Should the UMZCH have a low output impedance? Encyclopedia of radio electronics and electrical engineering

Encyclopedia of radio electronics and electrical engineering / Transistor power amplifiers

 Comments on the article

About reducing intermodulation distortion and overtones in loudspeakers

The difference in the sound of loudspeakers when working with different UMZCHs is primarily noticed by comparing tube and transistor amplifiers: the spectrum of their harmonic distortion is often significantly different. Sometimes there are noticeable differences among amplifiers of the same group. For example, in one of the audio magazines, the ratings given by the 12 and 50 W lamp UMZCH tended in favor of a less powerful one. Or was the assessment biased?

It seems to us that the author of the article convincingly explains one of the mystical reasons for the occurrence of transient and intermodulation distortions in loudspeakers, which create a noticeable difference in sound when working with various UMZCH. It also offers affordable methods to significantly reduce the distortion of loudspeakers, which are quite simply implemented using modern element base.

It is now generally accepted that one of the requirements for a power amplifier is to ensure that its output voltage remains unchanged when the load resistance changes. In other words, the output resistance of the UMZCH should be small compared to the load one, amounting to no more than 1/10,,,1/1000 of the resistance module (impedance) of the load |Z_н|. This view is reflected in numerous standards and recommendations, as well as in the literature. Specially introduced even such a parameter as the damping coefficient - K_d (or damping factor) equal to the ratio of the nominal load resistance to the output impedance of the amplifier R_O MIND. So, with a nominal load resistance of 4 ohms and an amplifier output impedance of 0,05 ohms K_d would be 80. The current HiFi standards require that high-quality amplifiers have a damping factor of at least 20 (and at least 100 is recommended). For most transistor amplifiers on the market, K_d exceeds 200.

Reasons for small R_O UM (and correspondingly high K_d) are well known: this is to ensure the interchangeability of amplifiers and loudspeakers, to obtain effective and predictable damping of the main (low-frequency) resonance of the loudspeaker, as well as the convenience of measuring and comparing the characteristics of amplifiers. However, despite the legitimacy and validity of the above considerations, the conclusion about the need for such a ratio, according to the author, is fundamentally erroneous!

The thing is that this conclusion is made without taking into account the physics of the work of electrodynamic loudspeaker heads (GG). The vast majority of amplifier designers sincerely believe that all that is required of them is to deliver the required voltage at a given load resistance with as little distortion as possible. Loudspeaker designers, for their part, seem to assume that their products will be powered by amplifiers with negligible output impedance. It would seem that everything is simple and clear - what questions can there be?

Nevertheless, there are questions, and very serious ones. The main one is the question of the magnitude of intermodulation distortion introduced by the GG when it is operated from an amplifier with negligible internal resistance (voltage source or EMF source).

"What does the output impedance of the amplifier have to do with this? Don't fool me!" the reader will say. - And he's wrong. It has, and the most direct, despite the fact that the fact of this dependence is mentioned extremely rarely. In any case, no modern works have been found that consider this effect on all parameters of the end-to-end electroacoustic path - from the voltage at the amplifier input to sound vibrations. For some reason, when considering this topic, we were previously limited to analyzing the behavior of the GG near the main resonance at low frequencies, while no less interesting things happen at noticeably higher frequencies - a couple of octaves above the resonant frequency.

This article is intended to fill this gap. It must be said that in order to increase accessibility, the presentation is very simplified and schematized, so a number of "subtle" issues remained unconsidered. So, in order to understand how the output impedance of the UMZCH affects intermodulation distortion in loudspeakers, we must remember what the physics of sound radiation from a GG cone is.

Below the main resonance frequency, when a sinusoidal signal voltage is applied to the winding of the GG voice coil, the displacement amplitude of its diffuser is determined by the elastic resistance of the suspension (or air compressed in a closed box) and is almost independent of the signal frequency. The operation of the GG in this mode is characterized by large distortions and a very low return of a useful acoustic signal (very low efficiency).

At the fundamental resonance frequency, the mass of the diffuser, together with the oscillating mass of air and the elasticity of the suspension, form an oscillatory system similar to a weight on a spring. The efficiency of radiation in this frequency range is close to the maximum for this HG.

