A STARTING POINT :
« Since I throw my ears into the high-fidelity world, I had to revise, while listening to all kinds of specialists, all the basics concepts I learned in my electronic classes. All these simples notions of Watt, impedance, resistance, etc… were linked with so many subjective values, so many misusage, that even the officials laboratories start to be lost.
So I wished, only to clean up my brain, to look back closer to all these fundamentals concepts and share with you all my discoveries.
In the generalist electronic and loudspeakers books, in the shops, the poor buyer faces so much contradictive notions about systems power. Let’s check back the usual terms used in our domain and give them their real value, and usage. »
Guy Le Cornec
Frequency, or pulse per second, is measured in hertz (Hz). The audible frequency bandwidth is, theoretically, for a human being from 20 to 20 000 Hz. The ear sensitivity, as much in level as in frequencies, diminishes with age.
The sound pressure level unit is dBspl. Its zero level is defined at the human audibility threshold:
2.10-5 Pascals. It uses logarithmic scale, as all decibel scales.
The dB(A) unit derivate from the dBspl scale, with a frequency-dependent weighting.
Comparative table of the noise levels in dB(A)
|ACOUSTIC PRESSURE Pascal (Pa)
|SOUND LEVEL PRESSURE
|NATURE OF THE NOISE
|< 2 .10-5
|Sailing ship, garden, cave
|Bedroom, low voice
|2 . 10-3
|Calm office or apartment calm voice
|6,3 . 10-3
|Noisy office or apartment calms, normal voice
|2 . 10-2
|Apartment with television, strong voice
|6,3 . 10-2
|2 . 10-1
|Train station, workshops
|6,3 . 10-1
|Lesions if 8H/day
|Press, motor bikes, machining
|Riveter, hammer rammer
|No conversation, club, concert
|Threshold of pain, disturb, deafness
The efficiency is the relationship between the generated energy and the energy received by the device. In our case, between the electrical power delivered by the amplifier and the acoustic power generated by the driver. We should notice that our hearing is not acoustic power sensitive, but acoustic pressure sensitive. In addition, at constant acoustic power, acoustic pressure increases if radiation angle decrease. So this data is hardly linked with subjective feelings.
The sensitivity is the sound pressure level delivered by the driver at one point of space, normally at 1 m on its normal axis, for a specified voltage (normally 2,83Vrms for a speaker of the normalized 8 ohms impedance). This measurement is usually given in half space radiation angle (2pi steroidal, or visually: half of a sphere).
The frequency response is the graphic curve of the driver sensitivity at each frequency. It’s normally measured from 20 to 20 000 Hz, for a 2,83V signal at 1m on driver axis. It should be smooth, progressive, without dips or bumps, neither steps. It’s an important characteristic of a loudspeaker.
Power. It’s the major misused term in our domain… There are two calculations of power which may interest us.
– The root main square power (in Wrms) delivered by an amplifier. The audio amplifiers are tension generator, then the power is calculated depending on the load in which the power is dissipated: lower is the load, more the amplifier is powerful, because it delivers more current (P = U x I). This is true until the power supply voltage drops, or the output stage components burn out suffering of overcurrent.
– The electric power handling of a driver, in Waes. It’s only the amount of power a driver may receive without being broken: either by ungluing the coil from its former, or by burning the wire enamel. It’s a thermal limit. A powerful driver is not better sounding, it’s only harder to break… a truck tire under 8,2 bars pressure is not more precise than a motorbike tire under 2 bars. This power unit is Watt AES, because the measurement conditions are precisely defined in a norm, the AES2-1984 (r2003) norm.
Old DIN and IEC norms are not used anymore. All the peak, pulse, music, average, peak music… power units are meaningless.
The maximum acoustic level, or dBspl Max, it’s only due to the volume of air moved at a specific frequency. This volume depends on the cone surface (Sd) and the coil maximum displacement Xmax of the driver. A large 8’’ full range speaker with 2mm Xmax will generate more acoustic pressure, so will be louder, than a 5’’ bass driver with a 5mm Xmax.
The displacement limit Xmax is the positive, or negative, ideally symmetric, displacement the coil may reach within field gap before the distortion, caused by the drop of Bl factor, reach 10%. The researchs of Dr Klippel shown it corresponds to 18% drop of Bl. Our Xmax are defined through this criteria, eventually leading to an Xmax greater than physical X, which is a simple substraction of coil height from field gap depth.
