Apples vs Apples: Introducing the Common Amplifier Format (CAF) Oct. 07, 2013, by Pat Brown

Transcription

1 Apples vs Apples: Introducing the Common Amplifier Format (CAF) Oct. 07, 2013, by Pat Brown Last year I was contracted to perform a barrage of measurements to be used to help a theme park select between several amplifier brands. They provided the list of tests, and I provided the instrumentation and the sweat. The amplifiers were similar with regard to their power ratings, but differed in how they performed during the testing. Today s smart amplifiers can behave very differently depending on the test stimulus, load impedance, and especially the duration of the test. I found myself needing to tweak the test differently for each one. In some cases, I was measuring how the amplifier is programmed to respond to an event, making many comparisons apples versus oranges and subjective. After days of testing, I was not satisfied that the measured results produced a meaningful comparison between the various models. This specification dilemma had reared its ugly head before, with regard to loudspeaker data. It was solved by the development of the Common Loudspeaker Format (CLF). It s been eight years now since the development of the CLF. It gives sound system designers the information needed to deploy loudspeakers, and this intended purpose kept the format from drifting in scope and becoming bloated. A hallmark of the format is the use of the Maximum Input Voltage (MIV) for characterizing the maximum allowable drive signal to the loudspeaker. The MIV is crucial for proper amplifier selection. It gives the maximum RMS voltage that the loudspeaker can handle, and it is the basis for determining the amplifier size required to deliver it. The CLF is a concise format that is appropriate for its intended task. It was intentionally designed to present system designers with what really matters. The CLF has been wildly successful, and is now the main data format for nearly all of the room acoustics/sound system modeling programs. It is supported by more than 70 loudspeaker manufacturers and developers. Setting Parameters The success of the Common Loudspeaker Format has inspired the development of the Common Amplifier Format (CAF). The CAF is a report format that gives the sound system designer the ratings they need at a glance, along with some fine details for those who are interested. We learned from the CLF that a group of motivated individuals can quickly build the tools they need to do their jobs, and the same development approach was used for the CAF. Like the CLF, the CAF is not a standard, and it was not developed like a standard. It was designed by a small, focus group of individuals that was later expanded to a wider audience as the format neared completion. The development team included manufacturers, consultants, and contractors, each bringing a different perspective on what the CAF should include. We knew what we needed, and we built it. The CAF must at a minimum: 1. Give the amplifier a meaningful power rating for use in design calculations. 2. Provide information on how the amplifier behaves with various test stimuli, and into various loads. 3. Define the test procedures so that they can be repeated by anyone with the appropriate instrumentation. 4. Present a number of secondary metrics of interest to sound system designers, including plots as well as one-number ratings. This information will allow proper deployment of the amplifier, and meaningful comparisons of amplifiers from various manufacturers. Given the nearly infinite number of variables in the testing of electronic devices, some guidance is needed. Since amplifiers are always used to drive loudspeakers, it makes sense to allow the amplifier measurement process to be influenced by some key loudspeaker attributes provided by the Common Loudspeaker Format. 1 of 8

