Power Amplifiers in Public Address Systems
An amplifier is the workhorse of a PA system. It doesn't look very exciting, and usually has fewer knobs, lights or buttons than very basic effects or processors (although some - e.g. Crown's xps and xti ranges, and d&b's D6 and D12 - incorporate functions that have historically been the province of Crossovers and Controllers).
A PA amplifier usually takes the form of a 19″ rack-mounted box. Most modern amps are in 2U boxes unless they are very high-powered (over 1kW per channel into 4 Ohms), in which case the box may be 3U or even 4U. A few manufacturers have tried 1U models, but you won't find many of these in professional touring PA systems.
Technically, any device that applies gain (even negative gain) to a signal is an amplifier. To distinguish it from other kinds of amplifier (such as preamplifiers or line amplifiers), the main amplifiers are often described as ‘power amplifiers’. A power amplifier takes a line-level signal and reproduces it in a form that will drive a loudspeaker. It converts a low-voltage, high-impedance waveform into a high-voltage, low-impedance waveform. Its main purpose is to reproduce a low-power signal at high power.
Most modern PA amplifiers use transistors to convert the signal from low to high power. Typically the work is shared by a number of transistors (a dozen or more in most cases), and these can get very warm when the amplifier is running at full power. For this reason, power amplifiers in PA systems are almost always fitted with cooling fans and ventilation slots.
Most PA amplifiers have two channels (Left & Right, A & B, or 1 & 2), providing independent level control over the signal on each.
If all else fails, read the manual!
Before you start, make sure the amplifier has enough ventilation: if the slots are obstructed, or the amplifier is put in a small cupboard with the door shut, there is a danger it will overheat. If you are using more than one amplifier in a single rack, make sure they all ventilate in the same direction: some makes and models ventilate from back to front, others from front to back, and if you mix the two, each will be trying to cool itself using the hot air from the other.
Connect the line-level input signals (using signal cable) to the inputs, and connect the outputs (using speaker cable) to the speakers. Always power up with the volume controls set to minimum. Always increase the volume on each channel gradually.
PA amplifiers generally have only two controls - a volume knob for each channel - on the front panel. Many also have one or more switches for other functions on the rear panel, and some - e.g. Crown's now discontinued Power Base range (also sold under the brand name Amcron) - put the volume knobs on the back as well. Some also have internal switch or link options, which are obviously not intended for everyday adjustment.
Switchable functions often include:
Parallel Inputs. This routes the signal on one input (usually the left channel) to both channels. This is useful if you want the same signal to go to more than one speaker or speaker cluster (e.g. summed mono bass), as it saves having to use adaptors to split the signal.
Bridged Mono. This takes a single signal (again, usually on the left input), and provides two copies at the outputs, one in opposite polarity to the other. If the speaker leads are connected across both channels (using only the positive terminal on each), the combined power of both channels can be used to drive a single loudspeaker.
Limiter On/Off. Some amplifiers have inbuilt limiters, often independently selectable for each channel. The limiter's threshold is set by the manufacturer to prevent the amplifier being driven into clipping. There are few good reasons to switch it off (although some engineers prefer to leave the limiters off on sub bass amps).
High Pass. Some amplifiers have a switchable high pass filter (sometimes the frequency is adjustable or selectable too) to protect bass speakers from subsonic frequencies, or to protect midrange speakers from very low bass frequencies.
Earth/Ground Lift. In systems using balanced connectors throughout, it is good practice to disconnect the screen at one end of any cable linking earthed equipment (e.g. crossover and amplifier). The earth lift switch disconnects the cable screen, saving users the bother of rewiring if earth loop problems surface.
Sensitivity. Usually stated in volts or decibels (dBu), sensitivity describes the level of input that will yield an output equal to the amplifier's stated average power. Some amplifiers have switches that allow the user to select this, enabling the amplifier to be matched with differing equipment (e.g. −10dBV sources), or with other amplifiers that have fixed sensitivity. Selecting a lower voltage or decibel value increases sensitivity, and will make the amplifier seem louder (as well as making it more prone to clipping). If in doubt, higher values are safer.
Manufacturers use a variety of labels for the volume controls (‘Gain’ is relatively common), but usually they are NOT making the signal bigger: they are attenuators. Set fully clockwise, they allow the full input signal - at a level determined by the previous device in the signal path - to pass to the output amplifier stage, which has a fixed amount of gain. In any other position, they have the effect of reducing the amplifier's sensitivity. This may sometimes be desirable (if, for example, your amplifiers are fed directly from the mixer's outputs, you will be unable to use the mixer's maximum undistorted output - often higher than +20dbu - unless you limit the amplifier's sensitivity in some way).
