EE-93 Special Topics in Recording Engineering (1 Credit)

1

EE-93 Special Topics in Recording Engineering (1 credit)

Thursdays 6 - 7:15pm, plus 3 - 4:15 TU and THU in Granoff 252;

Tom Bates, instructor; George Nagel, teaching assistant

WEEK 4

Speaker Design: We will limit our discussion to dynamic speaker systems, which make up 95% of all speakers made, sold and used. But we should acknowledge that there are other forms of speakers operating on other principles. The limits or weaknesses of these other designs explain why they have not taken over the marketplace.

For example, there are ribbon speakers, and electrostatic speakers. These are often less powerful than is needed for some listening experiences, and often more complex in their construction and more difficult to repair. Electrostatic speakers are difficult for amplifiers to drive.

Dynamic speakers present their own difficulties, but are generally understood. There are many possibilities to buy drivers and crossover components.

Starting out:

The first step is to decide on a design for the box and woofer. There are many possible approaches, and mid range and high frequency drivers are generally employed in uniform ways, independent of the box design.

Box choices include:

·  No box or baffle (table radio)

Check holding woofer in air – check using a cymbal

·  Infinite baffle (early movie speakers)

Check sealed box large

·  Acoustic suspension

Sealed box small (Hi Fi of the 60s)

·  Horn Loaded

Klipsh Horn

Voice of the Theater

·  Ported

Most contemporary studio monitors

·  Passive Radiators

Choice of design approach is carried out with simultaneous search for low frequency driver. Information needed to find the best driver includes whether utilizing a 2-way or a 3-way system design. Motional feedback?

At this point you are utilizing your computer box and speaker design software such as

(Search URL: www.speakerbuilding.com)

All designs for commercially distributed product utilize speaker and box design software.

Choice of drivers utilizing Thiele Small parameters from library. All suppliers of drivers have these parameters available for expanding the software’s library.

Passive crossover choices:

Simple design – all the way down to a single capacitor.

Mid level design – cascaded Butterworth filter stages.

Complex design – Linkwitz Riley multipole design with time alignment.

Crossovers are also designed with computer help, including Linkwitz Riley.

Choice of physical delay of tweeter signal or electronic delay line

Design of active crossover, and possible built in amplification

Choice of interior baffling material – i.e. rockwool, fiberglass, etc.

Choice of exterior finish and subsequent distraction of wave propogation along the cabinet front

No matter how carefully you plan and implement, there is no way to know in advance the precise nature of the room acoustics in the space the speakers are to be installed. Grrrrr!

The most successful design is one that is adaptable to almost any acoustic environment.

Therefore, the best performance in a multi-driver speaker design comes from audio processing (either analog or digital) to compensate for weaknesses in the design, the drivers, or special circumstances in the environment. The processed audio is then sent to on-board amplifiers designed specifically to power the drivers used in that speaker model.

Amplifier Quick Overview:

A loudspeaker’s cone movement is proportional to the current in the voice coil. Likewise, the amount of heat in the amplifier output components is also proportional to the current. However, it takes voltage to make current flow, so a power amp must deliver both high voltage and high current.

For example, a 200 watt amplifier driving an 8 ohm speaker is designed to deliver 40 volts to an 8 ohm load, resulting in 5 amps of current. If we wish to double the cone motion, we must double the current from 5 amps to 10 amps and the voltage to 80 volts. Therefore, the amplifier rating must increase from 200 watts to 800watts (80V multiplied by 10 A). We can see that the power demands escalate rapidly.

Power amplifiers generally consist of three major stages, plus a power supply, plus protection circuits. The stages are an input buffer with volume control, a voltage gain stage, and a current gain stage.

A traditional amplifier power supply consists of a power transformer, often toroidal, whose output is determined by the secondary windings (i.e. the turns ratio). The output of the transformer goes to a bridge rectifier consisting of four high current diodes, which sends positive power to the positive bus and negative power to the negative bus. Then a LARGE scale capacitor (several, actually) acts as a storage device, holding the electrical energy ready until the amplifier needs to use it.

If the storage capacitors are considerably larger than what is needed for most amplifier activity, they charge up to a higher value during periods of low demand, and can deliver momentary bursts of power above their normal rating. This is called dynamic headroom, and can add 2 or 3 dB of peak undistorted power –which is equivalent to having up to 100% more wattage!

An alternative form of power supply is found in switching power supplies. These are based on the principle that high frequency transformers are much smaller and more efficient than low frequency transformers. In order to take advantage of this principle, however, we must free ourselves from attachment to the 60 Hz power lines.

