First a couple of points of clarification:
Amplifiers and speakers need to be considered together because the behaviour and performance of each of these components of the system is heavily dependent on the other. In any particular system, the selection of these two components therefore needs to be well co-ordinated.
There are two basic approaches to achieving this co-ordination:
Speakers can be categorised into the following six types, according to their use. Most of these may be either passive or powered types (see the section above).
Audio frequencies are generally considered to fall into the following ranges (or "bands"):
The boundary frequencies of the ranges are somewhat arbitary, and you will find other similar figures quoted. Sometimes the mid-range band is divided into upper mid and lower mid bands.
The range of frequencies that a speaker system must handle depends very much on the application. Speakers intended only for announcements in public buildings, for example, may only require a range of 200 Hz to 10 kHz, whereas good quality reproduction of music requires at least 80 Hz to 15 kHz. Good response down into the 'sub-bass' region will allow the music to be felt as well as heard, and good response up towards the upper limit of human hearing (20 kHz) will add increased clarity and crispness to the sound.
There are two basic approaches to handling a wide audio frequency range:
It is important to understand that the distribution of audio power is not uniform across the frequency spectrum − most of the power falls into the bass and lower-mid ranges. As you look at higher and higher frequencies, less and less power is present. This has important consequences in the design of speaker and amplifier systems − especially when separate speakers are used for the different frequency bands. For example, a 5 kW 3-band system might be made up as follows:
Some caution is needed in interpreting the frequency range figures quoted by speaker manufactures and suppliers. The norm is to quote the frequency at which the sensitivity has reduced to 3 dB below frequencies around the centre of the band for which it is intended (i.e. the power level of the reproduced sound has halved). However, sometimes "usable" frequencies are quoted − which may mean a reduction of as much as 10 dB (i.e. one tenth of the sound power) at the quoted frequency, compared with the speaker's mid-band sensitivity.
Firstly, it tells you only how much electrical power the speaker is able to accommodate without sustaining physical damage and/or causing undue distortion of the sound. This gives only a very rough guide as to the sound level (volume) that it is capable of producing, because the sound level obtained from a given number of watts of power depends upon the sensitivity of the speaker, which varies significantly between different models.
Secondly, because of the fluctuating power level of most sound signals, the proper specification of power handling capacity requires a standardised way of measuring the power level that a particular speaker can accommodate. Unfortunately there are several different methods of measurement in use, so when looking at the power rating quoted for speakers you need to be certain of which method is being used, and compare only "like with like". The most common methods are:
These different measurement methods must also be taken into account when matching the power rating of a speaker with the output power rating of the amplifier that is going to drive it. (When looking at amplifier rating plates, be careful not to confuse the mains input power requirement with the audio output power rating.) Assuming that the same power rating methods are being used for both the speaker and the amplifier, one possible approach is to ensure that the speaker power handling capacity is amply adequate to produce the highest sound level required (taking into account the speaker's sensitivity), and then choose an amplifier whose output power rating into the overall speaker impedance to be driven (see below) is around 30-50% higher than the total power handling capacity of the speakers to be connected to it. (This is a very basic approach − a more thorough one is detailed in Amplifier and Speaker Selection for required SPL, later on this page.)
For example, if two 100 W RMS rated speakers having a combined impedance of 4 ohms are to be connected to a mono amplifier, then the amplifier output power rating into 4 ohms should be around 260 to 300 W RMS if the maximum useful sound level is to be safely obtained from the speakers. The reason for the 30-50% margin is that the short-term power handling capacity of a speaker is well in excess of its RMS value, whereas that of a high-quality amplifier is usually only marginally above its RMS value. So to be able to utilise the available capacity of the speaker fully, without risk of speaker damage due to the amplifier being driven into overload during short-term peaks in the sound level, it is necessary to use an amplifier with an RMS output power capability that is considerably greater than the total RMS capacity of the speakers that are to be connected to it.
