The Sound Story

The program Motion does not include the ability to use sound, though it is something that some of the members of the mailing list have talked about.  I had been thinking about this for a while and started researching the possibilities.  I would have to attack this problem on two fronts.  One, build a stereo preamplifier to allow dual microphones to feed the sound card and two, find a way to make Motion capture sound.

Little did I know that finding a preamp circuit would take months of testing circuits on a breadboard before I found a circuit that would work well with my Aureal 8820 sound card on my aging Linux box.   I built prototype preamps with single transistors. I built preamps with a couple ICs and a few transistors. I think I have looked at almost every microphone preamp on the web and I have built many of them on a protoboard.  With all of the preamps I built,  I found myself going back to a simple circuit that was adapted from the data sheet for the LM386N audio amplifier integrated circuit. This circuit stayed built on a protoboard as a mono preamp so that I would have sound on the mpegs I collected each day.  In between builds of other preamps I messed around with this one trying to get rid of the noise in the circuit.  In the end, I got rid of most of the offending noise and even though it was by no means the most powerful of the preamps I built, it was enough to get the job done.  I also came up with a simple way to get rid of the nine volt battery that was used in these circuits.

I came up with the power circuit idea after being frustrated by the fact that a 9 volt battery would only last a couple of days running 24/7 and running an auxiliary power supply introduced a lot of hum into the sound card.  I was thinking about the problem one evening and thought that using the power supply for the computer that the sound card was in would mean that the ground circuit was common to both the amp and the sound card.  Since grounding issues are what usually introduce hum into an audio circuit, this just might do the trick.  I didn't have a 9 volt voltage regulator on hand, but did have a LM317 adjustable voltage regulator.  I built a circuit using the schematic on the data sheet and was pleasantly surprised when it turned out to be as quiet as using a 9 volt battery.  Pretty cool!  One major problem solved.

I went through a couple different ideas on microphones.  I started with a couple condenser mics from Radio Shack.  I mounted the mic capsule inside a metal tube for the first try.

This ended up sounding - well, like a microphone in a tube!  Who would have thought?  I then tried a different approach.  I used the reflectors from a couple cheap flashlights and put the microphone button in the area that would have housed the light bulb.  I cut out a mounting plate from some scrap fiberglass and using an old radio controlled car builder's trick, I dyed the fiberglass with clothing dye.  Not only does it look better, it works better as well.

This mount will attach behind the camera with the microphones below the camera housing.  The mics are angled so that they face out from center.  This helps a bit in being able to separate the left channel from the right.  My friend John tells me that the setup looks a bit lake a brassiere, but I figure that's just wishful thinking on his part!  The mics are rubber mounted in the fiberglass and this does a bit to isolate the mics from the rumble of the stepper motors.  It also makes it possible to angle the mics about 15°  from perpendicular to the mounting plate.

With the mics in a better state, it was back to the preamp circuits.  I came across a kit at Jameco called the Super Snooper and had high hopes for it, as it was designed to listen to voices at a distance.  Since my goal is to be able to incorporate the ability to point the camera in the direction of sounds coming from my property, I was hoping that this circuitry would be sensitive enough to serve my purpose.  The PDF file has a decent shot of the circuit board, so it was pretty easy to duplicate the circuit.  This circuit uses a LM386 audio amp and a LM1458 OP-Amp along with a condenser mic.  Unfortunately, it didn't produce a whole lot more sensitivity than using the LM386 alone.  I was a bit surprised that the Jameco kit did not make use of jumping pin 1 and 8 of the audio amp to increase its output - though the circuit board does have traces provided to add a 10µF cap to increase the voltage gain to 200. 

So it was back to the drawing board, or in this case Eagle software's  Eagle 4.14 light to draw up another circuit.  At the same time, I started doing some reading on how to etch circuit boards.  There are a couple types of kits available for etching circuits on copper clad circuit boards.  Radio Shack sells one with a Sharpie permanent marker so you can draw your layout, then etch away the un-needed copper with the supplied Ferric Chloride solution.  This may work well for very simple circuits, but the pen doesn't draw a sharp enough line for circuits where the traces will be very close together.

Another choice is to use one of the kits from Jameco or Allied Electronics that use a photo-sensitive coating on their copper clad blank boards.  For this type, you design your circuit using a tool such as Eagle 4.14 lite, then print it to a transparency, then place the transparency over the board and expose it with either sunlight or a UV lamp, then develop the board and finally etch it with either Ferric Chloride or Ammonium Persulfate.  This supposedly gives much better results, but I had read about using circuits printed on special paper with a laser printer, then ironed on to the circuit board blank.  This works because the toner used by the laser printers is a powered plastic.  When heated, the powder melts and forms a thin plastic film over the paper in the shape of each character.  As long as you use a paper with a slippery finish, such as photo paper, applying an iron to the back side of the paper will transfer the plastic image to the circuit board blank. This sounded like it would be the easiest way to go.

I purchased the Radio Shack etching kit.  Once I opened the package and looked at the board blanks, I noticed that instead of being totally smooth, they had a texture to them.  I though that this might prevent the process of ironing the printed circuit to the board from working properly.  The "fix" was to use some 400 grit wet and dry sandpaper, then some 800 grit, to sand the board as flat as I could get it.  This took about 15 minutes of sanding with water to wash away the copper particles to get the 3.5 X 5" board nice and flat.  The sanding is also a good idea to clean the oxidation from the board before trying to etch it.  Below are the before sanding and after sanding shots.  It was tough getting the bumpy surface to reflect with a flash, so I tried without one.  If you enlarge the pictures, you will get some idea of the difference.

