Enjoy, my DAC feast

My DAC Complex

Since I was fascinated by vinyl and original editions, I have disdained CD products. Could it be that the advanced digital equipment brought by the progress of the times is not as good as the outdated analog equipment? Although the old analog sound source has its advantages, it is difficult to meet the needs in terms of durability and content diversity compared to today. Digital sound source is the only choice . If the effect standard of LP is used to require CD, there is no difficulty in theory, but in actual application, there are very few products that can meet this requirement. What is the reason? So I also want to know some relevant answers through practice. The three major D/A chips TDA1541, AD1865, and PCM63 are said to be the three classics handed down from generation to generation . They are quite popular. I have always listened to many different versions of products, and the taste is difficult to express in words. There are imported machines and DIY products. The good and durable ones are all borrowed from friends. I am not very satisfied with the machines I own, haha.

Let's start with these classic chips and make them according to our own ideas. We can try to see what level of DIY works we can make and whether we can reach the level of the masters. I personally don't have much confidence because I rarely know about digital circuits. I have to cram for relevant knowledge and learn from the experience of netizens. I try to combine various chips with different uses into circuits.

Success requires hard work. After some hard work and adjustment, I have completed the following two things (time is limited, I will continue with the third one later).

1: First, I made the TDA1541, which is the most popular and well-known in the forum. It is said that it is not easy to make this DAC. I have also bought several different versions of finished PCBs and assembled them myself. I won’t talk about the final effect. There are good and bad ones. I think the most difficult thing about this D/A is the sense of space or the depth of the sound field and the bottom noise. For these DIY operations, some versions of the sound card test indicators are very good, but the listening experience is not ideal. I remember seeing Ming Daxia and other netizens mention the knowledge and skills of wiring more than once. I have read some related theories and think that this aspect is indeed the premise of good sound. These experiences need to be learned. Power quality is the cornerstone of sound quality. This seems to be a consensus among the majority of DIYers, but the various circuit forms are dazzling. LM317 and 337 voltage regulators are currently used by most designers, but I have a little experience: I prefer the sound expressed by the voltage regulator effect of the 78 series (personal opinion, please don’t criticize me). Another reason is that I have a lot of such parts on hand, and it is a waste to throw them away. IV/LPF is said to be an important part of DAC, which determines the final sound style of the product. It takes some thought to choose what form of LPF. There are op amps and Marantz-like HDMA in the forum. In order to save the trouble of debugging, I still chose the op amp combination (see the figure below). This output circuit looks like a one-armed old man. A careful analysis of the circuit seems to show that the designer intentionally omitted a lower arm resistor, otherwise the output midpoint could not be 0 potential. However, adjusting the 1.5k resistor can still adjust the current of the output tube, which is why this circuit was chosen. The midpoint output is not at 0 potential, which is not necessarily a bad thing when there is a coupling capacitor. I have seen some foreign players deliberately add a certain bias voltage to the capacitor inside the speaker crossover with a battery when making a speaker crossover. The advantage of doing this is that the phase distortion of the capacitor can be reduced (I don’t know if it is true or not). Therefore, the midpoint of the circuit here is not at 0 potential, which may be the original intention of the designer. The digital circuit is a standard combination, everyone's is the same, there is nothing special about it. After connecting the circuit diagram, I started drawing the board. The software I used was the simplest Sprint-Layout. It was also the first time I did thermal transfer. It was all the result of diving and learning in the forum. I won't repeat the method and process here. Please look at the picture... After power on, everything is normal, haha, no BUG, ​​the field tube is K170, the output tube adjusts the static current to 8ma, the midpoint is 3.9v. 2: The second one is the famous AD1865N-K. It is said that this is an R2R D/A chip full of vinyl charm, which is used by many expensive European and American brand CDs or decoders. Unfortunately, I have never had the opportunity to listen to it. I have only heard some DIY works, and they are NOS methods. Here I will also talk about the NOS method. It does not use oversampling digital filtering. The listening experience is indeed special, but in my opinion, this method is just a preference for a certain taste or the need for equipment matching. It does not fully express various sounds (personal opinion, please be merciful). However, the choice of which chip to use for oversampling digital filtering has also troubled me. The models with better reputation are roughly SM5843/5813, DF1700, PMD100 (HDCD). Combined with the current market price , from the perspective of cost performance, I chose SM5813APT, so that the digital circuit CS8412CP+SM5813APT+AD1865N-K is built. There is no change in the power supply part, and the same circuit as TDA1541 is used. AD1865 has an internal IV, and it can produce sound normally by directly connecting LPF. However, considering the influence of IV and the unknown performance characteristics of its internal IV, combined with the DIY production of other netizens, it is better to abandon its internal IV and use the current output type connection method, plus IV conversion circuit and LPF, which increases the playability of the circuit. As for which form of IV/LPF circuit to choose, for simplicity, op amp and second-order filtering are still selected. The reason why a more complex form is not selected is that 8x sampling digital filtering has been used before, and in order to obtain a more natural sound and keep the characteristics of AD1865 as much as possible. The board is also made by thermal transfer. Haha, it is better the second time... I will talk about it in the picture above..

