In the world of electronics, there are few devices that hang around for a decade, much less a generation. That microprocessor in your brand new whizbang computer was obsolete the day it left the factory, no mind about next year. Your new laser printer? Try and buy the same model next Memorial Day. Television sets? Fagedddaboudit. If you don’t have 3D with quadraphonic sound, you might as well have a crystal set.
But every now and again comes along a device that isn’t sexy, doesn’t have a dozen bells and whistles, but is the perfect combination of simple, cheap and easy to use. In the world of aviation do the DC-3 and the C-172 come easily to mind?
In the 1920s through 1940s, a vacuum tube called a #45 was the audio workhorse. Nothing special, just reliable and inexpensive. Some estimates place the number of these devices manufactured at four million-this at a time when a days wages were measured in pennies.
Along comes WW-II, and we needed something just a bit better than the #45, so the 6V6 vacuum tube arrived. A refinement of earlier devices, the 6V6 held sway until the swan song of the vacuum tube sometime in the early 1970s. Even today there are overseas factories turning out the 6V6 by the (literally) millions. The folks who are into music reproduction (sometimes called audiophools when they buy electroless copper gold-plated speaker wire) say that the 6V6 provides the nearest thing to a perfect audio amplifier that you can make. While I have a tin ear from too much gasoline being converted to noise in my near vicinity, back in the day I could certainly tell the difference between a pair of 6V6 output devices in an ultralinear design and the same music being pumped into the same speakers by a pair of 50C5 tubes. But that is ancient history. Except for the aforementioned music types, everything we do today is done in silicon without vacuum tubes-solid state. Transistors. Integrated circuit chips. Stuff that doesn’t need to get hot to work.
Several audio power amplifier devices have been around for a while, but none are still on the market other than the faithful LM386. I have schematics going back to 1976 that show the 386 being used, and by my math that means the little rascal is at least 34 years old; prototypes and sample quantities were probably at least two years older than that. My 1972 books make no mention of it, so I think we can give the 386 the benefit of the doubt as being 35 years old. That compares favorably with the venerable #45 vacuum tube that lasted 25 years and the 6V6 that lasted as a popular everyday device for about 35 years. And the 386 shows no signs of going away anytime soon.
So what does this little thing do and how can we use it to make our aviation life a little easier? Quite simply, it is an audio power amplifier. It will drive both speakers and headphones easily. It will work with an aviation 12-volt system as well as the ubiquitous 9-volt transistor radio battery. It economically sips current like nothing else in the inventory. It will boost the input voltage somewhere between 20x and 200x depending on how you do the circuit. It can be used as a straight amplifier, a summing amplifier or a power oscillator. It is stable, less than half a buck (Mouser, 42 in onesie-twosies) and reliable as a rock. The -4 version has an absolute maximum rating of 22 volts; the other versions have an absolute maximum rating of 15 volts, which is flirting with my recipe for disaster when run directly from an aircraft supply that can get up to 14.5 volts on occasion. I use the -4 version exclusively.
How do we use these little things? Ill let your imagination take wing. What you use for audio input and what sort of noisemaker (speaker, headphones, piezoelectric transducer, etc.) you use for output is up to you. All Im going to do is show you three configurations of amplifier. Wind it up and let it fly.
Figure 1 shows the basic amplifier. Give it a power-supply voltage of between 5 and 15 volts, an audio signal at audio input and you will get noise in the speaker. As mentioned, the speaker can be anything from a real speaker to a set of earphone elements to another sort of audio transducer. How do you calculate the values of the components? You first have to select what low frequency you want to start to cut off. For communications, we generally say this is 300 Hz. For hi-fi this is 20 Hz. The input capacitor (10n) shown as 10 nanofarads is calculated as follows. If the lowest frequency you want to hear is f and the input impedance (inside the chip, you cant change it) is 50 (50 kilohms), then the capacitor c in nanofarads is: c = 3000/f. Please don’t ask why unless you want another four pages. The output capacitor? Same drill, different numbers, but with c in microfarads instead and r the value of the speaker: c = 240,000/(r x f).
How much output can you expect? You have a power supply of v volts. You have a load of r ohms. The maximum you can get out of this amplifier is going to be: w = (v/2.8)2/r, where w is the output power in watts. For example, a 12-volt power supply with a 4-ohm speaker could have a maximum theoretical undistorted output power of 4.5 watts before clipping and distortion set in. With this little amplifier, that wont happen.
The maximum power available before the amplifier runs out of steam will be about a watt. That isn’t enough for you? I used this little amplifier at Oshkosh a few years ago, and the back rows complained that it was too loud. We are too well conditioned by the more is better crowd selling watts by the hundredbuck to understand that a watt is a lot of audio power into a decent speaker.
The 10-ohm resistor in series with the 100 nanofarad capacitor on the output? Its a simple way of ensuring stability and prevention of spurious oscillations that muddy the sound.
If a gain of 20 isn’t enough, see Figure 2. A resistor and capacitor in series from pins 1 to 8 (observe polarity on the capacitor) let you set the gain anywhere from the intrinsic 20 (very large Rg) to 200 (Rg = short circuit).
If you have a lot of noise on your power-supply bus (say, an unfiltered 12-volt aircraft supply), see Figure 3. A single capacitor from pin 7 to ground gets rid of nearly all noise.
Summing a lot of audio inputs into a single input is a piece of cake. Modify the input circuit (Figure 4) to give yourself a summing junction, which is nothing more than a relatively low-value resistor (1.0k) fed with a relatively high value resistor (22k). This cuts the gain of the circuit from output to input so that the gain is reduced from 20 to 1 (a volt at any one of the inputs gets you a volt at the output). If you need that gain back, see Figure 2. I show three inputs, but you could have 300 if you wish with no changes.
Tricks and Treats
There’s a trick that Ive exploited a few times. If you need the difference of a couple of audio inputs, take a look at Figure 5. Ive put polarity signs on the two inputs. Pin 2 is what is called an inverting input and pin 3 is a non-inverting input. The inputs are absolutely identical with the exception of the fact that a signal at pin 2 appears at the output shifted by 180…what goes UP at the input goes DOWN at the output exactly mirrored. Conversely, that same signal input to pin 3 has zero phase shift, so what goes UP at the input goes UP at the output. Feed the same signal to both inputs simultaneously, and you get ZERO output.
Heres the deal. You have a signal full of noise. If somehow you can get the noise without the signal and then feed the noise to one input and the signal plus the noise to the other input…poof! The noise vanishes and you are left with a clean signal. Sorcery? Nope, just the magic of noise cancellation. Did I hear somebody say they were going to try this to make a noise-cancelling headset? Go for it. And all of this for half a buck.
Where do we go from here? Dunno. Ive looked back over the last 33 years (!) of my articles and noticed a fundamental gap in my writing. In all that time Ive probably mentioned transistors a thousand times, but I cant find any articles where I tell you what a transistor is or what it does. Let’s spend a couple of months discussing the two main types of transistors, bipolar and field-effect, and how we can use them to our advantage. After that, who knows?