Most audio power amplifiers suffer from a defect known as "crossover distortion". This distortion is particularly troublesome at low output levels. At low power levels, the crossover distortion can rise to a high percentage of the output level and become the dominant source of distortion.
Dick Olsher once said that "the first Watt is the most important Watt". We agree!
At normal listening levels, in a studio control room or home environment, an amplifier spends much of its time delivering about 1 Watt of power. Musical peaks may demand 50 to 100 Watts, but most of the musical content is reproduced using a few Watts. For this reason, 1st-watt performance is crucial to the overall sound of the amplifier.
Amplifier THD+N is often only specified at high output levels. It is relatively rare to see specifications for the performance at 1 Watt. An amplifier that is relatively clean at high output levels may deliver poor performance in the 1st Watt, but this will not be reflected in the published (high output level) specifications. Look for both low and high-level specifications when selecting a power amplifier.
Crossover distortion occurs every time an amplifier output device turns on or off. In most amplifiers, these transitions occur every time the output current reverses. One output device is dedicated to pushing the speaker cone outward, while the other is dedicated to pulling the speaker cone inward. We call this combination of output devices a push-pull output stage. Larger amplifiers use multiple push and pull output devices running in parallel, but the operating principle is the same.
When an amplifier attempts to damp the motion of a speaker, the speaker may cause a series of current reversals that increase the occurrence of crossover distortion events.
Crossover distortion is particularly insidious in that it is not usually harmonic. The timing of the 0-current crossover events is a function of all of the frequencies being reproduced by the amplifier. This interaction causes intermodulation distortion (IMD) which is much more audible than harmonic distortion (THD). The distortion tones produced by IMD do not occur where our ears expect to hear distortion. IMD can cause two musical instruments to interact in a way that produces distortion tones that are never naturally produced by either instrument. For this reason, IMD sounds unnatural and unmusical to our ears.
In contrast, harmonic distortion tends to be less noticeable than IMD because it resembles the distortion produced by musical instruments. Each musical instrument produces a rich spectrum of harmonic distortion in a unique pattern that distinguishes it from all other types of instruments. A flute and a violin may play the same note but they sound very different because they produce very different patterns of harmonic distortion. Variations in harmonic distortion give musical instruments their unique voice.
If an amplifier adds some harmonic distortion, this may be small relative to the harmonics produced by the musical instruments, and the amplifier's distortion may go unnoticed. If the harmonic distortion produced by the amplifier is high enough, it may cause a noticeable change in the sound of some instruments. Some listeners enjoy the "warmth" or "color" added by an amplifier's harmonic distortion. In extreme cases, harmonic distortion can make instruments difficult to identify.
Crossover distortion has a strong IMD component and this makes it much more audible and objectionable than harmonic distortion. For this reason, amplifiers use a number of topologies and techniques to minimize the crossover distortion produced by push-pull output stages.
There are many types (or "classes") of linear amplifiers that use push-pull output stages. These include:
Most class-D (switching amplifiers) also use push-pull output devices. The push-pull devices in these switching amplifiers make abrupt transitions between full-on and full-off states. These switching transients occur at a high repetition rate and generally contribute to an overall increase in distortion and noise.
In a class-B amplifier, only one output device is on at a time. To make matters worse, there is usually a dead-zone where neither device is turned on. This dead-zone is crossed whenever the current to the speaker reverses. Every crossing causes a significant ripple in the output waveform. Class-B amplifiers perform very poorly at low output levels, but waste no power when idle.
Class-AB amplifiers attempt to reduce crossover distortion by overlapping the operating regions of the push and pull output devices. This overlap (or bias) allows a gradual transition between the push and pull devices and this greatly reduces the crossover distortion. But this overlap means that the push and pull devices are lightly pulling against each other when the amplifier is idle. This overlap increases the idle power consumption of the amplifier and the amount of overlap must be carefully controlled. Too much overlap will cause overheating. Too little overlap will cause a rise in crossover distortion. The distortion performance of a class-AB amplifier is largely a function of how well this overlap is executed. Temperature, loading, and aging can take a class-AB amplifier out of this sweet spot: If the overlap is lost, the class-AB amplifier operates as a class-B amplifier. If the overlap gets too large, the class-AB amplifier spends more time in a low-efficiency class-A mode.
Class-A amplifiers do not produce crossover distortion. The output device or output devices do not transition between on and off states. All output devices are always on in a class-A amplifier. This configuration wastes massive amounts of power, but it prevents crossover distortion. Class-A amplifiers are known for producing very little distortion at low levels.
Some low-power class-A amplifiers are single-ended. These amplifiers use an active device to push the output in one direction while the load pulls it back in the other direction. Single-ended class-A amplifiers tend to create substantial even-order harmonics, especially at higher output levels. These even-order harmonics are much more pleasing to the ear than the IMD distortion produced by crossover distortion, but they can add a substantial coloration to the music.
Push-pull class-A amplifiers attempt to reduce the even-order harmonics without causing crossover distortion. These class-A amplifiers tend to be much more transparent than single-ended class-A amplifiers.