The down-side of this is that it is possible to slightly over-drive the speakers, but this should be a rare occurrence because if the speakers are correctly rated then over-driving them would produce a sound level higher than what you required. In any case, most speakers will tolerate a degree of short-term over-driving without difficulty. If this is a concern, you can either take care to keep a close eye on the amplifier power meters, or you can start by choosing speakers with a rating somewhat higher than you expect to need (but this may increase the cost significantly). Alternatively you can buy powered speakers, in which case the manufacturer has made these difficult decisions for you!
The sensitivity of a speaker is a measure of the sound pressure level (SPL) that it will produce for a given amount of electrical input power (or voltage − see below). It is usually quoted in dB SPL at an input power of one watt, measured at a distance of one metre directly in front of the speaker, for a given type of programme signal. The type of signal is important because the sensitivity of a speaker varies with frequency. A typical sensitivity figure would be 100 dB SPL @ 1W @ 1m, which corresponds to an efficiency of approximately 10%.
The sensitivity figure can be used 'in reverse' to estimate the speaker input power required to achieve a particular SPL at a given distance, using the inverse square law. For example, suppose that we require a maximum SPL of 112 dB at a distance of 8 metres from a speaker having the sensitivity figure quoted above. From the inverse square law, the SPL will be 118 dB at 4 metres, 124 dB at 2 metres and 130 dB at 1 metre. Therefore, the maximum power input required to the speaker is 30 dB greater than 1 W, which is 1 kW. (For a more comprehensive explanation and example, see Amplifier and Speaker Selection for required SPL later on this page.) However, this is only an estimate because many complicating factors such as room acoustics and grazing have not been taken into account. Of course, if the particular speaker being used is unable to handle this power level (see Power Ratings above), then several can be used together − but the maximum power input to them will still need to total 1 kW. (Refer to the Impedance section below for information regarding the connection of several speakers to a single amplifier.)
In practice, an amplifier supplies a controlled voltage, not a controlled power, and the power taken by the speaker depends upon its impedance, which varies with frequency. Therefore it is becoming increasingly popular to specify speaker sensitivities in voltage-related terms. For nominally 8 ohm speakers, the value is specified in dB SPL @ 2.83 V @ 1m, 2.83 volts being the voltage required for a power of 1 W in 8 ohms. (For 4 ohm speakers the voltage required is 2 V.) So, the sensitivity value specified in this new way is effectively equivalent to a value specified in the old power-related way.
For a general definition of impedance, see its glossary entry. The impedance of a speaker is important for three main reasons:
It should be noted that the impedance value quoted for a speaker is a nominal value, and in practice the value varies with frequency. The minimum value is often of particular interest; this is likely to be around three-quarters of the nominal value, so an 8 ohm speaker is likely to have a minimum impedance of around 6 ohms.
A special case of speaker impedance arises in the case of so-called '100 volt line' speakers, which have a much higher impedance than conventional speakers. This allows many of them, located over a large area (such as throughout a public building), to be connected to a common amplifier using moderate gauge cable. The amplifier used must be one that has a 100 volt line output. These speakers are generally of low power rating (5 to 50 watts RMS each), and are usually equipped with a means to adjust the power level drawn from the 100 volt line by each speaker. As the usual application of this arrangement is for speech announcements and/or 'musac', the sound quality of these units is often not high − particularly as regards bass response.
Bridging is a technique to improve the matching between the impedance of a speaker (or the overall impedance of several interconnected speakers) and the optimum load impedance of the available power amplifiers, so as to increase the maximum amount of power that the amplifiers can provide to that speaker(s). It is most useful when it is required to use more of the power-handling capacity of the speaker(s), or more of the power-providing capability of the amplifiers, than could be achieved with a simple non-bridged connection of the speaker(s) to the amplifier.
In a bridged arrangement, two identical amplifier channels are driven by signals of opposite polarity, and the speakers are connected between the "hot" (i.e. non-earthy) terminals of their output connections. This is properly referred to as a 'bridge-tied load' (BTL) connection − or, less formally, as 'bridging the amplifiers' − and effectively doubles the voltage available to the speakers. Some models of amplifier have a switch to select this mode of operation; it is sometimes additionally necessary to use a different speaker output connector of the amplifier, or to 'manually' arrange the appropriate connections to its output terminals.
