Sharing tips on designing, debugging, and tuning smart speakers

From time to time, I see some audiophiles spending a lot of money or a lot of time to DIY speakers, but the results are not what they want. Below I will list some of the small experiences I have summarized based on my work, hoping to help those who are interested.

Some tips on speakers

1 In principle, the speaker is basically equivalent to a high-pass filter. In the usual sense, the speaker design is mainly aimed at the bass part, and the mid-high part basically only depends on the speaker unit, crossover and box shape.

2 For each specific bass or mid-bass speaker unit, there is at least one suitable cabinet form. In general, the specific cabinet form is selected according to the ratio of the resonant frequency of the speaker unit and the total quality factor of the speaker unit. It is generally believed that speaker units with a Fs/Qts ratio of less than 40Hz are suitable for labyrinth speakers. Speaker units with a Fs/Qts ratio between 40Hz and 80Hz are suitable for closed speakers. Speaker units with a Fs/Qts ratio between 80Hz and 120Hz are suitable for bass reflex speakers. Speaker units with a Fs/Qts ratio greater than 120Hz are suitable for exponential horn speakers. However, this is not absolute. Generally speaking, everything is conditional. It is of little practical significance to avoid conditions and talk about conclusions. In modern speaker unit design, due to the guidance of T/S parameters, under normal circumstances, only some parameters need to be optimized to obtain the expected effect.

3 As far as the bass reflex speakers are concerned, each low-frequency or mid-low frequency speaker unit has a suitable upper and lower limit of the net volume of the cabinet. Generally speaking, the smaller the net volume of the cabinet, the higher the resonant frequency, and the lower end of the frequency response tends to bulge. The larger the net volume of the cabinet, the lower the resonant frequency, and the lower end of the frequency response tends to decrease. The best net volume of the speaker usually mentioned is only for the design of the flattest frequency response. It is often seen that some improperly debugged bass reflex speakers have a large net volume of the cabinet, but the Q value is often adjusted low, which leads to soft and weak low frequencies. On the contrary, the net volume of the cabinet is small, but the Q value is often adjusted high, which leads to the bulging of the low end of the frequency response, and the speaker is prone to rumbling and humming resonance sounds, and the transient response of the low frequency becomes very poor.

4. When the area of ​​the inverted hole increases, the resonant frequency of the inverted hole increases, and the output sound pressure level increases; when the area of ​​the inverted hole decreases, the resonant frequency of the inverted hole decreases, and the output sound pressure level decreases. When the length of the inverted hole increases, the resonant frequency of the inverted hole decreases, and the output sound pressure level decreases; when the length of the inverted hole decreases, the resonant frequency of the inverted hole increases, and the output sound pressure level increases. The shape of the inverted port has a great influence on the radiation of low frequencies. The use of a graceful shape can achieve a good coupling effect with the air outside the inverted hole. The air friction distortion caused by the inverted hole will be greatly reduced, and the output sound pressure level of the low frequency will be improved. For specific examples, please refer to the relevant patents of BOSE and B&W.

5 The damping of the speaker and the horn should be appropriate, or in other words, there should be an optimal compliance ratio and tuning ratio between them. The compliance ratio and tuning ratio of the speaker unit with different Qts values ​​and the speaker cabinet are different. For details, please refer to the Nomo diagram in the design method of ANThiele and RHSmall. This can also be regarded as a theoretical basis for the double peak debugging method of the bass reflex box.

6 The Q value of a speaker usually refers to the Q value of the entire speaker system, which includes the Q value of the speaker, cabinet, crossover, wire, and power amplifier. Among them, the Q value of the speaker, cabinet, and crossover is the main part, and the influence of other parts is usually small.

7 Generally speaking, high Q value means high damping, and low Q value means low damping.

If speakers are divided according to their most basic uses, there are two most basic models: appreciation type and monitoring type. In daily life, we are most exposed to appreciation type speakers. Since it is a speaker positioned for music appreciation, there are certain targets and trade-offs in design and debugging. For speakers that mainly serve families and individuals and are used to appreciate music, different designers have different cultural backgrounds and cultural environments. The speakers they design and debug are also very different and wonderful in temperament, personality, and sound. This is also the main cause and source of different styles of sound such as British sound, American sound, and Nordic sound. Due to work, the author has the honor of contacting various speakers with different sound bases, many brands, and well-known or unknown. There are both popular brands, HiFi and HiEnd brands in Europe and the United States, as well as some inferior brands from South America, Africa and other countries. Since the factory I serve mainly provides EMS, ODM and OEM services, I have also come into contact with customers' various sound base requirements. Moreover, since we are providing EMS, ODM and OEM services, it is natural that we have to do everything we can to meet the customer's requirements. During the design, debugging and sound adjustment, we have to do some things and not do some things, and more often we have to make compromises and compromises.

I am not talented, but I dare to put forward a "good sound" suggestion here, hoping that those who come after me can avoid detours and pay less tuition. When designing and debugging appreciation speakers, I prefer to adjust the sound to be lively, relaxed, natural and cheerful. I don't like dull, dragging and lifeless sounds. On this basis, the sound is adjusted to be sweeter and more moist, and the frequency response of the high and low ends is further expanded. The density and layering of the sound are adjusted as much as possible, and the service angle of the speaker is adjusted to be wider, so that the sound sounds more open and lively. Finally, the phase response and group delay of the speaker are adjusted to a linear relationship, so that the positioning and sound field of the speaker are better. At this time, it is particularly important not to adjust it to a non-linear relationship. When there are difficulties, it is better to compromise on the flatness of the frequency response rather than the non-linear relationship.

Many audiophiles tend to think that the more expensive the materials and components, the better the sound. The author's factory experience tells me that under the pressure of cost, we can only choose suitable and applicable ones instead of expensive ones. Not only audiophiles have this idea, but many engineers also have similar misunderstandings: they think that only expensive components can be used to design and manufacture "good-sounding" audio equipment. In fact, in companies with good design, sophisticated and complete manufacturing processes, radios worth dozens of yuan can still make beautiful sounds to impress listeners when they are well debugged. For general user products (audio-visual appliances), first adjust the listening experience to balance, then adjust the sound to be softer and more beautiful, so that the listener feels comfortable and relaxed in the listening experience, and if possible, adjust the frequency response of the high and low ends to be wider. After such treatment, it can basically meet the requirements of most ordinary consumers.

The basic requirement for debugging and adjustment is to avoid increasing costs as much as possible. The specific approach is: use the electronic part to compensate for the acoustic part, or conversely use the acoustic part to deal with the difficulties of the electronic part, and use design and structure to balance and compensate for the deficiencies of the speakers or speakers.