The reason why this thing is put in the first place is because it is too important. Our common motherboard is ATX motherboard. It is made of printed circuit board (PCB). It is made on insulating material by electronic printing process. There are mainly four-layer boards and six-layer boards in the market. There are four common boards. The motherboard designed with 6-layer PCB is not easy to deform and its stability is greatly improved. If you are lucky enough to buy a 6-ply board, it is definitely worth it! Ha! Every layer of the motherboard is covered with circuits, so if the PCB board is burnt out, it can be repaired by wiring with our engineers' superb technology. If it is serious, the life of this motherboard will be over! The parts on the motherboard look dazzling, but they are all arranged in a well-organized way. Mainly includes a CPU socket; North bridge chip, south bridge chip, BIOS chip and other three chips; Front-end system bus FSB, memory bus, graphic bus AGP, data exchange bus HUB, peripheral bus PCI, etc. There are seven interfaces: FDD, USB and IDE.

First, the main chips on the motherboard

1, the north bridge chip MCH is a memory control chip on the left side of the CPU socket, also called the north bridge chip, and generally has an aluminum heat sink on it. The main functions of Northbridge chip are data transmission and signal control. On the one hand, it exchanges signals with CPU through the front-end bus, on the other hand, it exchanges signals with memory, AGP and South Bridge. After the North Bridge chip is broken, most of the phenomena are not bright, and sometimes it will crash when it is bright. If the engineer decides that your north bridge chip is broken and your motherboard is old, there is basically no maintenance value.

2. South Bridge Chip ICH4 South Bridge Chip is mainly responsible for data processing and transmission of external devices. There were ICH 1, ICH2, ICH3 before ICH4, but it didn't support USB2.0. While ICH4 supported USB2.0, it was easy to distinguish them: there were 82801ab82801bb82801cb8280/kloc-0 on the South Bridge chip. Most of the phenomena after the South Bridge chip is broken are not bright, and some peripheral devices can't be used, such as IDE port and FDD port, or the South Bridge is broken. Because the chips of the North-South Bridge are expensive and the welding is special, it needs special BGA instruments to take them down, and the North-South Bridge can't be repaired in general maintenance points.

3. The 3.BIOS chip FWH solidifies some direct hardware information in the read-only memory. It is an important interface between software and hardware. When the system starts, some hardware information is first called from it, and its performance directly affects the compatibility of system hardware and software. For example, some early motherboards do not support hard disks larger than 20G, which can be solved by upgrading the BIOS. Some problems encountered in daily use that are incompatible with new equipment can also be solved by upgrading. If your motherboard shuts down suddenly and the CPU fan is still running, then you should first consider whether the BIOS chip is damaged.

4. The system clock generator CLK has a crystal oscillator in the middle of the motherboard, which will generate a series of high-frequency pulse waves. These original pulse waves are input into the clock generator chip, shaped and divided, and then distributed to various frequencies needed by the computer.

5. Super I/O interface chip I/O is generally located at the lower left or upper left of the motherboard. The main chips are Winbond and ITE, which are responsible for converting serial data of keyboard, mouse and serial port into parallel data. At the same time, the data of parallel port and floppy drive port are also processed. In our maintenance site, some peripherals, such as keyboard and mouse ports, printer ports, etc., can't be used, most of which are I/O chips, and sometimes even cause the phenomenon of not lighting up.

6. Sound card chip Because most motherboards now have integrated sound cards, and most of them are AC'97 sound card chips. Of course, there is CMI's 8738 sound card chip. If your integrated sound card has no sound, it is most likely that it is broken here.

Second, the main socket on the motherboard

1, CPU socket At present, all motherboards adopt socket series zero-pull sockets. Socket370 socket was used in P3 in the early days, now socket478 socket is used in P4, and socket423 socket is also used in P4 in the early days. In "Serious Warning: It is forbidden to publish other people's contact information in this forum. Server CPU such as Xeon uses socket603 socket. In "serious warning: it is forbidden to publish other people's contact information in this forum, and offenders will be sealed with ID". The constant change of CPU packaging format has made us fans send him a lot of money! However, I recently heard that in "Serious Warning: It is forbidden to publish other people's contact information in this forum, and offenders will be sealed with ID". The packaging format of the next generation CPU is still socket478, which is good news for DIYer who is constantly pursuing performance.