I am no electronic designer, but through all of the reading about audio circuits I have been doing for the last couple of months, I am beginning to come up with some designs that  do what I need.  The design I just finished drawing up puts the capacitors and the volume control pots very close to the IC chip to keep down the noise a bit.  I am getting a little better with positioning components on the board to keep the layout compact, but I still have a way to go in that regard.  I enjoy working with circuits and think that I will spend some more time learning when I finish up with this project.  I have read a bit about PIC microprocessors and think I would like to spend some time getting familiar with them.  Ah, so many interests and so little time!

The next circuit I decided to build was fairly simple compared to some of the earlier attempts, but it was a design that had worked well at the breadboard stage, unlike the "Super Snooper circuit that I borrowed from Jameco and the two transistor amp circuit I found on an audio site.  Below is the Super Snooper circuit on a breadboard on the left, a quick and dirty perf board two transistor circuit and the final board design for the stereo LM386 board on the right.

Back to the etching of the circuit board.  After getting the board flat, I cleaned it with steel wool and an automotive prep solvent which is used to prepare body work for paint.  It strips off waxes and whatever else without being too harsh.  I had placed some drilling guides on the board and drilled four holes to allow me to line up the top and bottom circuit templates.  This is a good idea if you want the top and bottom circuits to line up with each other.  I used a size #62 drill bit which is .038" (.9652 mm).  Be forewarned, they are quite easy to break.

My wife took the artwork to work along with a couple of sheets of Epson All-Purpose Glossy Paper, # S041654, and used their laser printer to print my circuits to the shiny side.  I then punched 4 holes in the bottom template, which didn't need to be reversed for the printing, and in the top template, which did need to be flipped over by Eagle 4.14 to print properly.   Starting with the bottom, I lined it up using toothpicks and pressed the iron in the center of the paper to get the toner to stick to the copper.  Once the center is stuck down, removal of the toothpicks allowed me to finish ironing the circuit pattern to the copper.  I used just below linen on the iron's heat scale and pressed down hard with the tip of the iron.  It is possible to gently pull back the paper to see if the toner is going on completely or whether there are still bits attached to the paper and then return it to its aligned place, but it is very hot.  Once the transfer has been achieved, you can move to the top circuit. Do NOT remove the paper from the bottom while ironing the top or the toner will attach itself to whatever is under the board.  The paper has a wax-like coating on it, so the paper doesn't stick to the board.  Unlike some of the instructions I read on the web, I didn't need to soak the paper in water to get it off of the copper.

Once the ironing is complete, let the board cool and remove the paper.  Clean up your mess (It's probably not YOUR iron, right?) and then touch up any areas where the toner did not adhere with a Sharpie or other permanent marker with a fine tip. 

I then tied some string through one of the holes I had drilled and stuck the board into the ferric chloride that had been warmed up to about 100°F by sticking the bottle under some hot tap water.  I continuously agitated the solution and checked the progress every few minutes.  Even though this was the second time I had used the solution, the etching went pretty quickly.  I would guess that it took about 10 to 12 minutes for the copper to be totally dissolved.

After cleaning the board under running water and some dish soap, I cleaned the toner and Sharpie ink from the traces with some lacquer thinner.  The Radio Shack kit comes with some solvent to clean off the ink, but I already had the lacquer thinner out and it worked so well, I didn't bother trying their solvent.  I'm pretty impressed with using photo paper and an iron to transfer the toner to the blank circuit board.  It did such a good job that the printing that I placed on the front of the board was sharp and clear.  That's pretty neat considering that the letters were only 1/10th of an inch tall.  For my first attempt at using this method, I consider it a complete success and can't even think of anything I could try the next time to make it better - except for planning better circuits.  It turned out that my plan to solder both volume control pots directly to the board wouldn't work.  I had miss measured the depth of the case by an inch.  I have no idea how I did that, but the board would not fit if I soldered the pots to the bottom of the board as planned.  I ended up having to mount the board to the bottom of the case and attach the pots to the face of the case and run short lengths of shielded wire between them.  It's not quite as clean looking as I had hoped - even though no one but me will ever look inside the case. 

Stuffing the board with parts went quickly and the soldering was easy, thanks to finally replacing the worn out tip of my soldering station.  My concerns that having to move the two pots from being soldered directly to the board to being attached by a couple inches of shielded cable would introduce some noise into the circuit were unfounded.  The circuit is remarkably quiet.  The sensitivity and clarity of the sounds it picks up are pretty good.  With the pair of mics mounted on the camera outside the computer room window and that window shut, I can hear the radio being played at moderate volume inside the room.  I asked my wife and son to have a conversation on the driveway and had no problem listening in, though that's only about 40 feet away from the microphones.  The separation between the channels is good.  It is easy to tell what direction cars are moving when they pass the house.

Even though this phase of the project, or should I say the project within a project, has taken a lot longer than I had originally envisioned it would, I'm quite pleased with the outcome.  Besides accomplishing the task of building a microphone preamplifier, I've learned quite a bit about amplifier circuits, become pretty proficient working in Eagle 4.14 lite, and learned a new (for me) method for producing a pretty good looking circuit board. 

Below, left to right: 1) You can see how well the top and bottom layers of the circuit board line up. 2 & 3) The preamp takes its place in the little 1U rack case, along with the BiStep A06 stepper motor controller and a multiple output voltage regulator.  There's still a lot of room in this case.  I just have to figure out what else needs to go in there.

© Fager 7-31-05