Above the main resonance frequency, the inertial forces of the diffuser, together with the oscillating air mass, turn out to be greater than the elastic forces of the suspension, so the diffuser displacement turns out to be inversely proportional to the square of the frequency. However, the acceleration of the cone in this case does not theoretically depend on the frequency, which ensures the uniformity of the frequency response in terms of sound pressure. Therefore, to ensure the uniformity of the frequency response of the HG at frequencies above the frequency of the main resonance, a force of constant amplitude must be applied to the diffuser from the side of the voice coil, as follows from Newton's second law (F=m*a).

The force acting on the cone from the voice coil is proportional to the current in it. When the GG is connected to a voltage source U, the current I in the voice coil at each frequency is determined from Ohm's law I (f) \uXNUMXd U / Z_г(f), where Z_г(f) is the frequency-dependent complex impedance of the voice coil. It is determined mainly by three quantities: the active resistance of the voice coil R_г (measured with an ohmmeter), inductance L_Mr. The current is also affected by the back-emf that occurs when the voice coil moves in a magnetic field and is proportional to the speed of movement.

At frequencies much higher than the main resonance, the back-emf value can be neglected, since the cone with the voice coil simply does not have time to accelerate in half the period of the signal frequency. Therefore, the dependence Z_г(f) above the fundamental resonance frequency is determined mainly by the values ​​of R_г and L_г

So, neither resistance R_g, nor inductance L_г are not particularly stable. The resistance of the voice coil is highly dependent on temperature (TCR copper approx. +0,35%/^оC), and the temperature of the voice coil of small-sized mid-frequency GG during normal operation changes by 30 ... 50 ^оWith and rather quickly - in tens of milliseconds or less. Accordingly, the resistance of the voice coil, and hence the current through it, and the sound pressure at a constant applied voltage change by 10 ... 15%, creating intermodulation distortion of the corresponding value thermal signal compression).

Inductance changes are even more complex. The amplitude and phase of the current through the voice coil at frequencies noticeably higher than the resonant one are largely determined by the value of the inductance. And it very much depends on the position of the voice coil in the gap: with a normal displacement amplitude for frequencies that are only slightly higher than the fundamental resonance frequency, the inductance changes by 15 ... 40% for different GGs. Accordingly, at the rated power supplied to the loudspeaker, intermodulation distortion can reach 10 ... 25%.

The above is illustrated by a photograph of sound pressure oscillograms taken on one of the best domestic mid-frequency GG - 5GDSH-5-4. The block diagram of the measuring setup is shown in the figure.

As a source of a two-tone signal, a pair of generators and two amplifiers were used, between the outputs of which the test GG was connected, installed on an acoustic screen with an area of ​​about 1 m² . Two separate amplifiers with a large power margin (400 W) are used to avoid the formation of intermodulation distortion during the passage of a two-tone signal through the amplification path. The sound pressure developed by the head was perceived by a ribbon electrodynamic microphone, the non-linear distortion of which is less than -66 dB at a sound pressure level of 130 dB. The sound pressure of such a loudspeaker in this experiment was approximately 96 dB, so that the distortion of the microphone under these conditions could be neglected.

As can be seen on the oscillograms on the screen of the upper oscilloscope (upper - without filtering, lower - after HPF filtering), the modulation of a signal with a frequency of 4 kHz under the influence of another with a frequency of 300 Hz (with a head power of 2,5 W) exceeds 20%. This corresponds to an intermodulation distortion of about 15%. It seems that there is no need to remind that the threshold of perceptibility of intermodulation distortion products is much lower than one percent, reaching hundredths of a percent in some cases. It is clear that the distortions of the UMZCH, if only they are of a "soft" nature, and do not exceed a few hundredths of a percent, are simply indistinguishable against the background of distortions in the loudspeaker caused by its operation from a voltage source. Intermodulation distortion products destroy the transparency and detail of the sound - a "mess" is obtained, in which individual instruments and voices are heard only occasionally. This type of sound is probably well known to readers (a good test for distortion can be a phonogram of a children's choir).

However, there is a way to drastically reduce the distortion described above, caused by the variability of the speaker head impedance. To do this, the amplifier driving the loudspeaker must have an output impedance that is much greater than the components of the impedance R_{g and} X_г (2p fLg) GG. Then their changes will have practically no effect on the current in the voice coil, and, consequently, the distortions caused by these changes will also disappear. In order to demonstrate the effectiveness of this method of reducing distortions, the measuring setup was supplemented with a 47 Ohm resistor (i.e., an order of magnitude greater than the impedance modulus of the studied GG), connected in series with the GG. To maintain the same sound pressure level, the signal levels at the outputs of the amplifiers were correspondingly increased. The effect of switching to the current mode is obvious from a comparison of the corresponding oscillograms: the parasitic modulation of the high-frequency signal on the screen of the lower oscilloscope is much smaller and barely visible, its value does not exceed 2 ... 3% - there is a sharp decrease in HG distortion.