2 Constant Voltage Amplifiers are universally interfaced to loudspeakers using a constant-voltage interface. In a constant voltage interface, the output voltage of the source (the amplifier) is largely independent of the load impedance (the loudspeaker). The interface is optimized for voltage transfer, not power transfer (which is optimal into a matched load impedance). Think of it as an interface that yields controlled power transfer to the loudspeaker without over-taxing the amplifier. The CAF is designed to reveal, at a glance, the minimum impedance into which the amplifier is constant voltage, for various types of signals. What The Amp Sees All the amplifier knows about the outside world is the load impedance to which it is connected. If it s 4 ohms it does not matter to the amplifier if it is single 4-ohm box, (2) 8-ohm boxes in parallel, or (4) 1-ohm boxes in series. The lower the load impedance, the higher the current that the amplifier must source to sustain its output voltage. When it can t source the demanded current, the voltage sags and the distortion rises. These are the effects of amplifier overload. The loudspeaker s impedance is complex, having both a magnitude and phase. A plot is required to show how the loudspeaker s impedance changes with frequency (Figure 1). For sound system design purposes, a rated, or nominal impedance is used in lieu of the actual impedance of the loudspeaker. Figure 1 This rated impedance is found by taking the lowest impedance in the loudspeaker s pass band times 1.2. It is sometimes rounded to a typical value, such as 8 or 4 ohms, and it s assumed to be a pure resistor. Why? Simplification. In audio, rated or nominal means the actual value is too complicated to deal with, so we are going to call it this. The complex impedance curve of the loudspeaker becomes an x ohm resistor for the purpose of power rating calculations. Note that the output voltage of the amplifier will be the same for either load shown in Figure 1 (average or nominal), but its output power will be different, being higher for the rated (nominal) impedance than the actual impedance. This means that your loudspeaker can t actually handle its rated power crazy, but necessary for practical reasons. Amplifiers are tested into purely resistive loads, typically 8, 4 and (maybe) 2 ohms. The amplifier s output voltage is measured, and the power is calculated from the power equation into these resistive loads. Read the previous sentence again, please. This means that there is a disconnect between the actual power from the amplifier into a real loudspeaker and that calculated using a rated loudspeaker impedance. In all cases, the rated 2 of 8

3 impedance is lower than the loudspeaker s true average impedance, which means that for the same voltage, the amplifier is producing more power when driving a dummy load than when driving a real loudspeaker. That s actually good news for the amplifier, and it makes the power rating given the amplifier conservative, which is good engineering practice. Smart Amplifiers Some advanced amplifiers can monitor both the voltage and current at their output, and display the actual power being produced into the load. It s tempting to think that your loudspeaker is safe if the amplifier is putting out less watts than the loudspeaker is rated to handle. However, if you try to run the amplifier up to your loudspeaker s power rating, you will likely burn up the loudspeaker. This is because the loudspeaker s power rating was calculated using a rated impedance that is actually lower than its actual average impedance. An example is in order. Let s consider that the loudspeaker shown in Figure 1 has an MIV = 24 Vrms. I ll connect it to an amplifier that can easily produce 24 Vrms without clipping the signal. Figure 1 shows the power into the loudspeaker from the amplifier producing 24 Vrms of pink noise, using both the rated and average loudspeaker impedance. This loudspeaker would have a rated power of about 85 watts, even though it is only handling about 42 watts at its MIV. So, your 85 watt loudspeaker can t really handle 85 watts, and actually producing 85 watts into it (as monitored by your amplifier-control software) would require that you exceed the MIV of 24 Vrms, resulting in smoke and a funny smell. Does your head hurt yet? Just remember, a loudspeaker power rating is not an actual measure of the power being fed to the loudspeaker. It is the calculated power that would be produced by applying the MIV to a dummy resistor equal to the loudspeaker s rated (nominal) impedance. So, given the way that loudspeakers get their power ratings, you actually need to track the applied voltage to be sure that you don t overheat it by exceeding the MIV. The good news is that you can monitor this with an RMS voltmeter. As clever as real-time power monitoring is, it s not much use for avoiding loudspeaker damage. Maximum Input Voltage When sizing an amplifier, what I really need is a unit that can supply the MIV to the loudspeaker without clipping. This assumes that I want to drive the loudspeaker to it s MIV, with is probably a bad idea, since by definition at the MIV a loudspeaker is approaching thermal failure. Remember, it s the maximum input voltage, which means that something bad had to happen for that to be determined. This is typically an increase in the loudspeaker s impedance, caused by heat. As it gets hot, it s impedance rises, reducing the current draw from the amplifier. Since the current is what moves the voice coil, this means that above the MIV the SPL from the loudspeaker no longer tracks the applied voltage. Increasing the drive voltage by 3 db may only increase the SPL by 2 db. Where did the other 1 db go? You guessed it: heat! It s always desirable to stay as far below the MIV as possible, and there are no sonic penalties for doing so. The idea that under driving a loudspeaker causes damage is a myth. The Common Loudspeaker Format gives the maximum RMS voltage that the loudspeaker can handle. Our attention now turns to the Common Amplifier Format. It gives the output voltage of the power amp for various signal types and load impedances. Used together, the CLF and CAF can assure that an appropriate amplifier has been selected for a loudspeaker. What Must I Know? The Main page of the CAF report (Figure 2) offers some specifics regarding the physical characteristics of the amplifier, along with some one number ratings, including input sensitivity, voltage gain, dynamic range, latency, common-mode rejection ratio (CMRR), input impedance, output impedance and damping factor. It also includes navigation links to other parts of the report. 3 of 8

