Websites we've worked on: Dr Andy Moorhouse on piano ... BBC radio docs on sound:- presented by Prof Trevor Cox
A Guide to Loudspeaker Specifications
This article aims to discuss loudspeaker construction, principle workings, sensitivity, frequency response, power rating and directivity.
A loudspeaker is a type of transducer, i.e. is a device for transforming one type of wave, motion, signal, excitation or oscillation into another. The follow shows various types of transducer including microphones, guitar pickups and accelerometers.
Loudspeakers use electro-mechano-acoustic transduction, i.e. they convert electrical energy into mechanical energy, which in turn is converted into acoustic energy. Conversely microphones convert acoustic energy into mechanical energy and hence electrical energy. Curiously the mechanisms are interchangeable – in other words you can sought of use a microphone as a loudspeaker and vice versa, though I wouldn’t recommend it if you don’t want to ruin expensive equipment.
So how does all this work? Well it stems from initial experiments by Victorian physicists such as Michael Faraday who demonstrated and utilised the curious relationships between electrical current, magnetic fields and mechanical force.
Force on a Conductor
A conductor of length l carrying a current i at angle to a magnetic field of constant flux density B experiences a force F where
B: Tesla (T) (Webers/m2), i Amperes, l Metres
Note that the force, current and flux are at right angles to each other. This can be given by "Flemming's left hand rule" where - the thumb(Force), index (flux density field) and ring (current) fingers of the left hand show the relationship between the vectors.
Electrical energy passed from an amplifier is converted into kinetic energy as the loudspeaker cone moves in and out via electro-mechano-transduction.
The voice coil is attached to the cone, which is in turn held in a steel basket and fixed to a rigid support. Flux is generated by a magnet and guided by steel plates and a pole piece effectively forming a magnetic circuit. As a time varying current is passed through the coil the interaction with the flux generated by the magnetic causes a force. Generally a spider is needed to centre the voice coil in the gap, these are generally made of steam pressed cotton. The coil and magnet are also kept clean by a protective dust cap. The cone is generally moulded from steam pressed paper though polypropylene, kevlar and even carbon fibre cones can be used. The resulting piston-like action of the cone causes changes in the air pressure in front of it, producing sound energy. A well-designed speaker may only be around 5% efficient, in other words, 20 watts of electrical power from an amplifier may only result in 1 watt of sound power.
Most of the functions of an enclosure are NOT related to acoustics:
The above is a Marshal Stack – many were smashed up by Jimmy Hendrix.
Spherical design reduces high frequency distortion though also serves to distinguish the brand (www.minipod.com).
The acoustical functions of a loudspeaker include:
low frequency: control back load
isolate rear radiation
high frequency (to a limited extent): define baffling
An un-enclosed loudspeaker radiates sound as an acoustic "dipole". This gives rise to a poor l.f. response (since sound from the back of the diaphragm cancels sound from the front) and highly directional radiation.
To avoid these problems, we can mount the loudspeaker in an infinite baffle , in which case it radiates into the "half space" in front of the baffle as a monopole. Unfortunately, this is sometimes impractical (an INFINITE baffle would not fit into a finite listening room).
To deal with this impracticality, the infinite baffle is "folded" around the back of the loudspeaker, forming an ‘infinite baffle’ enclosure (a fancy name for the "closed box").
We specify sensitivity of a loudspeaker in terms of dB SPL for 2.83 V input.The sound pressure level is measured on-axis in anechoic conditions at a distance of 1 metre from the loudspeaker.
2.83Volts corresponds to the voltage across a standard 8 speaker driven at 1Watt.
So if we measure 0.2 Pa at one volt what is the Loudspeakers Sensitivity?
Change in SPL if 2.83volts applied is =9.04dB
i.e. 89.04 dB @ 1 meter with 2.83V input
The graph below shows typical sensitivities of a wide range of modern speakers
80dB/2.83V@1m flat panel speakers
95dB/2.83V@1m professional monitors
Using peak response or measuring with wide band noise can give misleadingly high sensitivities. Measuring with
B-weighting gives more accurate picture taking into account subjective loudness.