Class-H and class-G amplifiers are class-AB amplifiers with additional push-pull output devices that activate when large voltage peaks are encountered. When these extra devices turn on or off, they create additional crossover distortion transients. A class-AB has one crossover region near 0 output current, but class H or G amplifiers have 3 or 5 crossover regions. These additional crossover regions increase the crossover distortion problems by a factor of 3 or 5 relative to a class-AB amplifier. The advantage of switching between low-level and high-level output devices is that the efficiency can be greatly improved relative to a traditional class-AB amplifier. Class H or G amplifiers offer a good alternative to class-D amplifiers when high efficiency is required.
Most amplifiers use a substantial amount of feedback to reduce the distortion at the output of the amplifier. A feedback circuit compares the amplifier output to the input signal and produces a difference signal that is added to the input. This feedback corrects distortion after it starts to occur. If a distortion event happens quickly, the feedback network may be too slow to correct the distortion.
Crossover distortion can create fast transients that are too short to be corrected by a feedback network. The feedback network can actually extend the period of time over which the error occurs. Most class-AB amplifiers will show crossover distortion artifacts that have been stretched in time by ringing in the feedback network. This ringing occurs because the feedback network must settle after encountering the transient caused by the crossover event. Feedback networks correct distortion after it starts to occur.
Benchmark's AHB2 power amplifier has a unique patented feedforward error correction system that virtually eliminates all traces of crossover distortion. Like a feedback system, the amplifier output is compared to the input and a difference signal is created. But, instead of feeding this signal back to the input, it is added to the output to cancel the crossover distortion. The distortion is corrected before it reaches the final output of the amplifier. There is no ringing and no settling time because we do not need to wait for a feedback loop to respond. The correction happens in perfect synchronization with the error, and the entire error is removed.
A deeper look into the AHB2 will reveal a sophisticated combination of feedback and feedforward systems. It is important to note that it is the feedforward system that handles the difficult task of removing the fast transients produced by crossover distortion. The feedforward system eliminates crossover distortion before the feedback is applied. For this reason, the feedback systems are fully isolated from the crossover transients.
The AHB2 is cleaner than the best class-A amplifier and it is almost as efficient as a class-D switching amplifier. The feedforward system is so effective at eliminating crossover distortion that we were able to leverage some of the efficiency features of class-B and class-H amplifiers. To achieve this, we reduced the bias current in the class-AB output stage to achieve a standby power consumption that approaches that of class-B amplifiers. We also leveraged the efficiency provided by class-H tracking rails. Without the feedforward error correction, these efficiency enhancements would have produced higher crossover distortion. The feedforward system removes all traces of the crossover distortion that would have been produced by these enhancements.
Sometimes it is possible to have the best of two worlds.
Secrets contributor Sumit Chawla recently caught up with Benchmark’s VP and Chief Designer, John Siau to get a little more in-depth on several subjects.
Q: "Benchmark is one of the few companies that publishes an extensive set of measurements, but you also balance that with subjective testing. Can you talk about the equipment, the listening room, and the process for subjective testing?"
Q: "Was there ever a time where you learned something from a subjective test that was not captured by measurements?"
Q: "You conducted some listening tests to determine whether distortion in the “First Watt” was audible. What test material did you use for this, and what did you find?"
Q: "The AHB2 amplifier incorporates THX Audio Achromatic Amplifier technology. When and how did the partnership with THX come about?"
Q: "Linear power supplies have been and remain quite popular in high-end devices. You favor switch-mode power supplies. When and why did you make this switch?"
... and more!
At Benchmark, listening is the final exam that determines if a design passes from engineering to production. When all of the measurements show that a product is working flawlessly, we spend time listening for issues that may not have shown up on the test station. If we hear something, we go back and figure out how to measure what we heard. We then add this test to our arsenal of measurements.
Benchmark's listening room is equipped with a variety of signal sources, amplifiers and loudspeakers, including the selection of nearfield monitors shown in the photo. It is also equipped with ABX switch boxes that can be used to switch sources while the music is playing.
Benchmark's lab is equipped with Audio Precision test stations that include the top-of-the-line APx555 and the older AP2722 and AP2522. We don't just use these test stations for R&D - every product must pass a full set of tests on one of our Audio Precision test stations before it ships from our factory in Syracuse, NY.
Paul Seydor of The Absolute Sound interviews John Siau, VP and chief designer at Benchmark Media Systems. The interview accompanies Paul's review of the LA4 in the December, 2020 issue of TAS.
"At Benchmark, listening is the final exam that determines if a design passes from engineering to production. But since listening tests are never perfect, it’s essential we develop measurements for each artifact we identify in a listening test. An APx555 test set has far more resolution than human hearing, but it has no intelligence. We have to tell it exactly what to measure and how to measure it. When we hear something we cannot measure, we are not doing the right measurements. If we just listen, redesign, then repeat, we may arrive at a solution that just masks the artifact with another less-objectionable artifact. But if we focus on eliminating every artifact that we can measure, we can quickly converge on a solution that approaches sonic transparency. If we can measure an artifact, we don't try to determine if it’s low enough to be inaudible, we simply try to eliminate it."
- John Siau