4 Figure 2 Voltage-Power Matrix The heart of the CAF is the Voltage-Power Matrix (Figure 3). It gives the amplifier s response to burst stimuli, sine waves and pink noise into typical load impedances as approved by the manufacturer. The burst stimuli reveal the amplifier s output voltage to signal transients. The sine waves reveal the amplifier s ability to sustain a continuous tone. The pink noise tests are more like the music and speech that most amplifiers are used to reproduce. A separate matrix is given for each supported operating mode of the amplifier, such as 1-channel driven, all-channels driven, bridged-mono, parallel-mono, etc. 4 of 8

5 Figure 3 The matrix is divided into three main sections. The Audio section gives the input and output voltage to the amplifier for its approved load impedances. The Supply section gives the current and power drawn from the utility company to produce the audio output. The Waste section gives the byproducts of this process, namely heat production and fan noise. This gives the sound system designer the metrics they need for audio output, circuit breaker sizing, and cooling system requirements. It also reveals how noisy the amplifier is under various load conditions, something that is often found out too late after the amplifier has been installed in a quiet meeting room or sound booth. The matrix is presented in two forms. The first one uses linear units (e.g., volts, watts). The second, which is calculated from the first one, presents the same data as levels in dbv, dbu, dbw and dbm. Since there exists no universal agreement on what to use, the CAG gives them all, but in a way that does not clutter the report. Frequency Domain Plots Some specifications are best given by a plot rather than a single number rating. The CAF includes many. There are frequency domain plots for magnitude response, phase response, group delay and distortions. Most of the plots are for multiple load impedances, so that you can see how the amplifier s frequency response changes depending on the load impedance (Figure 4). There is also a plot of the amplifier s noise floor spectrum (Figure 5), including how it changes when a square wave (60 Hz at 80 ma) is applied to the shield connection of its input connector (Figure 6). This so-called Pin 1 Test is described in detail in the AES standard on interconnections, titled AES48 - Grounding and EMC Practices - Shields of connectors in audio equipment containing active circuitry. The CAF is the first amplifier specification to contain this information. 5 of 8

6 Time Domain Plots There is a time domain plot for the 1 khz square wave response (Figure 7). This is one of my favorite old school tests, and one of the most difficult for modern amplifiers. The final plot reveals amplifier s short-circuit behavior (Figure 8). The short is actually 0.5 ohms, which is representative of a short at the end of 50 feet of AWG12 cable. This test is last, because it sometimes smokes the amplifier. Figure 4 Figure 5 6 of 8

7 Figure 6 Figure 7 7 of 8

8 Figure 8 Coming Up Next The significance of the fact that amplifiers are interfaced as voltage sources to loudspeakers has eluded many audio practitioners. Applying a power mindset to a voltage interface creates misconceptions about what matters and what doesn t. A more power is better mentality leads to specifications that are exaggerated, misleading and in some cases simply wrong. Power ratings for both amplifiers and loudspeakers are practically useless if the voltages and impedances used to determine them are not given. The CLF and CAF present specifications in a way that allows them actually serve their purpose, which is to create a harmonious relationship between the amplifier and loudspeaker. Next time I will present the CAF for a popular amplifier and provide some design examples for deploying it for several loudspeaker types. Pat & Brenda Brown lead SynAudCon, conducting audio seminars and workshops online and around the world. For more information go to 8 of 8