1: First, I made the TDA1541, which is the most popular and well-known in the forum. It is said that it is not easy to make this DAC. I have also bought several different versions of finished PCBs and assembled them myself. I won’t talk about the final effect. There are good and bad ones. I think the most difficult thing about this D/A is the sense of space or the depth of the sound field and the bottom noise. For these DIY operations, some versions of the sound card test indicators are very good, but the listening experience is not ideal. I remember seeing Ming Daxia and other netizens mention the knowledge and skills of wiring more than once. I have read some related theories and think that this aspect is indeed the premise of good sound. These experiences need to be learned. Power quality is the cornerstone of sound quality. This seems to be a consensus among the majority of DIYers, but the various circuit forms are dazzling. LM317 and 337 voltage regulators are currently used by most designers, but I have a little experience: I prefer the sound expressed by the voltage regulator effect of the 78 series (personal opinion, please don’t criticize me). Another reason is that I have a lot of such parts on hand, and it is a waste to throw them away. IV/LPF is said to be an important part of DAC, which determines the final sound style of the product. It takes some thought to choose what form of LPF. There are op amps and Marantz-like HDMA in the forum. In order to save the trouble of debugging, I still chose the op amp combination (see the figure below). This output circuit looks like a one-armed old man. A careful analysis of the circuit seems to show that the designer intentionally omitted a lower arm resistor, otherwise the output midpoint could not be 0 potential. However, adjusting the 1.5k resistor can still adjust the current of the output tube, which is why this circuit was chosen. The midpoint output is not at 0 potential, which is not necessarily a bad thing when there is a coupling capacitor. I have seen some foreign players deliberately add a certain bias voltage to the capacitor inside the speaker crossover with a battery when making a speaker crossover. The advantage of doing this is that the phase distortion of the capacitor can be reduced (I don’t know if it is true or not). Therefore, the midpoint of the circuit here is not at 0 potential, which may be the original intention of the designer. The digital circuit is a standard combination, everyone's is the same, there is nothing special about it. After connecting the circuit diagram, I started drawing the board. The software I used was the simplest Sprint-Layout. It was also the first time I did thermal transfer. It was all the result of diving and learning in the forum. I won't repeat the method and process here. Please look at the picture...

After power on, everything is normal, haha, no BUG, ​​the field tube is K170, the output tube adjusts the static current to 8ma, the midpoint is 3.9v.

2: The second one is the famous AD1865N-K. It is said that this is an R2R D/A chip full of vinyl charm, which is used by many high-priced European and American brand CDs or decoders. Unfortunately, I have never had the opportunity to listen to it. I have only heard some DIY works, and they are NOS methods. Here I will also talk about the NOS method. It does not use oversampling digital filtering. The listening experience is indeed special, but in my opinion, this method is just a preference for a certain taste or the need for equipment matching. It does not fully express various sounds (personal opinion, please be merciful). However, the choice of which chip to use for oversampling digital filtering has also troubled me. The models with better reputation are roughly SM5843/5813, DF1700, PMD100 (HDCD). Combined with the current market price , from the perspective of cost performance, I chose SM5813APT, so that the digital circuit CS8412CP+SM5813APT+AD1865N-K is built. There is no change in the power supply part, and the same circuit as TDA1541 is used. AD1865 has an internal IV, and it can produce sound normally by directly connecting LPF. However, considering the influence of IV and the unknown performance characteristics of its internal IV, combined with the DIY production of other netizens, it is better to abandon its internal IV and use the current output type connection method, plus IV conversion circuit and LPF, which increases the playability of the circuit. As for which form of IV/LPF circuit to choose, for simplicity, op amp and second-order filtering are still selected. The reason why a more complex form is not selected is that 8x sampling digital filtering has been used before, and in order to obtain a more natural sound and keep the characteristics of AD1865 as much as possible. The board is also made by thermal transfer. Haha, it is better the second time... I will talk about it in the picture above..