WARNING: The output voltage of bridged high-power amplifiers can be high enough to cause electric shock. The speaker cable used must be suitable for the voltage and current supplied by the amplifier. Speakers and/or amplifiers can be seriously damaged by inappropriate use of bridging, or by making incorrect connections.
The following lists provide some basic guidance regarding when it might be appropriate to use bridging, but they are not exhaustive.
Reasons NOT to bridge
Reasons to bridge
None of the conditions in the above list apply, and:
Bridging is possible only when each channel of the amplifier is capable of driving an impedance of one-half of the overall impedance of the connected speaker load. It should only be employed if approved by the amplifier manufacturer and if the speakers are suitably rated, otherwise serious damage to the amplifier or speakers may occur.
When a speaker (or a set of interconnected speakers) is connected to a single amplifier, the full power rating of the amplifier can be delivered to the speakers only if the (combined) impedance of the speakers is that which the amplifier is optimised to drive. For example, if the (overall) impedance of the speakers were four times the optimal value for an amplifier, then it would only be possible to obtain about a quarter of the amplifier's rated output. This is because the amplifier would be unable to provide sufficient voltage to force through these speakers the maximum amount of current that the amplifier is capable of supplying. So, assuming that the speakers are suitable for handling the full rated power of the amplifier, the speakers would be very under-utilised.
However, by using two identical amplifier channels driven by signals of opposite polarity, and by connecting the speakers between the "hot" (i.e. non-earthy) terminals of the outputs of the two channels, the voltage available to the speakers is doubled. Since (for a given impedance) the power obtained is proportional to the square of the applied voltage, this arrangement increases the power supplied to the speakers by a factor of four − which would nicely compensate for the problem with our example speakers whose impedance is four times too high. In this example, we have obtained the power rating of just one of the amplifier channels − but at the cost of using a second channel (typically, both halves of a 2-channel unit).
Since, in this example, we needed two amplifier channels to get the rated power of just one of them, each channel was clearly only working at half-power. The speakers were fully powered but the channels were only half-loaded, which is still not a satisfactory situation. So, we can connect a second speaker (or set of speakers), of the same impedance and power rating as the first, to our bridged channels − and double our total output power. Now, both channels and both (sets of) speakers are fully utilised.
From a technical perspective, the reasoning behind this is that, in a bridged configuration, the load impedance "seen" by each channel is half of the actual connected load. In our example the load was four times too large, so even in the bridged arrangement each channel still "saw" twice its optimum load impedance. Therefore, only half of the available current was being taken and so only half the power was being supplied. By connecting a second set of speakers, the total actual load became twice (instead of four times) the optimal load for a single channel. So now, in the bridged configuration, each channel sees its optimal load and therefore is able to deliver its full rated power output.
To recap using some example figures, suppose we have a pair of 16 ohm speakers, each suitable for being driven by a 100 W amplifier, but our pair of 100 W amplifier channels are both optimised for connection of 4 ohm speakers. If we simply connect one speaker to each channel, we can only get a maximum of 25 W (or perhaps a little more) from each channel − a total of 50 W. But if we bridge the channels and connect both of the speakers (in parallel, as usual) to the bridged output, then we have connected an 8 ohm load in a way that causes each channel to see half of it − the amplifiers' optimal value of 4 ohms. So each channel is able to deliver its rated 100 W and we get a total of 200 W (four times what we got without bridging!) shared equally between the two speakers. (Put another way, we have effectively created a single 200 W amplifier with an output optimised for 8 ohms − and connected 8 ohms to it.)
Note, however, that bridging is not a way to get more out of an amplifier than it was designed for. It's just a way of summing the power capability of two identical amplifier channels in a way that effectively creates a single amplifier with an optimum load impedance which is double that of each individual channel.