Connoisseurs may object that there are many ways to reduce the variability of the impedance of the voice coil: this is filling the gap with a cooling magnetic fluid, and installing copper caps on the cores of the magnetic system, and careful selection of the core profile and coil winding density, and much more. However, all these methods, firstly, do not solve the problem in principle, and secondly, they lead to the complication and increase in the cost of the production of HGs, as a result of which they are not fully used even in studio loudspeakers. That is why most mid-frequency and low-frequency GGs have neither copper caps nor magnetic fluid (in such GGs, when operating at full power, the liquid is often ejected from the gap).

Therefore, powering the GG from a high-impedance signal source (in the limit - from a current source) is a useful and expedient way to reduce their intermodulation distortion, especially when building multiband active acoustic systems. In this case, damping of the main resonance has to be performed purely acoustically, since the intrinsic acoustic quality factor of mid-frequency GGs, as a rule, significantly exceeds one, reaching 4...8.

It is curious that it is precisely this mode of "current" power supply of the GG that takes place in lamp UMZCHs with a pentode or tetrode output with a shallow (less than 10 dB) FOS, especially if there is a local FOS for current in the form of resistance in the cathode circuit.

In the process of setting up such an amplifier, its distortions without a general OOS usually turn out to be within 2,..5% and are confidently noticeable by ear when included in the break of the control path (comparison method with the "straight wire"). However, after connecting an amplifier to a loudspeaker, it is found that as the depth of feedback increases, the sound first improves, and then there is a loss of detail and transparency. This is especially noticeable in a multi-band amplifier, the output stages of which drive directly to the corresponding loudspeaker heads without any filters.

The reason for this, at first glance, a paradoxical phenomenon is that with an increase in the OOS depth in voltage, the output impedance of the amplifier decreases sharply. The negative consequences of powering the GG from an UMZCH with a low output impedance are discussed above. In a triode amplifier, the output impedance, as a rule, is much less than in a pentode or tetrode, and the linearity before the introduction of feedback is higher, so the introduction of feedback on voltage improves the performance of a single amplifier, but at the same time worsens the performance of the loudspeaker head. As a result, as a result of introducing an output voltage feedback into a triode amplifier, the sound can actually become worse, despite the improvement in the characteristics of the amplifier itself! This empirically established fact serves as inexhaustible food for speculation on the topic of harm from the use of feedback in audio power amplifiers, as well as arguments about the special, tube-like transparency and naturalness of sound. However, from the above facts, it clearly follows that the point is not in the presence (or absence) of the OOS itself, but in the resulting output impedance of the amplifier. That's where the "dog is buried"!

It is worth saying a few words about the use of negative output resistance UMZCH. Yes, positive current feedback (POF) helps to dampen the GG at the fundamental resonance frequency and reduce the power dissipated in the voice coil. However, one has to pay for the simplicity and efficiency of damping by increasing the influence of the GG inductance on its characteristics, even in comparison with the operation mode from a voltage source. This is because the time constant L_г/R_{r is replaced by a larger one, equal to} L_г/[R_г+(-R_{output PA})]. Accordingly, the frequency decreases, starting from which the inductive reactance begins to dominate in the sum of the impedances of the "GG + UMZCH" system. Similarly, the influence of thermal changes in the active resistance of the voice coil increases: the sum of the changing resistance of the voice coil and the constant negative output resistance of the amplifier changes more in percentage terms.

Of course, if R_out.PA in absolute value does not exceed 1/3 ... 1/5 of the active resistance of the voice coil winding, the loss from the introduction of the POS is small. Therefore, a weak current POS for a small additional damping or for fine tuning of the quality factor in the low-frequency band can be used. In addition, the current POS and the current source mode in the UMZCH are not compatible with each other, as a result of which the current supply of the GG in the low-frequency band, unfortunately, is not always applicable.