Where volume controls are marked, the scale will normally be stated in decibel values from maximum attenuation (fully counter-clockwise, −∞) to no attenuation (fully clockwise, 0 dB). Intermediate points will generally be given as minus values, showing the amount of attenuation (or negative gain) in decibels. With the controls set fully clockwise, the amplifier's full voltage gain will be applied to the input signal (see previous paragraph; also see the notes on Gain and Sensitivity below).
In most live applications the amplifiers will be near the stage and the mixer will be in the audience, so it is impractical for the Front-of-House engineer to adjust amplifier settings during a show. For this reason, many engineers set the volume controls on each amplifier to maximum (0 dB), and control the overall level from the mixer. However, to make optimum use of the system's output and headroom, the amplifier's volume controls should ideally be set in conjunction with sensitivity and crossover settings, so that maximum input and output levels are achieved in all devices at the same time (otherwise, the undistorted output of the whole system is restricted by whichever device first reaches its maximum level).
If your amplifiers have switchable gain and/or sensitivity settings, these should be set in conjunction with your controller/crossover settings.
If the amplifier has limiters, USE THEM. Most decent power amplifiers specify their undistorted output power, and are capable of producing at least double that amount of power when driven to distortion. Most speakers will fail quickly if driven by a lot more power than they are designed to handle. The limiters are there to protect your equipment, and only operate when maximum safe levels have been reached.
Only if you want your passive loudspeakers to work.
Most modern rack-mounted power amplifiers will have balanced XLR input connectors, and Neutrik ‘Speakon’ (usually NL4) output connectors. Most will also have alternative input connections using 1/4″ jack sockets (many also have connector strips for hard-wiring the inputs), and alternative outputs on Pomona (dual banana jack) and/or binding posts. Similarly, most have IEC mains power inputs (although some have hard-wired mains cables). Older power amps may have different combinations of connector. If you get one that won't accept your existing input and output connectors, you will also need to make or buy new cables or adaptors.
An amplifier's output power is determined by the quality and capacity of the power supply. Amplifiers generally have one of two kinds of inbuilt power supply: linear, or switched-mode (also known as ‘switch-mode’, ‘switching-mode’, or simply ‘switching’).
power supplies need a larger, heavier transformer to convert mains voltage to the voltage required by the amplifier circuits (the transformers used in linear power supplies are invariably toroidal - that is, they are in the shape of a torus: they look like a large doughnut - with the conductor wires wound round a ring-shaped magnet; this minimises the effect of the magnetic field on nearby signal conductors). In a linear power supply the transformer and main rectifier capacitors are key components. Low-grade components cannot deliver high-grade performance: don't expect a linear-powered amplifier with a small transformer to provide sustained high output power. Amplifiers with linear power supplies are usually larger - and always substantially heavier - than equivalent switched-mode models.
power supplies are lighter and more efficient (they generate less heat than linear supplies). The term ‘switched-mode’ refers to the fact that they use switching (rather than resistive) devices to regulate current flow and output voltage. Although they use a transformer, this is much smaller and lighter than the kind of transformer used in linear power supplies. Historically they have been less capable of sustaining high output current (so were generally less suitable for low-frequency output, which requires higher current). They are also more expensive to manufacture, and have a reputation for being less reliable. However, modern switched-mode amplifiers will generally do anything an equivalent linear-powered amplifier would do. Typically (watt-for-watt), they are about half the weight and twice the price of linear models. Lighter weight can be a considerable advantage (unless you think overloaded vehicles or back injuries have some peculiar merit - I suppose you could use the money you saved to pay the fine or join BUPA). There is little to distinguish the two types in terms of audio performance. However, few manufacturers produce switched-mode models with output power of more than 1kW or so (into 4Ω) per channel, so if you need more power you will find there is less choice.
The main thing you require is POWER. If in doubt, go for more power than you think you will need. If the manufacturer of your loudspeakers recommends that you use amplifiers within a range, go for one as high in that range as you can afford: an under-powered amplifier cannot produce maximum available output from the loudspeakers without distortion. Some manufacturers publish guidelines on which of their amplifiers will suit a specific power requirement (see e.g. the QSC Amplifier Selector).
For the best ‘hi-fi’ performance from your amplifier/speaker combination you should use an amplifier matched to the peak capacity of your loudspeakers, and use a limiter to limit the RMS voltage to their continuous average capacity. This will ensure that instantaneous peaks will be handled cleanly, while continuous power levels do not exceed the speaker's continuous average power-handling capability. In practice, however, an amplifier that produces an average output of between 1.5 and 2 times the speaker's continuous capacity will usually be quite adequate. Power in excess of the speaker's peak capacity has no useful purpose.