Thought we were finished? Well let's just briefly look at this another way. Suppose we have a single really powerful speaker (or a set of interconnected speakers that for some reason it's just not practical to split up). We don't have an amplifier powerful enough to drive this speaker to capacity, but we do have a pair of identical channels of a lower power rating. Can we use bridging to help in this situation? As we saw above, it's all a question of impedances. The first thing to check is the minimum acceptable load impedance for the two individual channels. For a bridged arrangement to be acceptable, the impedance of the connected speaker (or set of speakers) must be no less than twice this figure because, when bridged, each channel will "see" only a half of the actual connected load impedance. The next thing is to determine if there's enough advantage to be gained from bridging to make it worthwhile. To do this we compare how much power the channels can provide into this halved impedance (when both amplifiers are driven), as compared to how much can be provided into the full speaker impedance with just one of them driven. (For two-channel amplifiers in which a common power supply is used for both channels, the available output power with both channels driven can be significantly less than that available when using just one channel − check the manufacturer's specifications.)
To see how this approach works in practice, let's look at some example figures. Suppose we have a single 4 ohm speaker suitable for driving by an 800 W amplifier, but what we have is a "300+300 W" two-channel amplifier. First we check the amplifier specifications to see if it's channels can drive 2 ohm loads (half of 4). So far so good. Now let's say that further checks show that when both channels are driving 2 ohm loads (the bridged situation), the maximum output is 250 W per channel (a total of 500 W), whereas if we connect our 4 ohm load to just one channel of the amplifier (and don't use the other channel), the amplifier can deliver 350 W on that one channel. So, in this example, bridging has given us only an extra 150 W of drive power into our speaker, an increase of just 1.5 dB − possibly because this amplifier was optimised for 4 ohm loads. Unless you're desperate for that extra 1.5 dB, or bridging is very easy to do (see the next paragraph), it probably wouldn't be worth it.
Some two-channel amplifiers have a switch that configures the pair of channels into bridging mode, which is far more convenient than having to arrange for it yourself. But if you have to do it yourself, you will need a means of inverting the polarity of the signal fed to one of the two channels (easy if the amplifier accepts a true balanced input − just swap the 'hot' and 'cold' conductors of the input connection − pins 2 and 3 of an XLR). You will also need some specially wired cables to allow your speakers to be connected between the 'hot' connections of the outputs of the two channels. Beware though, not all amplifiers are suited to bridging − check the manufacturer's information to avoid possible damage to your expensive equipment!
To describe the various possible modes of internal operation, reference is sometimes made to the 'class' of a power amplifier. This is a design parameter, and rarely impacts significantly on how the amplifier is used within a system, but for completeness and technical interest this section gives a brief explanation of the most common classes encountered. Some of these are not relevant to PA applications, but are still listed.
As a starting point, it is necessary to understand that all power amplifiers have at least two 'output devices' (usually transistors of some kind, though valves are sometimes still encountered) − one device supplies the current to the speaker(s) on positive excursions of the waveform, and the other device supplies the current on the negative excursions. (High power amplifiers have several devices operating together to perform each of these two functions, but this has no bearing on the amplifier class.)
Simply put, the class of the amplifier is a description of how the two power output devices are utilised to supply the required output current, particularly how they are coordinated in their operation in order to achieve the crossing over of the output voltage waveform from positive to negative and vice versa. The amplifier class has a bearing on the linearity and the efficiency of the amplifier. 'Linearity' relates to distortion levels − good linearity means less distortion (and vice versa). 'Efficiency' relates to how much power is wasted as heat in the amplifier − high efficiency means that little power is wasted, meaning less mains power is required and that the amplifier can be much smaller and lighter, and requires less cooling.
All speakers require some kind of enclosure to house the drivers. Enclosures are usually either:
However, in the case of full-range, bass and mid-range speakers, the enclosure doesn't simply house the drivers − it has a significant effect on the sound produced. This is primarily because of the damping effect of the air trapped within the enclosure, which is often modified by carefully designed 'porting' arrangements − one or more holes which may be fitted with internal tuning tubes or ducts.
These are large, heavy units, usually with 12", 15" or 18" diameter cones. As a general rule, the larger the diameter the lower the frequencies that can be reproduced, and the heavier the magnet the greater the power handling capability.
