Why HiFi? pt 6

Amplifiers

I figure it is time to talk about amplifiers. Oddly enough, there has been very little change in amplifier technology in the last 40 years, that is to say, the basic designs have remained the same, so we are looking at mostly simple refinement of what was already there. This, of course, is why many audiophiles use amplifiers that were made in the ‘50s, ‘60s, and ‘70s and are perfectly happy, especially when talking about tube amps since the best sounding tubes were made in the ‘50s and ‘60s because due to the rise of the semiconductor, manufacturing of tubes became no longer profitable. Yet, there are still hundreds of amplifier companies promoting their particular take on one or more of these technologies to claim your buying dollar, so what does it all mean, and why one over another?

To begin with, an amplifier serves two functions, to increase gain (signal amplitude or volume) and amplify power (driving power/Watts, signal strength). There are many amplifiers that only amplify power without increasing gain, most preamps fall into this category, in order to increase dynamic headroom and push signals over long cable runs. On the other hand, many power amplifiers have multiple gain stages in order to provide enough volume for low-efficiency speakers.

Amplifier Class

So, we will start with amplifier classes as this applies to both solid-state and tube designs, and makes up for the majority of the differences in amplifiers. The oldest and simplest design is Class A, which is pretty much universally considered the best sounding, mostly because it uses a single device (tube or transistor) to reproduce all 360º of signal oscillation. The drawback of Class A is waste heat and inefficiency.

 Which brings us to Class B which uses two devices, one to perform the top half of the sine wave and the other to perform the bottom half of the sine wave allowing the reciprocal device time to cool, reducing heat waste and increasing efficiency. The drawback here is what is called switching distortion, caused by the lag time of a cold device turning on and off.

The answer to the switching noise issue was Class AB in which each device extends beyond its 180º of the sine wave in essence never completely shutting off reducing switching noise.

A further refinement was Class A/AB in which the amplifier operates in Class A for a given amount of the time and Class AB when higher outputs are required.

Next in the line of Audio amplifiers is Class D, often referred to as digital amplifiers as they employ pulse width modulation. Class D Amplifiers are non-linear (each device only reproduces a small sample of the sine wave) switching amplifiers and can theoretically reach 100% efficiency. Long felt only suitable for subwoofer amplification due to the high levels of switching distortion, several breakthroughs were made in the early 2000s allowing them to be used for general amplification. While not as musical as a pure Class A amplifier and generally lacking the current capacity of linear designs, they allow for higher outputs at a lower manufacturing cost.

The last of the Classes that are applicable to audio are Class G and Class H (which are similar) both of which are almost as much power supply designs as amplifier designs modulating the power supply voltage to meet signal demands, in essence, providing power to the amplifier more efficiently producing less heat and greater output. While not popular, I have satisfactorily used a Class H amplifier for my Home Theater for over 40 years and it is still one of the best sounding solid-state amps I have heard.

Single-Ended Vs. Bridge Vs. Differential

Ok, this is pretty simple, single-ended means that you are using a single amplifier circuit that operates in reference to ground. A bridge amp or push-pull amp consists of two equal amplifiers 180º out of phase of each other in a balanced configuration (IE: one amp is positive the other is negative). The advantages of this are several, first and foremost (ideally, given sufficient power supply) you get four times as much power for a given power supply voltage (double the power for the use of two amplifiers vs. one and double again for half the impedance seen by each amplifier). Again ideally, given two identical amplifiers, amplifier noise and distortion are self-canceling thus significantly lowering the noise floor and distortion. Unfortunately, the same effect is the major drawback, since no two amplifiers are exactly the same, differences in noise and distortion are additive as are differences in signal reproduction. The result being, that single-ended amplifiers tend to be warmer and more musical, whereas push-pull amps tend to be quieter and more analytical. A differential amplifier is a single-ended amplifier that the signal path is isolated from ground giving it some of the advantages of a balanced circuit (tube amps that employ output transformers have differential output).