You also need to consider the load your amplifiers are required to drive (the lower the impedance, the greater the load). Most amplifiers are designed to drive loads of 4Ω (and will driver higher impedances, but not lower). A speaker with a nominal impedance of 4Ω will usually have a lower impedance in part of its frequency range†. For this reason, driving multiple speaker arrays with a combined nominal impedance lower than 4Ω is not recommended, even if the amplifier's technical specifications claim it will drive 2Ω. At best, the amplifier's damping factor will be reduced (see below). At worst, the amplifier will overheat and shut down or burn out during a show. Avoiding low load impedance is most important at low frequencies (where power demands are highest, and where loudspeaker cone mass - requiring tighter damping control - is greatest). The higher the load impedance, the greater the damping factor (see below). Generally, this means that audio quality will deteriorate with lower impedance loads: in ‘hi-fi’ terms, your amplifiers will probably perform at their best when driving 8Ω loads.
†IEC specifications - if these are stated for your loudspeakers - stipulate that minimum impedance may not be less than 80% of the nominal impedance, but this may still mean that two 4Ω loudspeakers in parallel can present a load of 1.6Ω. The safest measurement to use, if you don't care what it sounds like as long as the room is warm, is the loudspeaker's DC resistance, as its impedance will not be lower than this regardless of frequency.
Specifications often given (and which you may require) are:
The power (in Watts) that the amplifier can produce into a stated load. Usually this will be described in one of three ways:
The DC equation translates in an AC circuit to
where P is the power in Watts, V is the voltage in Volts, and R is the resistance or impedance in Ohms). The result is average power, not RMS power!
Nevertheless, ‘RMS’ is used by some manufacturers to describe the average amount of power an amplifier can produce. Where this derives from a test, the test actually measures voltage, not power, and the power value given is (probably) an average derived from the RMS voltage.
As RMS power is not a valid technical term, the value given may not have been obtained by any valid technical method (and without further details you can't assume it has any validity at all).
The only useful power is undistorted power. The important point in both FTC and EIA ratings is that they state the amount of Total Harmonic Distortion (THD) at the amplifier's claimed average output power. Harmonic distortion is what happens when an amplifier approaches clipping. THD of less than 1% is generally considered to be inaudible. THD higher than 1% usually means the output signal's peak-to-peak voltage is approaching (and may be starting to exceed) the amplifier's power-rail voltage. The amplifier can't accurately reproduce anything requiring an output greater than its power-rail voltage, so the peaks of the signal are clipped off (this is what ‘clipping’ means). Driven to hard clipping, an amplifier can typically deliver double its undistorted power, as well as brief peaks of up to four times its undistorted power. Without giving THD information, the manufacturer of an amplifier could truthfully claim its peak output power was 1,000 Watts. However, the same amplifier might be rated at only 200 or 250 Watts using FTC or EIA definitions. Where FTC or EIA ratings are given, it means an amplifier's capability can be realistically compared with others that use the same standard. If the manufacturer's specifications don't tell you what standard is being applied, they don't tell you what you need to know.
A good amplifier will produce its rated power into 4Ω with less than 0.1% THD (in higher-end amplifiers, less than 0.05% THD at full power is commonplace). While many manufacturers claim their amplifiers will handle a 2Ω load (and may publish 2Ω power ratings), amplifiers NEVER work at their best with very low impedances (see FTC ratings, above). It may nevertheless be worth getting an amplifier with 2Ω capability, as it is designed to withstand higher current demands, and so will more comfortably manage 4Ω nominal loads at high power without distortion or overheating. Use it to drive 2Ω nominal loads, however, and the best you can ever expect is that it will survive the abuse.
The difference between the rated power and audible clipping. Usually this is not very much (3dB at most).
Ideally the relationship between input voltage and output voltage should be constant, regardless of the input frequency. Most modern amplifiers produce an almost ruler-flat response between 20Hz and 20kHz (often the range is much greater). However, this measurement is normally taken from amplifiers driving a dummy (purely resistive) load. The frequency response is not usually quite so flat when the amplifier is driving a reactive load (such as a loudspeaker, in which impedance varies with frequency), which is why different makes and types of amplifier often sound different. Listening to an amplifier through the speakers you intend to use can be useful.
