Playing with phase in AM signal generation

I needed to test a UHF linear amplifier’s ability to deliver a low distortion AM signal output. Unfortunately, the output distortion requirement was better than the distortion limits of any commercial AM signal source I could find. It was necessary to improvise an AM signal source suitable for the purpose.

I needed a source of low distortion AM to feed the amplifier’s input but every signal generator I could find (Hewlett-Packard, Fluke, Anritsu, etc.) had AM output distortion specifications that were worse than the distortion requirement that the amplifier itself was supposed to meet. To make a source whose AM distortion was low enough for test purposes, I had to gin something up along the following lines shown in Figure 1.

Figure 1 Block diagram for an improvised AM signal source. Source: John Dunn

The plan was to take a double balanced, ring diode mixer and inject a carrier and a modulating audio signal into that to create a double-sideband (DSB) mixer output signal with a suppressed carrier. I needed to follow that with a purchased commercial amplifier to make up for mixer loss. I then added the DSB signal to the original carrier to create the required AM signal having the carrier and the two AM sidebands.

This worked just fine.

To illustrate the signal processing principles however, a SPICE model can be used where the double balanced mixer is represented by a four-quadrant multiplier (Figure 2 and Figure 3).

Figure 2 SPICE simulation of the designed AM signal source delivering 50% modulation. Source: John Dunn

Figure 3 SPICE simulation of the designed AM signal source delivering 100% modulation. Source: John Dunn

One thing to keep in mind is that you want zero phase shift between the original carrier signal and the reinjected carrier signal at the summing point. With no phase shift at all, we can expand the time scale (shown in Figure 4) and look at the nicely even timing of the zero crossings of the AM waveform which is what we want.

Figure 4 Even zero crossing timing of the AM waveform with no phase shift. Source: John Dunn

However, as shown in Figure 5, If the reinserted carrier is phase shifted with respect to the original carrier, we get some unwanted effects.

Figure 5 Zero crossing timing with 90° phase shift with respect to the original carrier showing unwanted effects. Source: John Dunn

The phase shift of 90° in Figure 5 is of course quite extreme, but for demonstration purposes, going to that extreme makes it easy to see how badness arises.

Figure 5 shows that the zero crossings are no longer timed the same which means that there is now unwanted phase modulation (shown in Figure 6) of the signal output and degradation of the amplitude modulation effect we’d been seeking.

Figure 6 Unwanted envelope degradation of the signal output with 90° phase shift. Source: John Dunn

When we return to our previous time scale, we also see with the 90° phase shift that the AM envelope has been degraded as well. Putting our phase angle back to zero degrees again, we see very nice AM signal generation all the way up to 100% modulation (Figure 7).

Figure 7 AM signal source delivering 100% modulation with 0° phase shift. Source: John Dunn

Keep the phase, baby.

John Dunn is an electronics consultant, and a graduate of The Polytechnic Institute of Brooklyn (BSEE) and of New York University (MSEE).

Related Content

Modulation basics, part 1: Amplitude and frequency modulation
Quadrature modulation: The signal behind digital communications
The device model matches the published data—that ain’t necessarily so
The y-, z-, h-, and s-parameter alphabet soup
<!–

–><!–

div-gpt-ad-inread

–>

The post Playing with phase in AM signal generation appeared first on EDN.