 10, 2014

Science Fair Experimental LM386 Amp


          You may have seen my "Souped-Up" amp already - my first shot at DIY amp building, cloned from the Smokey amp.  I did mention how these projects are addictive, and now here we are with my science fair project.




          In case you have no idea what you're looking at (it is a pretty weird little amp) it is the son of a Noisy Cricket and a cigar box amp.  Put together one LM386, one 4" speaker, six potentiometers and a wooden craft box and this is what you get.

Lots more after the break.





          What makes this amp interesting is that since I really was just experimenting around with what semi-randomly placed potentiometers could do I found some effects that no person with real knowledge of electronics would have thought to be a good idea  :)


          You'll have to forgive my less-than-perfect schematic; besides the fact that it's hand drawn and not technically standard/correct, I thought I did a pretty smashing job.  If you are good with schematics you'll notice it has a lot in common with the Noisy Cricket amp, shown below.



          One difference is that the Cricket has a buffer and some extra filters and an LED.  Surprisingly it doesn't have any major problems with noise despite lack of shielding and filters.  The main difference that makes my project different are the potentiometers, particularly the added output pot and the pot replacing the resistor in the Zobel network leading from pin 5.

          Maybe you already knew that, but in the event you were wondering what the heck the knobs do, here they are in order from left to right, roughly in order they progress through the circuit:

1     Input pot                        - Volume
2     Tone pot                        - Low pass filter
3     Gain pot                        - Gain (overdrive)
4     Zobel network                - Buzz, clipping, weird cool stuff
5     Grit mod                        - Adds a little noise/gain
6     Output pot                     - Volume (sort of)

Build in progress

     The first pot is 10k ohms and connected directly to the input.  It was the same as the Cricket but I feel it is redundant since it does pretty much exactly what the guitar volume knob does.

     The second pot went through a couple different wiring configurations and eventually ended up as an extremely mild tone changer.  All the way up is normal, all the way down rolls of a tiny bit of highs.

     The gain pot is nothing special and changes the gain - minimum being almost clean, max being overdrive going into distortion territory.  Very practical but not new here.



     The Zobel network mod is where things start getting real.  As I thought up what places I could put pots in the circuit I realized that some interesting stuff could possibly happen if I replaced the 10 ohm resistor normally in the filter with a 1k ohm pot.

      As it turns out, this is true.  With the resistor set to no resistance, the filter does not short circuit as I thought it might.  Instead this just turns off the filter and it has ever so slightly more noise/picks up radio waves etc.  However, turning it just a little so it is above about 10 ohms instantly changes the sound.  The effect is somewhat like the grit mod, but more pronounced.  Long tones start to sizzle and what was strong overdrive pushes over into distortion.  There seems to be a level, almost like it has diodes, where sounds marginally too quiet disappear, making fade outs break up and crackle.

     The burning question here for tone heads is "Does it sound good?".  I am obligated to say that there is no definitive answer - it is all in the ear - but honestly it may or may not have uses.  For low gain stuff it sounds pretty terrible, but if you're playing metal the additional gain and breakup sounds pretty neat.  Not always good per se, but neat :)

     The burning question for techies is "Why does it do this?", and I'll be honest and say that I don't know.  It could be a feedback loop in the ground, or it could compound with the grit network as explained next.  It may not work if you make a similar circuit without the grit network, or connect the leads differently than I did, but if you have any knowledge or theories it would be appreciated.  It is for science after all, and "I have no clue," never sounds good, even if the amp does.


     Next is the grit mod - a loop from the bypass pin to the output that increases noise and gain.  The difference here is that instead of using a switch I used a pot, more for consistencies sake than anything else.  It doesn't matter, the pot acts as a switch anyway - just a little resistance shuts it off.

     I found the grit mod really didn't do much, particularly compared to the Zobel network, but what is really interesting is that the two interacted with each other - activating one turns off the other effect.  Weird, but having no resistance on either pot turns on the grit mod, some resistance on both turns on the Zobel, and any other combination turns off both.  Now I really want answers from you electrical engineers out there :)

In the beginning...

     The final pot I installed was a 10k ohm right before the cap and the speaker.  The placement of it at the end of the circuit has a big difference compared to at the beginning.  Placed at the input, the pot attenuates the signal strength going into the LM386 - in other words, quieter also means less overdrive.  However, at the end, it attenuates signal without changing gain - the desired result.

     Of course there is a problem though; the output pot jumps dramatically between volumes rather than continuously changing.  At maximum it plays fine, but turning it down about 100 ohms or so instantly drops the volume to a whisper level (can play under normal conversation levels).  At first I thought this might have been the result of a bad pot, but now I am pretty sure it is either because 1)  The pot's value is too high       or 2)  The capacitor should go before the pot not after.  Again, wisdom is always appreciated.


     I suppose I should wrap up with some concluding thoughts and comments on the sound.  Overall the sound quality was surprisingly good, relative to the "Souped-Up" amp.  The sound was much bassier and full, which I attribute to both the bigger speaker and the acoustically better enclosure.  As to whether or not I would consider the whole project a success or not... It was very interesting, but being a very unfinished-seeming experimental project, it was not something I planned on keeping long term.  I already dissected the parts and have put them towards two new amps-in-progress (coming soon)  Even so, I learned a lot from this and hopefully you have too.  If you haven't please share your knowledge with me, I'll put it to good use ;)

No comments :

Post a Comment

Questions, comments, concerns, complaints? Leave your thoughts below!