This is usually given as a number (e.g. 200, or >200), and represents the ratio between the load impedance and the amplifier's output impedance (damping factor = load impedance ÷ output impedance). The damping factor will therefore be greater where the load impedance is higher, so - whatever amplifier you are considering - the damping factor will be twice as high using 8Ω speakers as it would be using 4Ω speakers.
In theory, damping factor determines the ability of an amplifier to limit unwanted movement of a speaker cone. Ideally, the cone's movement should correspond exactly with the signal waveform, and should cease abruptly when no signal voltage is present. In reality the loudspeaker cone has both mass and resonance, and may oscillate after controlling voltages have ceased (booming or ringing may be symptoms of this, but can also arise from other causes).
As well as being a motor, a loudspeaker is a generator: movement of the coil induces a small voltage across the loudspeaker's terminals. A short-circuit across the terminals - placing a very high load on the generator - makes the cone and coil much harder to move (it has the effect of damping cone motion), so an amplifier with a very low output impedance - a virtual short-circuit - helps to prevent the cone from moving when no signal voltage is applied. You can demonstrate this effect for yourself with any loudspeaker driver by moving the cone gently with the terminals disconnected, then shorting them together and comparing the ease of movement.
In practice, the impedance across loudspeaker terminals includes cable resistance (as well as any components in passive crossovers), which limits the possible advantage of low amplifier impedance: with an 8Ω driver a 10-metre length of 2.5mm cable is equivalent to a damping factor of approximately 117. Adding this cable resistance to an amplifer with a damping factor of 200 will yield an effective damping factor of only 74, and even an amplifier with a damping factor of 1,000 or more will only yield an effective damping factor of a little over 100 when it includes this cable. Also, loudspeaker impedances can vary considerably (at some frequencies being much higher - and at the lowest frequencies lower - than the nominal impedance). This means that the true damping factor is affected by cable gauge and length, and will vary from speaker to speaker and with frequency. If an amplifier's damping factor is stated, the figure given will only apply to its nominal load, and assumes a cable of zero resistance.
Low impedance across loudspeaker terminals is not the only form of damping in operation: the stiffness of the cone and suspension and the acoustic loading provided by the cabinet will also have a damping effect. The importance of an amplifier's damping factor is therefore qualified by other factors (including driver selection and cabinet design). Nevertheless, higher damping factors (i.e. lower output impedances) are generally advantageous, especially with larger ported cabinets - like typical front-loaded bass cabinets - where acoustic damping may be less effective. For bass loudspeakers a damping factor of at least 200 is generally desirable. However, as already shown, the effect of cable resistance on an amplifier's actual damping ability is substantial (and usually more significant), so use cables with large conductor area (2.5mm or greater), and keep them as short as possible: note that an amplifier with a damping factor of 100 would provide an effective damping factor of 75 - marginally higher than in the earlier example - using 10m of 4mm cable instead of 2.5mm.
Usually expressed in volts per microsecond (V/µs) this describes the rate at which output voltage can change. To reproduce a square wave accurately, the output voltage must be able to swing instantly from fully negative to fully positive. However, audio signals do not usually consist of square waves (and indeed, any frequency above 20 kHz isn't generally considered to be audible), so all an amplifier needs to achieve in practice is to get its output voltage at full power to change faster than a 20 kHz sine wave. In an amplifier producing 1,000 Watts into 8 Ohms (e.g. QSC RMX5050 or equivalent) this equates to a slew-rate of less than 16V/µs. Even the cheapest budget amplifiers have a higher slew-rate than this (typically at least 30V/µs, and almost never less than 20V/µs), so ‘low’ slew-rates are unlikely to have much effect on audio quality. Conversely, if an amplifier's slew-rate is high it may respond to sudden big transient signals (the sort of signals you get from dropping a mic or turning the mixer on or off) faster than protective limiters can react, so very high slew-rates are not necessarily advantageous.
How much noise the amplifier contributes to the system's output. This is usually given in decibels, and represents the difference between the noise floor and the amplifier's rated output. Most amplifiers are noiseless compared to other components in the signal path (anything over about 80dB is not likely to cause a problem, and even budget amplifiers typically have much higher margins). The physical noise made by some cooling fans can be a problem in very quiet environments.
The ratio (sometimes given as a factor - e.g. 40× - but more usually given in decibels) between input and output voltage. This can normally be determined from (and will correspond fairly closely with) the amplifier's sensitivity and rated power*, but may be selectable on some amplifiers. Note that output power is determined by the amplifier's power supply, not by the amount of gain: it is possible for an amplifier with low power output to have high voltage gain (in which case modest input signals will drive it to clipping), or for an amplifier capable of very high output power to have relatively low gain (so a larger input signal is needed to produce full power).