To do this, we rectify wall current and store the output in a very large capacitor array. This is an extremely stiff supply! However, wall current is offset by 60 volts, more or less, so this must be removed to avoid hazardous shocks. This is most commonly used within sub-woofers were operators can’t get their hands on dangerous voltages.

Then we use a switching transistor to turn on and off at a high rate of speed (perhaps 100 kHz).

This high frequency AC is fed to a small high frequency transformer which isolates the secondary from AC shock hazard.

This high frequency AC is then rectified and filtered again, resulting in the final DC supply for the amplifier.

This design is more complex than the traditional passive power supply, but the weight of the components, and therefore the cost, is much less. There are amplifiers that switch the feed to the power supply, and some that switch the output from the power supply to the audio output. There are some that do both.

In addition to the primary benefit of greatly reduced weight, we can control the operation of the high frequency transistors to compensate for variations in AC voltage and load currents, thus improving both kinds of power-supply regulation.

Class of Operation

Class A:

A single or complementary pair of transistors drives the entire wave form of the audio signal. This results in very low distortion but demands lots of power during low level signals, especially during an idle state. The amplifier sends to the speakers one half its output power capability when in idle. Also, speakers must be protected from transients resulting from this idle state offset at turn on or turn off.

Class B:

Class B operation divides the audio signal’s wave form into a positive path and a negative path. This results in efficient use of drive current, but also results in zero crossing distortion or cross over distortion when the positive and negative halves are re-joined together. This is because it is almost impossible to have a smooth transition between the positive and negative paths of the amplifier circuitry.

Class AB:

Improvements in amplifier design have resulted in the employment of both A and B techniques simultaneously. For a short time, while the signal level is low, and when it is crossing from positive to negative (or vice versa), both halves of the amplifier circuitry are turned on, and work as a complementary pair in class A operation.

When the signal level exceeds a threshold, typically 5 to 10 watts, the amplifier reverts to class B operation. This results in much of the energy savings of class B, with some of the sonic purity of class A.

Class C, E, F:

These classes of operation do not apply to audio circuitry.

Class G:

This mode uses 2 or more sets of output transistors connected to different supply voltages. This allows the amplifier to run at high efficiency by having the delivered audio signal’s level always near the power rails. As the signal gets louder, the amplifier switches to higher sets of rails, and then switches back when the signal is lower in level.

Class H:

This is similar to Class G. This class uses a single bank of output transistors connected to a low voltage supply, along with some means of switching them to a higher voltage supply, when required. This method has the same thermal benefits as Class G, but it avoids the second bank of output transistors, thus reducing the size and cost of the amplifier.

Class D:

In Class D operation, we rapidly switch the output transistors on and off. This is another way of converting DC power into audio power, while reducing inherent heat losses. Because these losses occur when the output transistors are partially on, we avoid this state. We turn them fully on, and send all of the DC power to the load, or we leave them fully off so that no power flows. In both cases, little or no power is wasted in the transistor.

Note: How does on-off switching drive a loudspeaker? The magnetic field in the voice coil does not collapse instantly when the amplifier switches off. The loudspeaker continues interpolating a fraction of the waveform while the amplifier is switching. Remember that the amplifier is willing to switch very quickly. This class of operation is not good at driving electrostatic speakers.

Negative Feedback and Distortion

The output of real world circuitry is distorted – at least a little. Analog circuit components most often act in slightly unpredictable ways, and must be nudged back into compliance with the very exacting demands we require of our audio circuitry. The most common form of this correction is through the use of negative feedback. This can be applied over the entire circuit/amplifier, which is called global feedback. However, negative feedback can be applied stage by stage, resulting in effective elimination of distortion, but without the circuit lag which can result from components that respond too slowly.

Discussion of thermal stability: with Bipolar and FET output stages. Discussion of safety circuits. Discussion of driving difficult loads. Heat from reflected energy.

Note: Bipolar transistors would appear to be better choices for amplifiers than FET output stages. This is because bipolar transistors appear to have a lower impedance when conducting current, and so it would appear that they would easily make larger amplifiers (more watts). However, as bipolar transistors conduct and get hot, their impedance goes down and they start to conduct more power…..and then they lower their impedance some more and conduct some more…… and soon they destroy themselves as this phenomenon gets out of control. FETs, on the other hand, have a raised impedance when they get hot, so this reduces their conductivity and they become self regulating. You can drive a number of them in parallel without needing a special protection circuit like bipolar transistors need. Therefore, manufacturers can afford to put more FETs in the output stages of their amplifiers and have greater conductivity (more watts) than in a similarly priced bipolar model.

Future Developments:

I believe the greatest areas of promise for future development include the following:

Digital audio look ahead

Higher speed switching power supply

Class D operation with controls tied to the switching power supply

Further refinement of Class H operation