Simple Audio Amp, Check my schematic for me?

I'm fiddling with a simple "laser show" type audio visualization, like what's described here. By attaching a mirror to the focusing assembly from a DVD player and connecting the coils to an audio source, you can get a pretty nice effect when you reflect a laser off the mirror. I'm not worried about damaging the coils by overdriving (I probably will), but I'd like to make sure I'm using the LM386 correctly and that my circuit won't damage other input/output devices. The schematic has 3 "sections:"

1. Input summing and pass-through. The idea is the input is connected to J1 and the speakers to J2, so this device is placed between the source and the speakers. The input is summed to a mono signal for amplification. I'm not too sure about this summing, but from my reading it seems correct. I'm shooting for minimal effect on the second output device (speakers) so the resistors are scaled up by 10 from examples I found online to improve channel separation. I'm not driving any long lines and don't care much about the quality of the audio driving the coils.
2. Optional low-pass filter. It might be cool to have the laser react only to the bass notes, so the input can be passed through a simple RC low-pass before amplification.
3. Amplification. I've never used the LM386 before, so this is what I'm least sure of. Two main questions:
   
   • If I want a gain of 20, are R5 and C2 necessary? It's my understanding that without them the "default" gain is 20.
   • I'm not clear on what pin 7 (bypass) is for. The capacitor is included to prevent oscillations of some kind?

The "AUD_AMP" signal is what will finally be connected to the coils, in case that's not clear.

Schematic:
http://i.imgur.com/XtozU.png

Updated schematic:
http://i.imgur.com/BJEGm.png

This post has been edited by crimsondynamo329 on January 02, 2013 06:28 am

for the amplifier, just use the datasheet defult config.
The summing networks resistor may be.. 10K insted of 4.75K, the 200 K one prehaps 22K.

Prehaps a diode on the output will give it some threshold to make the laser blinking, not sure on thatone.

Summing section seems okay.

R5 and C2 are not needed for the default gain.
C5 is a bypass cap,for the input amp section internal to the LM386. It would be a good idea to use it. 10-22uF should be fine.
You might consider adding the R/C network (10ohm/0.1uf) on the output also.

This is all in the LM386 datasheet:
http://www.ti.com/lit/ds/symlink/lm386.pdf

That low pass filter... too small R, too large C

Thanks for the replies, exactly the type of tips I was hoping for.

I got those values from an online calculator. I put in 120Hz target cutoff and that's what it gave me . Why are these values bad, and what would be a good alternative? Just scaling each up/down by 10 for 1.3k and 1uF?

For practical reasons, I'd definately go up by a factor of 10 to start... 1.3K + 1uF... and use a film cap

Your "Q" will still be low, but should work better.

Considering R1-R2 are already providing a few kohms of series resistance, I don't think R7 matters. Nix it, and go with 0.1uF or so.

An RC filter is a very gradual thing, so don't worry about "Q" or cutoff frequency. It'll still let through plenty of recognizable melody, if you were to listen do it. This would be manifest in the display as very short loops and jags and squiggles. The mirror itself acts as a filter and may dampen much of this already.

Tim

Hmmm... I wonder how pieces of shiny foil glued directly to a small speaker would work. Or perhaps glued to a flexible strip in front of a speaker (or just use a strip of reflective mylar).

Meh, just a thought.

@Sch3mat1c: My thinking on including R7 is to make the low-pass optional, i.e. if R7 were removed then both options on S1 would be identical and the amp input would always be filtered. Including R7 will help keep the AUD_IN line separated from the filtered line, right? If the gain just needs to be turned up a bit when that option is selected, it's not really a problem.

@Mac: The idea actually comes from my car, where the headlights reflected in the rearview "bounce" in nice patters when the subwoofers in the back kick in. However, I would like this device to be silent so I had to look for some other way of getting those kinds of vibrations on a mirror. A mirror glued to a balloon stretched across a speaker (or the vent for a sub enclosure) works well, but makes distorted sound that takes away from the effect. This DVD focuser setup is (very nearly) silent and gets about the same effect.

PS here's an updated schematic:
http://i.imgur.com/DNGex.png

Yes, the extra series resistance helps isolate filtered from normal, but to do a reasonable job it must be many times the source impedance, i.e., over 20k easily. So now your impedance goes up even more. Simple solution, you only need an SPST to switch the capacitor in or out.

Tim

I can see where it would, and that's why I mentioned a strip rather than a complete covering, and something light (foil or mylar) rather than a mirror.

Thanks for the reply btw, here's hoping yours works great.

Funny how I miss the obvious solutions because I'm overthinking... So just to make sure I'm clear: with the updated schematic below, with the switch closed R1/R2 will isolate the filtered line from the input and with the switch open there's no filter to worry about. Makes sense to me.

http://i.imgur.com/BJEGm.png

Yup

Welcome to the "90% of us" club

It's the curse of the geek