With intermodulation distortion, we apparently figured it out. Now it remains to consider the second question - the magnitude and duration of the overtones that arise in the diffuser of the GG when reproducing signals of an impulse nature. This question is much more complicated and "thinner".

As is known, GH diffusers can be considered infinitely rigid only in a very rough approximation. In fact, when they vibrate, they bend significantly, and in a very bizarre way. This is due to the presence of a large number of parasitic resonant frequencies of the diffuser and the moving system of the HG as a whole. After the passage of the pulsed signal, free oscillations at each of the resonant frequencies do not die out immediately, generating overtones, coloring the sound and hiding clarity and detail, worsening the stereo effect.

There are theoretically two possibilities to eliminate these overtones. The first is to shift all resonant frequencies beyond the operating frequency range, into the region of far ultrasound (50...100 kHz). This method is used in the development of low-power high-frequency GGs and some measuring microphones. With regard to the GG, this is a method of a "hard" diffuser.

The second possibility is to reduce the quality factor of parasitic resonances so that the oscillations die out so quickly that they cannot be heard. This requires the use of "soft" diffusers, the bending losses of which are so large that the quality factor of parasitic resonances is close to unity. However, the non-linear distortions and the maximum sound pressure of a GG with a "soft" diffuser turn out to be somewhat worse than that of a GG with a "hard" cone. On the other hand, GG with "soft" cones, as a rule, win significantly in terms of clarity, uncoloredness and transparency of the sound.

So, a third option is also possible - the use of a GG with a relatively "hard" diffuser and the introduction of its acoustic damping. In this case, it is possible to combine the advantages of both approaches to some extent. This is how studio control loudspeakers (large monitors) are most often built. Naturally, when the damped HG is powered from a voltage source, the frequency response is significantly distorted due to a sharp drop in the total quality factor of the main resonance. The current source in this case also turns out to be preferable, since it helps to equalize the frequency response simultaneously with the exclusion of the effect of thermal compression.

As for the overtones arising from the free oscillations of the diffusers of the GG, then, since parasitic resonant frequencies are usually located much higher than the frequency of the main resonance, the operating mode of the GG - with a current or voltage source - has practically no effect on them. The only direct way to deal with parasitic resonances is acoustic damping. However, the probability of their excitation when the GG is powered from a current source is less, since these resonances become most noticeable when they are excited by distortion products. Both the absolute and relative amplitudes of these distortion products for this operating mode of the GG turn out to be significantly smaller.

Summarizing the above, we can draw the following practical conclusions:

1. The loudspeaker head operating mode from a current source (as opposed to a voltage source) provides a significant reduction in intermodulation distortion introduced by the head itself.

2. The most appropriate design option for a loudspeaker with low intermodulation distortion is an active multi-band, with a crossover filter and separate amplifiers for each band. However, this conclusion is true regardless of the GG diet.

3. The operation of the heads from current sources causes the need for acoustic damping of their main resonance, as a result of which some damping of parasitic resonances of the moving system is also achieved along the way. This improves the impulse response of the loudspeaker and helps to eliminate additional sound coloration.

4. In order to obtain a high output impedance of the amplifier and maintain a small amount of its distortion, OOS should be used not in terms of voltage, but in terms of current.

Of course, the author understands that the proposed method of reducing distortion is not a panacea. In addition, in the case of using a ready-made multiband loudspeaker, the current supply of its individual GGs without alteration is impossible. An attempt to connect a multi-band loudspeaker as a whole to an amplifier with increased output impedance will lead not so much to a decrease in distortion, but to a sharp distortion of the frequency response and, accordingly, a failure of the tonal balance. Nevertheless, the reduction of intermodulation distortions of the GG by almost an order of magnitude, and by such an accessible method, clearly deserves worthy attention.

The author thanks NIKFI staff members A.P. Syritso. for help with measurements and Shraibman A.E. for discussing the results.

Author: S. Ageev, Moscow; Publication: cxem.net

See other articles Section Transistor power amplifiers.

Read and write useful comments on this article.

