home > projects > bigaudio > ras_sub
Rapture A. Sys.
System Design
Electrostatic (ESL)
Halcyon DAC
Antithesis Sub.

Rapture Audio System

Text has not been updated or edited.


Purpose of Crossovers

Crossovers hold a very important role in loudspeaker design and operation. Each loudspeaker transducer, whether it is dynamic, electrostatic, planar magnetic, piezo, ribbon, or plasma is capable of producing only a limited range of frequencies. Even if a specific transducer is capable of producing sound at a given range, the efficiency, power, and response may be severely limited. In response, speaker systems are built using multiple transducers optimized for a specific frequency range, such that the system as a whole covers the entire audio spectrum. Crossovers direct the power from the incoming signal to the appropriate transducer by its frequency response characteristics. By nature, the crossover is a set of electrical filters. Crossovers also serve to protect loudspeaker transducers. A high power bass note can destroy an expensive silk dome dynamic tweeter, even if it is not capable of reproducing the sound. The crossover prevents unwanted signals from reaching the transducer.

Passive Crossovers

Common consumer audio equipment uses passive crossovers. These crossovers are placed between the power amplifier and the transducers of the speaker and are usually internal to the speaker. The crossovers are made from high power, high value resistors, inductors and capacitors (and hence called an RLC circuit). A high pass filter is used on the tweeter, a low pass filter on the woofer, and if the system contains midrange drivers, band pass filters for each midrange. Depending on the number of components and the filters can have a variety of characteristics. These characteristics include cutoff frequency, cutoff slope, pass band ripple, and phase response. This type of system has been used for years and at the outset, relatively cheap. However, a passive crossover has serious flaws.

Because the crossover is after the power amplifier, the crossover dissipates a great deal of power. The greater cutoff slope, the higher number of components needed and therefore even more power is wasted. Also, the tolerances of these large, high power parts are very high, which means the values are not very accurate. This leads to improperly tuned filters. Once constructed and installed, these filters are almost impossible to change, and therefore, the system cannot be tweaked. Sometimes, L-pads are included which allow attenuation of high frequency drivers, but this just leads to more wasted power. So many components also lead somewhat harsh and non-linear phase distortion that varies by frequency. The biggest problem with passive crossovers is that they isolate the drivers from the amplifier. This causes two problems.

First, the impedance of the loudspeaker varies with frequency because it is a reactive device (dynamic speakers are inductors, ESL's are capacitors). A passive crossover is designed for a resistive load that doesn't change properties due to frequency. With the load impedance changing, the response of the crossover changes, and falls rapidly away from the designed specification. The only remedy is to include an impedance-correcting network across the loudspeaker element, which serves only to waste more power and possibly introduce another source of distortion. The second problem with isolating the loudspeaker element from the amplifier is the dampening effects of the amplifier are lost. When a dynamic speaker cone moves, it creates inertia. This inertia will cause the cone to continue moving, even after the signal has reversed direction in a distortion causing condition called overshoot. In order the correct this, the amplifier must supply more immediate power to keep the voltage tracking. However, the inductors and capacitors in the crossover can store energy and effectively decouple the loudspeaker element from the amplifier. Now the amplifier does not know if the loudspeaker is accurately tracking. This destroys transient response in dynamic speakers and has sever effects on ESL efficiency and quality.

Active Crossovers

Active Crossovers are the solution to almost every problem that passive crossovers possess. Active crossovers are filters, like passive crossovers. However, active crossovers work at audio line levels before the main power amplifiers. The output from the crossovers are sent to individual amplifiers, which amplify each frequency range independently and then connect directly to the speaker elements. Active crossovers can be constructed without expensive and inaccurate inductors, and all the parts are small, low power, cheap and most importantly, accurate. Since the output is to an amplifier, the impedance is constant. This is important also because it is very easy to create a multiple pole filter with very sharp and precise response by using multiple op-amps in a chain while maintaining a steady input and output impedance. The properties of active crossovers can be easily changed, and depending on design, the change can be while listening to the system. This offers much greater control and customization. Because the loudspeaker elements are directly connected to the amplifiers, the dampening factor of the amplifiers is not lost, and no power is wasted. Even the amplifiers have an easier load because any single amplifier only has to amplify a specific range of frequencies, offering even more customization. The only tradeoff is that more than one stereo amplifier is needed or at least one amplifier channel per frequency range per channel.

Digital Options

Active crossovers still use reactive components, and therefore non-linearities in phase and/or frequency response exist. Digital technology comes to the rescue with Digital Signal Processors, capable of filtering and processing audio signals in the digital domain. The proliferation of digital technology has created a new way to process audio signals. A Digital Signal Processor is a very specialized computer with the single task of performing rapid real time calculations on a signal. Because the processor relies on a software algorithm to work it is almost infinitely flexible. The two issues of concern are digital conversion quality and processing delay. Although digital audio is claimed to be superior to analog systems, this is not always true. A digital audio signal alone does not degrade in quality, but conversion to and from analog and the use of improper mathematical operators can seriously effect the quality. Also the processor does not render an output immediately, there is an associative lag, which is increased by the time it takes to move signals to and from the digital domain. Therefore, the use of digital technology must be questioned on four aspects: use, effects, efficiency, ease.


The Power Block modules are the home to four different subsystems. Some subsystems have multiple parts but everything will fit together inside of one chassis in blocks for easy construction and upgrades. The subsystems that comprise the Power Block take the input signal from the Central Control Unit (CCU) and do the amplification and power supply for the output transducers, in addition to relaying information back to the CCU. Each Power Block will sit right next to ESL/Tline stack and be required to plug into electrical power. The plan for size and encasement is to build each Power Block inside a gutted ALR full tower computer case.

ESL Amplifier

The ESL amplifier has very special requirements. It must be very stable and capable of driving a highly capacitive load. Although the ESL dissipates no power, it charges to very high energy levels and this energy must be moved very rapidly leading to high power requirements in the amplifier. Also, the output of this amplifier must be high voltage, either by direct drive or a step up transformer.

ESL HV Supply

The ESL also requires a high voltage bias supply. This supply during testing must be variable through about 3-7kV. When the system is turned on or off it would be beneficial to slowly charge and discharge the system to decrease dust build up.

Sub Amplifier

The subwoofer amplifier has different requirements than the ESL amplifier. It must be able to drive a very low impedance load very hard for recreating heart-stopping bass for movies. It does not have to have a high bandwidth but it must still be accurate and well supplied. Also, the plan calls to use an accelerometer to build an active servo control loop into the very end of the bass driver/amplifier stage.

Digital Processing

The input audio stream is processed at the front end in a digital signal processor. The digital crossover function as well as room equalization take place here. The signals are streamed via differential high speed serial links to the power blocks where they are converted to analog directly before the amplifier stage. A control subsystem takes care of housekeeping tasks.