It is important to understand that the operation of a bass driver is heavily influenced by the speaker enclosure − its dimensions, construction and porting arrangements. Therefore, when replacing faulty bass drivers, it is essential to fit the correct type − the low-frequency response or power-handling capability of a speaker cannot usually be improved simply by fitting a higher-spec driver.
HF drivers fall into the following categories:
Piezo drivers are capable of little power output in comparison with dynamic ones, and so are used only in low power equipment, generally less than 200 W overall system power.
Because the amount of power at high frequencies in an audio signal is much less than the overall power of the signal, HF drivers are used which have a power rating much lower than the bass drivers. Therefore, the HF drivers are more prone to damage by overload than are the lower-frequency drivers in the speaker system. Although some systems provide some measure of internal overload protection for the HF drivers (see below), in order to prevent the possibility of serious damage it is important to avoid:
The crossover is the component that separates the signal into a number of frequency ranges, usually two or three, so that each driver receives only the frequency ranges it is designed to handle. 'Protection' refers in this case to features of the speaker design that aim to avoid damaging power levels being applied to the speaker drivers. As an internal component of full-range speakers, there are three typical crossover arrangements:
Note that, unless stated otherwise, all the sound levels and power levels and ratings that we refer to here are 'continuous average' values (often referred to as 'RMS' values, although this is not strictly correct). Do not substitute 'programme power', 'music power' or 'peak power' values (see Power Ratings above). For brevity we will just use the term 'continuous' here rather than 'continuous average'.
To illustrate the principles clearly, we will initially consider a single full-range speaker connected to a single amplifier channel. In summary, the procedure is:
Example Selection Procedure
Multiple SpeakersWhen connecting more than one speaker to an amplifier channel, there are two additional considerations:
Each front-of-house speaker should be positioned pointing directly towards the part of the audience to be covered by that speaker.
An important aspect to consider when choosing speakers is the angle, both vertical and horizontal, over which the sound needs to be dispersed. This will depend entirely upon the placement of the speaker in relation to the audience, and must be considered in conjunction with the number of speakers to be used. In general, short throw units have a wide dispersion angle, whilst long throw units have a narrow one. When the dispersion angle required to cover the audience is very large, the use of trapezoidally-shaped speakers can be of considerable benefit, as several of these may be neatly positioned side by side, in an arc formation, to give a very wide angle of coverage.
Typical support methods for front-of-house speakers are:
Whichever method is used, safety is always an important consideration, as speakers are heavy and can cause serious injury if they fall.
Monitor speakers are usually floor-standing wedge-shaped units (floor monitors), or are small stand-mounting units. On large stages side-fill (or cross-fill) monitors are also used, and occasionally monitors are flown above the front edge of the stage, appropriately angled downwards to point towards the relevant performers.
All wedge-shaped monitors should be correctly angled up from the horizontal; that is, they should have an angle which projects the sound towards the performers's head when he or she is located at their usual distance from the monitor. This means that on a large stage where the distance from the monitor may be several metres, a sound projection angle of 30° from the horizontal may be appropriate, whereas when the monitor is very close an angle of 60° from the horizontal may be more suitable. To accommodate this, some monitor speaker enclosures are designed to enable placement at one of several different angles.
In general, monitor speakers for vocalists should be placed so as to minimise the risk of feedback. This means placing them such that the sound from the monitors reaches the microphone at the angle at which the microphone is least sensitive − its "null point". This angle depends upon the polar response of the microphone you are using. For a cardioid microphone, the null point is on the rear axis, so the monitor should point directly at the back of the microphone. For a super-cardioid type, there are null points at about 55°, measured from the rear axis towards the front, on each side. A single monitor can be placed pointing at this angle at just one side of the microphone, but the vocalist will usually be happier with two monitors, one on each side, so that they are able to hear the monitor sound equally with both ears. With a hyper-cardioid microphone, the null points are at about 70° on each side of the rear axis, and the monitors (again, ideally two of them) should be placed accordingly.
Monitor speakers for musicians should be placed close to the musician concerned and be pointed towards them. They should preferably point away from the audience and not point towards any microphones (unless they aim at the microphone's null-point − see above).
This page last updated 24-Sep-2018.