Loudspeaker distortion is due to its nonlinear behaviors: motor strength Bl, coil inductance Le, surround and spider compliance Cms are not perfects, even sometimes asymmetric, depending on the current (i) flowing thru the coil, and coil position (x) in the field gap. All this demands a careful design of each constitutive parts of a loudspeaker.
At equal conditions, distortion increase with coil elongation. Indeed, at equal elongation, a well-crafted driver distorts less.
A BIT FORWARD…
The full range concept. Nowadays majority of loudspeakers use a multi-way configuration, passive filters dispatch frequencies between different specialized drivers. These filters have, already in theory, many limitations, multiplied by the non-linearity of actuals components.
In History the first driver was used as a full-range: without filter, in despite it was rather a midrange driver, as it was a voice transmitter.
As broadcast develop, loudspeakers improved their neutrality and broaden their bandwidth.
Only later, when high acoustic level demand increased, loudspeakers became frequency band dedicated. Technical solution constrains by a full range design, are not responding to Hell Fest live show needs.
In domestic conditions, a high quality full range driver will benefit from the absence of filter.
It’s a know-how Supravox has kept, transmit and perpetuate.
Full paper driver cone. At Supravox we only use full paper cone in our full-range drivers, from the coil bond to the surround. So it was since the first electro-dynamic loudspeaker patent.
By its homogeneous, but random, orientation of cellulosic fibers within its structure, the paper cone brings great rigidity/damping ratio through all the bandwidth. The accurate design of a cone, its profile, its thickness, will define its frequency response: broad and clean. The paper pulp formula will define the loudspeaker neutrality. Specific treatment of the paper surround allows us to control frequency response and sound.
We use the same exponential cone, from the same molds, and the same paper pulp since M. Liebert development around full range drivers during the 40’s.
Our drivers may use dust cap or bicone. Extremely rigid dust cap, treated for this purpose, directly glued onto the coil former will extend the driver frequency response beyond 12-15kHz. Bicone, through its shape and surface acts as a second driver cone, will already regenerate much energy from 4 kHz.
The choice is made during the careful design and optimization of each loudspeaker.
Which magnet ? A magnet, whatever the technology, is first of all a permanent magnetic energy source. Anyway, as any earth material, nothing is perfect. Each magnet is defined by four main behavior:
– Its magnetic remanence, its capacity to maintain persistant field strength.
– Its coercivity, which is its capacity to resist external demagnetizing forces.
– Its maximal usage temperature.
– Its temperature rate, or the reversible field strength loss with temperature increase.
Sadly none of the magnets is perfect in all these aspects, and cost is also a criteria.
Ferrite magnet is the cheapest, with good remanence, high coercivity, comfortable usage temperature of 225°C, but the worst temperature loss.
Neodymium is expensive, with very high remanence and coercivity, good temperature coefficient, but low usage temperature of 80°C for standards grade.
Alnico, one of the oldest magnet formula with its Ticonal cousin, offer high remanence, almost perfect temperature ratio, very high usage temperature of 450°C, but low coercivity for old and usual grades. It’s a very good, and expensive, magnet which needs careful design of the magnet assembly to exploit its qualities.
Another magnetic energy source is field coil. In this configuration the magnet is replaced by a large coil, which create a magnetic field proportional to the current flowing through it, it’s called: induction. If the power supply is a current source, then the field in the field gap is invariant.
Acoustic level, efficiency and power. We seen that the acoustic pressure is only function of driver cone displacement multiplied by its surface. We also noticed that a high sensitivity loudspeaker needs less power for the same sound power.
However, the power dissipated in the coil increases its resistance in proportion to the increase of its temperature, it diminishes the efficiency of the driver by reducing the current flowing through the coil, as the amplifier is a voltage source i=U/Re. We call this phenomena: thermal compression.
Moreover, more power means more components within the electronic path, more thermal dissipation through them, so more electronic and thermal non-linearity, and then more distortions. If we need fewer watts, we may use simpler circuitry; even class A schematic, more linear, with lower distortion and renowned musicality.
We understand that higher sensitivity indirectly induce lower distortions in real listening conditions, even in the associated electronic stages.
From this screed you should retain three essential notions regarding the choice of your loudspeakers:
– The power handling of a loudspeaker reveals nothing of its qualities.
– The subjective qualities of a driver are intimately linked to the profile and paper pulp of the cone. It’s a rare and precious know-how.
– High sensitivity driver let us use a low power amplifier, either tube or solid-state, intrinsically more linear and therefore more musical.