Fosgate Audionics FAA 1000.5
The key to this product is technical innovation. The new Fosgate Audionics FAA 1000.5 power amplifier brings together a host of technical innovations, to give you high power (200 watts times 5 channels) and good sound, in a convenient, easily handled package, and at a bargain price ($2800) compared to other home theater power amplifiers with similar power.
The story of this amplifier is the story of its innovations. Just as three streams join forces to form a mighty river, so three proud audio heritages have joined forces to bring you this mighty amplifier. First is Jim Strickland, one of the most imaginatively creative innovators in audio. As founder and technical director of Acoustat, he introduced the innovative Acoustat electrostatic speakers, which featured direct drive tube power amplifiers and later twin step up transformers (for more linear spectral coverage). Then Jim developed the FET-based Acoustat TransNova series of power amplifiers, giving tubelike sound from a solid state power amplifier circuit, which was a startling innovation back then. Since Rockford bought Acoustat, Jim has continued to contribute his engineering talents to the Rockford group, and he brings his latest thinking, his latest fresh innovations, to this new Fosgate Audionics power amplifier.
The second stream is the heritage from the Rockford Hafler group. With input from both Jim Strickland and engineer Mark Albers, the Rockford Hafler group has produced a line of pro audio amplifiers and powered subwoofers. In the sound reinforcement world of pro audio, the key requirements are high power and ruggedness, combined with portability, so the Rockford Hafler group has gained considerable experience in making power amplifiers that can put out lots of power yet still be in a package that is convenient to handle.
The third stream is the heritage from Fosgate Audionics, also bought by Rockford. Under the leadership of Jim Fosgate and Charles Wood, Fosgate Audionics was one of the pioneers in multichannel or surround sound decoders, and also featured musical sounding tube electronics. So it is natural that this new 1000.5 Rockford power amplifier, a multichannel product with its emphasis on musically good sound for the home (rather than lesser sound quality requirements of the sound reinforcement pro audio world) be given the Fosgate Audionics name.
Let's look now at all the technical innovations that have joined forces in this power amplifier, and discuss the benefits that each of these innovations brings you.
Heat and Efficiency
As we have discussed in previous reviews, the chief enemy of a home theater power amplifier is heat. Heat must be dissipated, and that gets very problematic in a power amplifier that attempts to put five or six high powered channels in one conveniently manageable (and affordable) chassis package. Dissipating lots of heat requires lots of exposed heat fin area, and that in turn would make a multichannel amplifier unmanageably heavy, unmanageably large, and unmanageably expensive.
It would be great if some means could be found to reduce the amount of heat produced by each powerful channel. But how? One obvious tactic would be to employ a different type of circuit for the amplifier's power output stage (which is where most of the heat is generated), a type of circuit which is more efficient. A more efficient type of output stage circuit generates less surplus heat. Indeed, the very definition of power amplifier efficiency involves the amount of power that is actually delivered to your speaker, compared to the amount of power that is wasted by generating unwanted surplus heat.
Why then don't all manufacturers use the most efficient possible type of output circuit for all their home theater power amplifiers, in order to generate the least amount of unwanted surplus heat, and thus be rid of this heat villain? There's a fly in the ointment. Generally speaking, the more efficient the type of circuit, the worse it sounds. The best sounding type of output circuit, class A, sounds best because it keeps each device in the output stage turned on all the time, but it is also the least efficient. Indeed, it generates so much excess heat that it is rarely used in high power monoblock and stereo power amplifiers, and never in a high power multichannel power amplifier (where the multiplication of excess heat by the multiple number of channels would create an impossible heat dissipation problem). The next class of output circuit, class B, is more efficient because it keeps each device in the output stage turned on only half the time, and operates in push-pull (in order to reproduce the total audio waveform from two devices [or groups of devices], each turned on only half the time). But class B output stages don't sound nearly as good as class A, in part because devices suffer time delays and distorting small signal nonlinearities when they are first turned on.
In between class A and class B is a type of output circuit called class AB, which operates in push-pull like class B, but which keeps each output device turned on for somewhat more than half the time. Class AB sounds better than pure class B, since the output device that is still turned on (by virtue of being biased to be on somewhat more than half the time) masks or covers up the distorting sins of the other output device just being started up. Class AB has an efficiency between class A and class B, so it gives you the benefit of less of a heat problem than class A, while also giving you better sound than class B. For this reason, most push-pull power amplifiers on the market are class AB. Even within class AB, there are engineering tradeoffs and compromises. As discussed in our Plinius Odeon review, class AB can be richly biased for better sound but worse generation of excess heat, or it can be biased lean for less generation of excess heat in a multichannel power amplifier, with some sacrifices in sonic quality (typically, this sounds less musically natural and more like artificial solid state, with fatiguing glare in the upper frequencies that sounds annoying in its own right and also blocks or clogs genuine musical information).
