Updated 12 March 2005

Power Measurements for a Sherwood XA-2105BP Car Power Amplifier

Rev (-) 12 Mar 2005

© 2005 By Joe Roberts

Introduction: This article shows some basic measurements of power capabilities for a small (by today's standards) car power amplifier. The amplifier used was a Sherwood XA-2105BP, it is one I have had hanging around (never used) for about 10 years. It is rated at 45 watts per channel at 4 ohms.

Method: The amplifier was connected up to test equipment on a lab bench. A regulated power supply (variable voltage) capable of about 50 amps was used to power the amplifier. A signal generator was used to apply an identical signal to both input channels. The outputs of the amplifier were connected to 4 ohm 100 watt resistive dummy loads. An oscilloscope was used to monitor the output waveform and a true RMS voltmeter was used to make voltage measurements of both the power supply used to power the amplifier and output channels of the amplifier. The output of the amplifier was monitored for various signal frequencies and input levels.

A photo of the somewhat messy lab bench showing some of the equipment used for these tests.

Results: I measured the power output of this amplifier (both channels driven) for various levels of DC supply voltage to the amplifier. In each case the amplifier was fed a 1000 Hz signal and the amplitude was adjusted so that the output of the amplifier was just at the threshold of clipping. The results are as follows:

What do these results show? Clearly, when the DC supply voltage to the amplifier is higher, the maximum output power increases. Normally, this car amplifier would be installed in a car (not a lab bench). When a car engine is off the battery voltage is closer to 12 VDC; when the engine is on and the alternator is running (charging the battery), the voltage is more likely to be in the 13.8 VDC range. So, your system can play louder when the engine is running. Basically, a 15% increase in DC supply voltage translates to a 25% increase in power.

I also measured (with the DC supply voltage at the standard 13.8 VDC level) the output of this amplifier at some frequency extremes. For each frequency, the input signal was increases until the amplifier just reached the threshold of clipping. The results are as follows:

Somewhat surprisingly, this particular amplifier actually puts out more power at 20 Hz then it does at 1000 Hz (this is not usually the case for most amplifiers).

Oscilloscope Displays for Various Clipping Conditions: Below are some photos of actual oscilloscope displays that show real world examples of clipping. These photos are useful in that they show how clipping behavior of real world amplifiers varies as a function of signal frequency. The photos below represent the clipping behavior of most (audio power) amplifiers available today. Each photo has descriptive notes.

Figure 1. Amplifier output with 1000 Hz signal just at the threshold of clipping. The signal is clean and shows no significant aberrations.

Figure 2. Amplifier output with 1000 Hz signal driven into slight clipping. Note that the waveform peaks have been flattened out. This level of clipping would be clearly audible but not too hideous (but then again few people listen to sine waves).

Figure 3. Amplifier output with 1000 Hz signal driven into significant clipping. The waveform peaks have been flattened out more. This level of clipping with normal music would start to become fairly noticeable (the music would sound compressed).

Figure 4. Amplifier output with 1000 Hz signal driven into very bad clipping. The waveform is starting to more closely resemble a square wave. The sound quality at this point would be objectionable to most any listener. Also, clipping of this magnitude is sending a big invitation to speaker damage!

Figure 5. Amplifier output with 20 Hz signal driven into slight clipping. Note how this plot differs from that of Figure 2. The amount of clipping is similar, but notice the waveform? The clipped portion is slanting downward with time instead of holding fairly flat. This is very typical of clipping behavior for low frequencies. Why? The 20 Hz waveform is so long (time wise) compared to the mid band frequency waveform that the power supply is having a hard time holding up the rail voltage. Basically the rail voltage is being sucked down as the storage capacitors drain their charge (in an attempt to provide the power needed). The very best amplifiers will hold this slant to a minimum. Testing amplifiers at this frequency really puts the power supply to the test!

Figure 6. Amplifier output with 20 Hz signal driven into bad clipping. The waveform clipping characteristic (clip slant) is quite a bit more pronounced now. Clearly the amplifier's power supply is having a harder time with this low (20 Hz) signal as compared to the relatively "easy" 1000 Hz signal used in the first test!

Figure 7. Amplifier output with 20,000 Hz signal driven into slight clipping. Note that this plot is different from both the plots for 1000 Hz and 20 Hz signals. Instead of a "flat top" or "slant" clip, the waveform is taking on other distortions. This too is quite common for many amplifiers. Frequency extremes basically tax the performance of most amplifiers. For high frequencies, more often the clipping behavior is similar to what is starting to occur above. In this case the power supply of the amplifier is not the major cause for the distortion; more likely the circuit design is the reason. Making things work (well) at high power and high frequencies simultaneously is harder to do.

Figure 8. Amplifier output with 20,000 Hz signal driven into bad clipping. Notice how different this waveform looks as compared to the ones in Figures 2 and 6. The amount of clipping is roughly the same, but the waveform sure looks different! Again, clipping at high frequency extremes often produces waveforms with strange looking artifacts. Ironically, if one were to listen to this signal (20,000 Hz, assuming they can hear that high), it might not sound too distorted! The reason? The distortion products would be too high for any human to hear. What you might hear instead are some lower frequency artifacts.

Conclusions: Measurements were made on a "typical" low to mid power amplifier intended for use in a car audio system. The behavior of this amp is similar to many (but not all) amplifiers intended for car audio use. We can conclude:

  1. The amplifier's maximum output power capability is higher when the DC supply voltage is higher (running on battery alone vs. when the engine is running).
  2. The amplifier performs best at mid band frequencies. Clipping behavior at frequency extremes shows that the waveform degrades in a less controlled manner than at mid frequencies. Clipping at low frequency extremes is largely due to the power supply being heavily loaded while at high frequency extremes circuit design is more of a factor.

Back links and E-mail

E-mail to Joe Roberts
Joe's home page...