Voltage Amp Vs. High Current Amp Vs. Voltage Mode Vs. Current Mode Amplification

Two very disparate concepts that are easily confused, to understand the first, you must first understand the relationship between voltage and current. Power (Wattage) is the sum of voltage times current and current is equal to voltage divided by resistance or in the case of amplifiers impedance (in the case of electromechanical motors, like speakers, impedance varies with frequency). Voltage is pressure and current is volume, a great analogy is water in a river, voltage is the speed of the water whereas current is how much water is in the river (to continue the analogy, resistance is the size of the channel through which the water must pass, and wattage is the amount of water that passes a given point), so you can see that the same amount of water can pass a given point in a fast-moving stream as in a slow-moving river. So in the same way, you can increase wattage by increasing either voltage or current. And similarly, if you have a large current flow, it can adapt quicker to changes in impedance, which is why high current amps have better bass control than low current amps with the same power rating and why high current amps tend to have more dynamic headroom. This is also why solid-state amps often have better bass control than tube amps which are generally voltage devices rather than current devices.

Voltage Mode Amplification vs. Current Mode Amplification is a completely different subject and has to do with how the signal is modulated. Simply put, in voltage mode, voltage is modulated to create a sine wave and in current mode current is modulated. Again the key is speed, current mode amplification is a magnitude faster than voltage mode (most microwave broadcast is done using current mode amplification). This of course means that current mode amplifiers have a much greater linear bandwidth (frequency response across the entire power spectrum, amplifiers tend to have a wider frequency response at half power than at full power), but that isn’t the main advantage of current mode. To really get a handle on this we need to talk about one of the pivotal subjects of HiFi, negative feedback. Essentially a small bit of signal is looped back out of phase to cancel out distortion making an amplifier much more stable. The problem with this is the amount of time it takes to make that loop creates a bit of signal distortion known as Transient Intermodulation Distortion (TIMD). It is TIMD that accounts for that hard bright edge associated with solid-state amplifiers (ironically, the slew rate [rise time] of tube amps is too slow to produce TIMD making it silly for non OTL tube amps to not have negative feedback), which is why many high-end solid-state amplifiers choose to forego negative feedback (which in turn makes production cost a magnitude higher as only about 1% of the parts manufactured can achieve a tight enough spec to overcome this lack). In a current mode amplifiers the negative feedback loop is at such a high frequency as to produce no audible TIMD. So why aren’t all solid-state amps current mode? well, the drawback to current mode amplification is that it requires a fixed impedance load, which of course, speakers are not. Luckily, in recent years (the last 15 or so) some clever amplifier designers have discovered a workaround for this problem, and current mode amplifiers are becoming a thing.

Tubes Vs. Solid-State

I know, this is where you expected me to start, but I felt it best to cover the basics first. Everything you have been told about why audiophiles prefer tubes over solid-state is probably wrong. The standard line put forth by electrical engineers who are not audiophiles (which sadly includes many tube amp manufacturers) is that tube amps roll off the highs making them sound warmer and they produce some magic distortion that audiophiles like the sound of. While there is a grain of truth in this myth (as all good myths have a grain of truth), it is entirely wrong. It is true that many poorly made tube amps roll off the high frequencies and a bit of the low frequencies also, but this is not a product of them being a tube amp as much as it is just poor design, but many solid-state amplifier manufactures roll off the highs and muddy the bottom end in order to make their amps sound more “tube-like”, but they really are just bad sound solid-state amplifiers no matter how popular they are. It is also true that tube amps tend to produce higher levels of Total Harmonic Distortion (THD) than most solid-state amps, it is pretty much universally accepted that they are inaudible levels of THD, but THD is something that is easy to measure and fairly easy to counter so solid-state manufacturers focus on that. The real story goes back to TIMD, which is both hard to measure (in fact most solid-state manufacturers don’t), significantly more audible, and harder to counter. It must also be admitted that in blind A/B tests subjects will often choose the amplifier with higher TIMD because it creates the appearance that you are hearing more (which of course you are, your hearing the TIMD). Ironically, tube amps by not producing TIMD, even though they have a slower slew rate (meaning by definition that they are less accurate), allow the listener to hear more resolution and inner detail that would normally be covered up by the TIMD.