There are also classes of output stage circuitry beyond class B. Class C is even more efficient than class B, and thus generates even less unwanted excess heat. Class C accomplishes this by turning on each half of its push-pull output stage less than half the time. Generally, class C is biased so that it reproduces only the high peak portions (positive and negative) of the audio waveform, completely ignoring that portion of the audio waveform near the zero crossing. Of course, this class C configuration would have truly awful sounding distortion, since it does not even try to reproduce the entire audio waveform. This means that class C is completely unacceptable for audio, and indeed this high distortion class C is employed only in narrow specialized applications, such as certain kinds of radio transmitters.
But wait. What if the high efficiency benefits of class C could somehow be married to the good sounding benefits of class AB? Hitachi, the makers of the MOSFET output devices used in many audio power amplifiers, created such a marriage when they developed what they called the class G type of output circuit. What is class G? Basically, it is a push-pull output stage that looks like a class AB output stage, but with an extra output device added in a series ladder to each class AB output device on each side of the push-pull circuit. Thus, where a simple push-pull class AB output stage would have two output devices, one on each push-pull side, the class G output stage has four devices, two per push-pull side, and on each side there are two devices basically connected in series. These two devices on each side, connected to each other in a series ladder, could even be exactly the same type of MOSFET output device, in the class G circuit as developed by Hitachi. This means that they would split the power supply rails voltage between themselves, each handling half the voltage. And this in turn means that the lower device could handle one quarter the total power capability (power is proportional to voltage squared), while the upper device could hopefully handle the remaining three quarters.
But how does this series ladder improve efficiency? After all, way back in the days when solid state output devices had only limited voltage and current capabilities, it was commonplace to arrange multiple output devices in arrays that were both series and parallel, but they all worked together in class AB, without yielding any higher efficiency. The secret to class G's greater efficiency lies in a clever trick. The upper device in the series ladder is biased so that it operates in highly efficient class C, reproducing and amplifying only the peaks of the audio waveform. But doesn't this produce awful sounding distortion? No, because the lower device in the series ladder operates in class AB, and thereby fully reproduces and amplifies the lower portions of the audio waveform, those portions nearer the zero crossing than the peak portions where the class C upper device in the series ladder joins the fray.
Let's look at a specific example to see how this works, and for an example let's use the same numbers that in fact pertain to this Fosgate Audionics 200 watt-per-channel amplifier. Assuming that the two devices connected in the series ladder on the plus side are the same type of device, they will naturally tend to split the rails voltage equally between themselves, which means that the lower device will naturally handle that portion of the audio waveform from the zero crossing (actually, from just below the zero crossing, depending on how rich its class AB bias is set) up to half the rails voltage, and this in turn means that it will handle up to 50 watts' worth of the (plus part of the) total output power of 200 watts, when that channel module is driven to its full 200 watt output power capability (and the same will be true for the minus side). Then the upper device in the series ladder is biased so that it doesn't even turn on until the audio waveform peak demands more than 50 watts' worth of output power. So this upper device joins the fray, to reproduce and amplify the audio signal, only on those peak portions when (and if) those peak portions rise above the 50 watt level. Note that during quiet audio passages, or if you keep the peak power level low even during maximum audio passages (because you have very efficient speakers, or because you have set the volume level low), then the amplifier won't ever be called upon to put out more than 50 watts peak, in which case the upper class C device will never even be turned on.
How does this configuration improve efficiency? First, the lower device in the series ladder, operating in class AB, never has to put out more than 50 watts' worth maximum (actually, that 50 watts is of course further split between the plus and minus sides). Thus, the heat sink requirements for this lower device can be very modest, the same as that for a common 50 watt multichannel receiver. Then, because the upper device operates in class C and is only turned on for a fraction of the time (only on peaks), its average excess heat generation over time is small, even though it is putting out high peak power. So the requirements are also modest for the heat sink for this class C upper device, in spite of its high power output, because heat dissipation by a heat sink into room air is a function of a long term average transfer of energy over time. In sum, you get a powerful 200 watts of power from each channel, with nearly the very modest heat sink requirements of a 50 watt receiver. This saving in heat sink requirement pays you big dividends in this amplifier's low cost, moderate package size, and low 42 pound weight (the pound weight and the differing kilogram weight listed in the owner's manual are both wrong). Indeed, the Fosgate Audionics FAA-1000.5 is so efficient that its modest heat sinks can be housed internally, just as in a 50 watt receiver. This means that the exterior housing of this amplifier can be attractive, smooth, and easy to handle, without the common sharp edged protruding heat sinks that cut your hands.
Class G Sound Quality
What about the sonic fidelity of the class G circuit configuration? Does class G manage a smooth transition, accurately tracking the audio waveform, from its class AB phase, when only the lower device is on, to its class AB plus class C in series phase, when both devices are turned on? The technically astute among you might well be wondering if the class C upper device doesn't commit distorting errors when it is first turned on. Well, it surely does, but, when this upper device is first turned on, the lower device is already putting out a full 50 watts of good sounding class AB loudness, and this easily masks the errors committed by the upper device at its first turn on. It's exactly the same kind of usefully beneficial masking that is employed by richly biased conventional class AB output stages, only in this case it is like a super rich class AB, since at the moment of first turning on the upper device the lower device is already putting out a very loud 50 watts. The Fosgate Audionics engineering team has also taken care that the upper device can turn on very fast, at microsecond speeds, so that it does not miss out on the beginning of even the fastest transient peaks. Incidentally, the upper device is also engineered to turn off more slowly on the trailing back side of a peak, in order to avoid sonic degradation from residual capacitance in the devices holding energy and causing a delayed overhang.