Note sensitivity is not necessarily a guide to how good a speaker is. For instance a PA speaker will have a sensitivity of around 115dB so sacrificing sound quality for high output. Conversely a good monitor may be designed to ensure a flat frequency response with components that also reduce sensitivity.
Older system specified efficiency in the rather unusual units of dB SPL/W/m where the power is the electrical input power. The older system is still quite popular so it’s worth knowing about.
A change in input POWER of Pratio would cause an increase in radiated sound pressure level of 10 log(Pratio)
because velocity and sound pressure are proportional to Voltage, whereas power is proportional to voltage squared).
This means for instance if a speaker with a sensitivity of 100 dB/W/m is powered by 50W, the resulting SPL is
100+10*LOG(50/1) = 100 + 17 = 117dB/W/m
Although still commonly used efficiency is not as useful as voltage sensitivity since input power is difficult to measure given input impedance is frequency dependant, so for instance a lower frequency will have lower impedance and suck more power, hence be less efficient. Also modern solid state amps are voltage sources supplying current on demand so power is less important than dB for volts.
The resulting SPL for a speaker powered across a range (sweep) of narrow band frequencies is the frequency response. Frequency response is often quoted as the maximum deviation in SPL within a specified range. For example…
Frequency response 3dB from 42Hz - 20kHz
Often manufactures will also provide a plot of frequency response…
dB against frequency.
Ideally a speaker will have a flat response across a wide range. Usually the more ragged the response to poorer the speaker.
Clearly different types of speaker will have different responses to suit their application for example
- Studio Monitors require a flat response
- PA Speakers require a high sensitivity at the expense of frequency response
- Sub woofers require an improved bass response at the expense of the higher frequencies
Improving Bass Response
The bass response of a loudspeaker can be improved by using back radiation. Unfortunately, the front and back radiation is in anti-phase - we need a "phase inverter" before it is possible to add the front and back radiation constructively.
Such a strategy is achieved by the "phase inverter" family of loudspeaker enclosures, which couple front and back radiation from the low frequency unit(s) through an acoustic phase inverting network.
The phase inverter family includes
- the transmission line enclosure
- the bass reflex enclosure
- the auxiliary bass radiator
The most important member of the phase inverter family is the:
Bass Reflex Enclosure
In a bass reflex enclosure, the loudspeaker internally excites a Helmholtz resonator. This is an acoustic network which has the amplitude and phase response of a second order resonant system (which gives a phase inversion above resonance).
The bass reflex enclosure is a closed box enclosure with a "port" in one of the enclosure walls. The port is a tube, open at one end to the inside volume of the enclosure and open at the other end to the listening space. Air in the port moves as a single "lumped mass" which bounces up and down on the spring formed by the volume of air trapped in the enclosure, giving the system a second order resonant behaviour.
From resonant frequency upwards, the port radiation and the speaker radiation are in phase (so they add constructively giving more bass). Below resonance, the port and speaker radiation are out of phase (so they add destructively giving less bass than a closed box) At the resonant frequency the port response much greater than the 'speaker response (because the back load presented to the loudspeaker is maximum at the resonance of the port.
The frequency responses of the speaker and port in a bass reflex enclosure are shown individually in the figure below....
Electric Power that can be input before unacceptable distortion
E.g. an 8W speaker rated at 35W has sensitivity of 85dB/2.83V @ 1m
Hence at 1W the speaker produces 85dB
So at 35W, maximum SPL is
At low frequencies sound propagation is omni directional (spreads out in all directions) for high frequencies sound propagation becomes much more directional. This holds true for speakers where the placement of sub woofers is less important than higher frequency monitors.
This can be shown for a loudspeaker by plotting a curve of equal SPL hence representing the propagating wave front.
As the frequency increases the propagating wave front splits into lobes. The forward lobe carries most of the sound energy.
This has an important bearing on stereo sound systems and mixing as the listener is less likely to care where low frequency sounds come relative to high frequency sounds.