*Power is determined by the formula P=(V2)/R, so that - for example - an amplifier delivering 400 Watts into 4Ω gives us:
400 = V2/4
V2 = 400 × 4 = 1,600
V = √1,600 = 40.
So the output voltage is 40 volts.
If the sensitivity is +4dBu (1.23 volts) then the gain (output voltage ÷ input voltage) is 40 ÷ 1.23, or 32.5 times, or approximately 30dB.
As this calculation shows, gain, sensitivity and output power are interrelated, and the third can be calculated from either of the other two.
The level of input - this may be stated in decibels (usually dBu), volts, or millivolts - that will yield the rated output. This may also be selectable. Common values are:
Note that lower sensitivity values will increase the level of output for any given input.
Class refers to the power supply and output-stage configuration. The most common varieties in audio amplifiers are:
Class A. The output signal is driven by a single (ground and positive) D.C. power supply. With no input signal the output-stage transistors are half-on, so the amplifier is drawing current and creating heat when no signal is present (in fact, it draws most current when there is no signal). This is a more common configuration in hi-fi amplifiers, and has a theoretical advantage in signal reproduction in that a single set of transistors is used (so there isn't any distortion as the output signal swings from plus to minus). If you find a Class A amplifier in a PA system it is probably antique.
Class B. The output signal is driven by a split (ground, plus and minus) D.C. power supply. Paired transistors are used, with one half of the pair driving the positive half of the output signal, and the other driving the negative half. With no input signal the output-stage transistors are off (so the amplifier draws very little current when no signal is present). There is a risk of distortion as the output signal crosses zero (as the signal is passed from the transistors handling the positive half of the signal to those handling the negative half). This kind of distortion is known as ‘crossover distortion’ (which has nothing to do with crossovers). Class B amplifiers are not usually seen in audio circuits, but are included here to explain Class AB (see next paragraph).
Class AB. Class AB amplifiers can be thought of as Class B amplifiers with a small overlap in the middle. This means that each of the paired transistors is passing a small amount of current at idle, but this is much less than in Class A amplifiers. However, the overlap eliminates the crossover distortion associated with Class B amplifiers. Most amplifiers in PA systems are Class AB or variations on class AB designs (e.g. see next paragraph).
Class H. Class H amplifiers are similar in design to Class AB, but use more stages in the power supply. Low-level signals are dealt with in the same way as in class AB amplifiers, but higher voltage power supply rails deal with bigger signals. Two-stage (low and higher voltage) power supplies are the most common, but some class H designs incorporate further stages.
Primarily developed for other applications, Class D amplifiers are also becoming more common in audio equipment. These use switching amplifiers (similar in operation to switch-mode power supply circuits) which exhibit greater efficiency than the other types mentioned above.
If an amplifier you want to buy is Class AB or Class H, don't worry (and if it is Class D it will usually also be OK as long as it has been designed for audio use). If it isn't, don't buy it unless you understand why it isn't.
All electrical equipment uses current. Most audio devices don't use much (a few milliamps - a few thousandths of an amp - in many cases), so you can generally run a bank of processors and effects units without overloading a single 13A power point.
Power amplifiers - the things that do the physical work - are the components in the audio chain that use the most current, although their current demands are usually less than you might expect. All the same, it is important to ensure that any load on the mains supply is less than the supply's rated current capacity. Where several amplifiers are used at the same time (e.g. tops, subs and monitors), their total current consumption should not exceed 13A if they are supplied through a single power point, or the current rating of the mains circuit - often 20A in domestic or ‘village hall’ circuits - if they are supplied through multiple sockets on the same ring. Larger touring systems usually use dedicated supplies of 32A or more.
Power is the product of voltage and current: Power (P) = Voltage (V) × Current (I). In theory, therefore, your 3kW system needs:
P (3,000) = V (230) × I
I = 3,000 ÷ 230 ≈ 13A.
In practice, however, amplifiers driving a musical signal never run at full power. If you tried to get them to run at full power with a musical signal, the signal would be badly distorted, and any amplifier with thermal or current overload protection would shut down within a few minutes. Any amplifier without protection would fail fairly soon.
Most amplifier specifications will give information about power or current consumption under some or all of the following conditions:
If you are designing a PA system, you should generally base your expected total current consumption on 1/3 power (and at least 1/8 power).
Note that power or current consumption at 1/8 power is greater than 1/8 of the amplifier's output power. Some of the power drawn from the mains is converted into output power, and the rest of it is converted into heat.
Useful extra features include: