AC Explorations - Need More Current!

published 2025-12-09

by Christopher Howard

Continuing to work through my workbook on AC currents and voltages, I was eager to reproduce some of my book learning with actual components. Series resonant LCR networks are exciting because it is possible to have a voltage across your capacitor or inductor that is larger than the supply current, which works out mathematically so long as the vector sum of all the voltages in the circuit equal the supply voltage. To get interesting voltage levels, though, you'll need not-insignificant levels of current, since E = I × X, the X here being the reactance of the reactive component.

This is where I ran into an obstacle at the workbench. My experiments were in the audio frequency range — the last network I set up being resonant at 5.3 Khz. I needed about 100 mA current, but the signal generator I was using to generate the input signal would start drawing back before that — somewhere around 40 mA I think, or maybe less.

I had the idea of using one of the open amplifiers built into my home-made analog computer. However, the LF365 chips those are based on are limited to 10 mA.

Some kind of more powerful amplifier was needed, but what could I find easily? My first thought was to repurpose some amateur radio amplifier, but surely those would be designed for RF frequencies, yes? Not audio frequencies.

Then I remembered I had one tube of LM386 chips, which is an audio amplifier IC. Some pertinent details on the LM386:

requires 5V to 12V supply
expects a speaker impedance between 4 and 32 ohm
can accept signals up to an amplitude of 400 mV
has a built-in gain of 20x

Based on those data points, you would think that it would be possible to draw a current of up to 2 amps. However, in the data sheet there is a limiting curve of supply voltage vs output voltage, which indicates that with a 4 ohm load, the output voltage plateaus at about 3.7 volts. So, you can't quite make it to one amp. But 900+ milliamps is still a nice step up from 40 or 50!

[Edit: I had misread Figure 3 and I see now that the Y-axis is in units "volts peak-to-peak", meaning that the RMS voltage output is quite a bit lower. With further experiments on the bench, it seems in practice that I'm not able to get an output of more than about 1.6 Vrms before it starts chopping off the peaks of the signal.]

To this end, I built the minimum parts design as indicated in the LM386 datasheet (Rev C), figure 10. However, when I first hooked it up, I was getting terrible HF noise and instability, which had me stumped for several days. Eventually, I was able to resolve the issue by putting one additional 47 uF capacitor from the power supply pin (6) to ground, and then another one from the input pin (3) to ground. I tested it and it is producing clean signals from 1 kHz to 20 kHz.

The signal sometimes becomes distorted if one of the capacitors gets a little loose, so I'm thinking I need to transfer the circuit from the breadboard over to a deadbug construction. Deadbug is a new skill set for me, but I'm getting it figured out.

Copyright

CC BY-SA 4.0 Deed