One answer that designers have tried to deal with the sonic disparity between tube and solid-state sound is Hybrid tube/solid-state amplifiers imparting some of the benefits of solid-state while retaining some of the musicality of tubes. Most employ a tube front end with solid-state outputs, though I have seen some that use a solid-state front end with tube finals. Why this solution works is a mystery to me (one would assume that you would just be adding the deficits of both technologies but for the most Hybrids do sound better than equivalent solid-state amplifiers).

Looping back to current mode amplifiers, though I do love them and feel they are more tube-like, I still find tube amps to be more musical, go figure.

The Fiddly Bits

Now we get down to the nuts and bolts, literally (well the figurative nuts and bolts). The core of any amplifier is its power supply, and an amplifier is only as good as its power supply regardless of the other technologies involved. The key parts of a power supply are the transformer (which converts the voltage from 120v or 220v to something more manageable), the rectifiers (which convert from AC to DC and regulates the voltage), and the capacitors. Capacitors serve two functions; storage for dynamic headroom and filtration for purity of signal. Different capacitor manufacturers excel at different functions and if you want to judge the specific components in your amplifier you would need to research that specifically as it would entail more than the entirety of this article, but the rule of thumb is more and bigger is better. As to rectifiers; that is an art and science all in its own and beyond my ken as well as the scope of this article. The preferred transformer on the other hand appears to be the toroidal transformer as it is more efficient, more compact, quieter (less hum) with less flux leakage (magnetic interference), cooler (less heat), lighter, and simpler (thus cheaper to make).

Moving on from power supplies but continuing with transformers, the real magic of a tube amp is the output transformer which allows the amplifier to match impedance with the speaker (converting current into voltage at the same time). According to David Manley’s book (yes he wrote the book on tube amps) a tube amp is only as good as its output transformer, which is why OTL manufacturers opt to not use them. At least one manufacturer used output transformers on their solid-state amplifiers (McIntosh) which had the benefit of protecting the speaker from DC and protecting the amplifier from direct shorts.

The holy grail of tube amplifiers is the single-stage single-ended triode amplifier. Essentially this is a single-ended Class A amplifier, with a single stage of amplification. The major drawback of these is power output (maxing out at about 10 Watts), meaning a very limited selection of available speakers. On the other hand, this is ideal for headphones and I sometimes wonder why anyone would bother to make anything else. (One of my all-time favorite headphone amplifiers is the $149 Schiit Audio Vali 2, though not as quiet as its big brother the Lyr 3, for the money it is a truly amazing and musical amplifier.)

Someone once told me that the power output of a solid-state amplifier is only limited by its ability to dissipate heat, which is why discrete circuits perform better than IC chip based amplifiers, so big heat sinks are a good thing (of course pointless without a sufficient power supply to support it). MOSFET (Metal Oxide Semiconductor Field Effect Transistor) transistors are considered the best sounding power transistor and JFETs (Junction Gate Field Effect Transistor) are considered the best for low-level circuits (they are quieter but not as high current). Direct-coupled amplifiers are considered superior to capacitance-coupled amplifiers for their purity of circuit but do have the disadvantage that they are susceptible to spurious noise and can pass DC (Direct Current) to the speaker (which is an almost instant speaker killer).

Conclusions

That about sums up my 60+ years of audio amplifier knowledge with the exception of the occasional spurious bit of technology like Super Feedforward (a kinda black box technology designed to one-up negative feedback eliminating TIMD). As always, while technology gives you guidelines in what to test, the only true test of an amplifier is how it sounds using your source, and in particular, your speakers (or headphones). It is always important to remember that every amplifier designer is building what they believe to be the best sounding amplifier out there, and given the wide range of ancillary components, there probably is no wrong answer. As I said in my first “Why HiFi?”, it is how all of the components come together that makes the difference.