<< Back

Latest news of science and technology, new electronics:

Machine for thinning flowers in gardens 02.05.2024

In modern agriculture, technological progress is developing aimed at increasing the efficiency of plant care processes. The innovative Florix flower thinning machine was presented in Italy, designed to optimize the harvesting stage. This tool is equipped with mobile arms, allowing it to be easily adapted to the needs of the garden. The operator can adjust the speed of the thin wires by controlling them from the tractor cab using a joystick. This approach significantly increases the efficiency of the flower thinning process, providing the possibility of individual adjustment to the specific conditions of the garden, as well as the variety and type of fruit grown in it. After testing the Florix machine for two years on various types of fruit, the results were very encouraging. Farmers such as Filiberto Montanari, who has used a Florix machine for several years, have reported a significant reduction in the time and labor required to thin flowers. ... >>

Advanced Infrared Microscope 02.05.2024

Microscopes play an important role in scientific research, allowing scientists to delve into structures and processes invisible to the eye. However, various microscopy methods have their limitations, and among them was the limitation of resolution when using the infrared range. But the latest achievements of Japanese researchers from the University of Tokyo open up new prospects for studying the microworld. Scientists from the University of Tokyo have unveiled a new microscope that will revolutionize the capabilities of infrared microscopy. This advanced instrument allows you to see the internal structures of living bacteria with amazing clarity on the nanometer scale. Typically, mid-infrared microscopes are limited by low resolution, but the latest development from Japanese researchers overcomes these limitations. According to scientists, the developed microscope allows creating images with a resolution of up to 120 nanometers, which is 30 times higher than the resolution of traditional microscopes. ... >>

Air trap for insects 01.05.2024

Agriculture is one of the key sectors of the economy, and pest control is an integral part of this process. A team of scientists from the Indian Council of Agricultural Research-Central Potato Research Institute (ICAR-CPRI), Shimla, has come up with an innovative solution to this problem - a wind-powered insect air trap. This device addresses the shortcomings of traditional pest control methods by providing real-time insect population data. The trap is powered entirely by wind energy, making it an environmentally friendly solution that requires no power. Its unique design allows monitoring of both harmful and beneficial insects, providing a complete overview of the population in any agricultural area. “By assessing target pests at the right time, we can take necessary measures to control both pests and diseases,” says Kapil ... >>

Random news from the Archive Moving sidewalks for cities 25.11.2016 Travolators or moving sidewalks can hardly be called an innovation. The first "moving sidewalk" was shown back in 1893 in Chicago (USA) at the World's Columbian Exposition, and in 1900 it premiered at the World's Fair in Paris. Travelators were already seen as the transport of the future, and H.G. Wells used them to describe London in the 1890nd century in A Tale of the Days to Come and When the Sleeper Awakes, written in the XNUMXs. Despite the optimism of a century ago and the development of technology, travelators have remained a niche vehicle used in airports, shopping centers, museums and zoos. The use of travelator as public transport is limited to underground metro stations in different cities of the world. Recently, employees of the Federal Polytechnic School of Lausanne (Switzerland) again returned to the topic of the travelator as urban transport. Using a mathematical model, they showed that in medium-sized European cities, moving sidewalks were able to carry 7000 passengers per hour on each lane more efficiently than buses. As a model, the researchers took Geneva, completely devoid of cars, in which passengers use the metro, trams, taxis, bicycles or city funiculars. It turned out that the most efficient configuration of a moving sidewalk in a city of this size is a small circle around the city center, in which there are no cars. The length of the sidewalk network of 47 links and 10 exits will be 32 km on the main roads. In addition, the system provides 37 intersections connecting "slow" travelators (speed up to 15 km / h) with high-speed "highways" using bridges and tunnels in order not to collide with conventional transport.

Other interesting news:

▪ Polymer LED indicators

▪ Plastic planes getting ready to take off

▪ Converting hydrogen to metal

▪ Self-cleaning film repels bacteria

▪ Giant accumulations of turbulent gas discovered in distant galaxies

News feed of science and technology, new electronics

Interesting materials of the Free Technical Library:

▪ site section Frequency synthesizers. Selection of articles

▪ article Study of control systems. Lecture notes

▪ article How dangerous are asteroids? Detailed answer

▪ Book article. Legends, cultivation, methods of application

▪ article Probe-oscilloscope. Encyclopedia of radio electronics and electrical engineering

▪ article Pop-up spoons. Focus Secret

Leave your comment on this article:

All languages ​​of this page

Arabic	Hebrew	Polish
Bengali	Hindi	Portuguese
Chinese	Italian	Spanish
English	Japanese	Turkish
French	Korean	Ukrainian
German	Malay	Vietnamese

Home page | Library | Articles | Website map | Site Reviews

www.diagram.com.ua
2000-2024