We purposely put this amplifier through a sonic torture test, throwing all kinds of audio signals at it, including high frequency, high energy transients. And we listened at many different levels. We were constantly on the lookout for any sign of sonic change or sonic degradation when the class C device joined in. In other words, we were probing to see if there was any sonic penalty to class G, either from the transition from class AB to class AB+C, or from the incorporation of the class C device's contribution into the audio signal at louder levels. We also tried this at the three different rails voltages available in this amplifier (see below for further discussion of this). The good news is that, with the native sonic capability of this amplifier at low volume levels as a reference baseline (q.v. discussion below), we did not hear any further sonic changes of significance from the addition of the class C device at louder volume levels, even on the most difficult transients. The bottom line is that class G works, as implemented by the Fosgate Audionics team in this amplifier. It does indeed give you more power for less money, with less heat, in a more convenient package, without further sonic penalty.
In fact, in some ways the sound of this amplifier should even improve slightly at louder levels, when the class C device turns on and joins the fray. How can this be? Once again, you can thank the creative imagination of Jim Strickland. The original Hitachi class G circuit proposed that the two MOSFET devices in the series ladder be exactly the same, and indeed this is the way they are configured in the Rockford Hafler pro audio amplifiers that already employ class G. But for the sonically superior home audio Fosgate Audionics amplifier, Jim wanted to do better. So, in this Fosgate Audionics amplifier, Jim developed an innovative way to employ slightly different MOSFET devices for the lower and upper devices in the series ladder, on each of the push-pull sides of this amplifier's class G output stage. Specifically, on one push-pull side, the lower device is a PNP MOSFET device, whereas the upper device is an NPN MOSFET device. Meanwhile, on the other push-pull side, these roles are reversed, with the lower device being an NPN and the upper device being a PNP. Thus, the entire push-pull output stage is complementary, and each class G series ladder on each side is uniquely also complementary.
How does this unique innovation help sonics? As we have previously discussed in other reviews, the complementary symmetry topology, employed in many higher end audio electronic products (preamplifiers as well as power amplifiers) promises the theoretical advantage of reducing overall distortion, and of virtually eliminating even order distortions (itself a controversial benefit, since it destroys the musically natural progression of even and odd order harmonics, and leaves behind only the uglier sounding odd order harmonics and intermodulation byproducts). But the practical implementation of complementary symmetry in solid state circuits fails to even realize in practice this theoretical advantage. That's because the PNP devices used in one half of the complementary symmetry topology are not in fact true mirror images of the NPN devices used in the other half, and they actually behave differently and amplify the audio signal differently. Typically, the PNP devices are slower, have different voltage gains, different time constants, etc. than the NPN devices. Thus, the two halves of these complementary symmetry circuits (both bipolar and FET) are not in fact truly complementary, and do not in fact balance each other out. Thus, these circuits still treat the audio signal asymmetrically, reproducing and amplifying the top half differently than the bottom half. This means that they still produce some even order distortions. Indeed, the introduction of these even order asymmetries may well also explain why many complementary symmetry circuits tend to have a richly warm, pleasantly musical sound, even with bipolar transistors (which otherwise tend to sound leaner and more sterile), and certainly with FETs.
In the traditional complementary symmetry output stages, the devices on one side of the push-pull output stage are all NPN, and those on the other side are all PNP. Thus, the aforementioned asymmetry persists. But Jim Strickland's innovative insight for this new amplifier was to intentionally mix the two types in each of the two push-pull sides. The plus side of the push-pull output stage employs both a PNP and an NPN output MOSFET device, and so also does the minus side. When all four output devices are brought to bear (as they are for all peaks between 50 watts and 200 watts), there is a much better, more truly symmetrical balance between the amplifying behavior of the plus and minus halves of the push-pull output stage, because each half is now employing both a PNP and an NPN device.
Consider voltage gain as a specific example of such behavior. Because the PNP MOSFETs have a lower voltage gain than the NPN MOSFETs, the conventional complementary symmetry practice of stacking only PNP types on one side of the push-pull output stage, and only NPN types on the other side, would result in asymmetrical voltage gain for the plus and minus halves of the audio signal, thereby creating even order distortion. But, by cleverly stacking mixed types on each side, Jim Strickland is able to achieve a more truly balanced amplifying behavior for the plus versus the minus side of the audio signal, hence less generation of even order distortions. Within each push-pull side, the PNP device achieves less voltage gain than the NPN device, but the two stacked together in their series ladder on one push-pull side achieve almost the same total voltage gain as the similarly mixed pair stacked in their series ladder on the other push-pull side.
This improved symmetrical balance is still not perfect, because the two push-pull sides are still somewhat different, the first side using an NPN device for the lower amplitude portions of the audio signal while the second side uses a PNP (and meantime that first side uses a PNP device for the peak
(Continued on page 78)