2. Memory bus sockets There are three kinds of memory we can see in the market: SDRAM, DDR SDRAM and RAMBUS. Due to the price reduction of DDR memory, SDRAM memory gradually faded out of the market. It uses 168 line socket, and there are two anti-blocking fractures in the middle and left side. DDR SDRAM has become the mainstream of the market because of its high cost performance. Using 184 line socket, there is only one anti-reverse plug fracture in the middle; RAMBUS memory performance is good, but the price has been high. In addition, support has been abandoned in "Serious Warning: It is forbidden to publish other people's contact information in this forum, and offenders will be sealed with ID", so its future is still just a suspense! Its socket adopts 184 RIMM socket, and there are two anti-reverse-insertion fractures in the middle. Some customers have repeatedly reported that the memory on the 845 motherboard is sometimes incomplete. This is because of Iin's serious warning: it is forbidden to publish other people's contact information in this forum, and offenders will be sealed with ID. The 845 series motherboard can only support 4 banks (a bank can be understood as one side of a memory stick). 845 series motherboards generally have three memory slots, and the second slot and the third slot * * share two banks. Therefore, if the banks inserted in the second and third slots are double-sided 256M, only one 256M can be identified.

3.AGP graphic bus socket It is located on the left side of the CPU socket and is brown. Its frequency is 64 MHz. In terms of speed, it is divided into AGP2X, and now most of them are AGP4X, and some motherboards already support AGP8X. Because different speeds require different voltages, some motherboards don't light up mainly because users plug the old AGP2X graphics card into the new AGP2X motherboard, thus burning out the AGP socket! Thankfully, some new motherboards have integrated automatic voltage regulators on the motherboard, which can automatically identify the voltage of the graphics card.

4. The 4.PCI bus socket is white, next to the AGP socket, which varies with the motherboard. Its frequency is 32 MHz. Insert more network cards, sound cards and other peripherals. 5, IDE device interface It is generally located below the motherboard. There are forty stitches and eighty threads. Two IDE ports are connected together, and sometimes one is green, which means IDE 1. Because the system first detected IDE 1, IDE 1 should be connected to the system boot hard disk. At present, most motherboards already support ATA 100, and some motherboards already support ATA 133, but higher-end motherboards already support serial ATA, which is a new technology with higher transmission speed and will be the mainstream technology of the next generation if the parallel transmission rate cannot be further improved. Having said so much in one breath, my mouth is thirsty. Let's take a look at our motherboard. Do you feel that they are much more familiar than before? Ha ha! It's time for us to say goodbye, that is, the motherboard. Today, I'll tell you another news about the motherboard. Our technical service center recently accepted a batch of boards for maintenance, and our engineers found it particularly difficult to maintain them. Later, it was pointed out by people familiar with the matter that the PCB edge of these motherboards had a hole the size of an eye of a needle. You can't tell the difference without looking carefully. Don't underestimate this small mouth, it is Lenovo's special mark on the scrapped motherboard! We have actually repaired many films, and we have to admire our technical level! This is not bragging! Therefore, everyone must be cautious when buying second-hand motherboards!

======================================================

cpu

======================================================

Mainly talking about frequency

1. Anyone who knows something about computers should be familiar with the word "frequency"! As the core of the machine, the frequency of CPU is of course important, because it can directly affect the performance of the machine. So, do you have a thorough understanding of CPU frequency? The so-called main frequency, that is, the clock frequency when the CPU works normally, theoretically speaking, the higher the main frequency of the CPU, the faster it will be, because the higher the frequency, the more instructions will be completed in a unit clock cycle, and the faster it will be. However, the same clock frequency does not mean the same speed because of the differences of various CPU internal structures (such as cache and instruction set). For example, PIII and Celeron, Thunderbird and Duron, Celeron and Duron, PIII and Thunderbird all have different performances under the same main frequency. At present, the main frequency of the mainstream CPU is above 600MHz, while P4 with the highest frequency (note that it is not the fastest) has reached 1.7GHz, and AMD's Thunderbird has reached 1.3GHz, and will continue to improve.

After the emergence of 486, the working frequency of CPU has been increasing, but some other devices of PC (such as cards, hard disks, etc. Limited by technology, it can't bear higher frequency, which limits the further improvement of CPU frequency. So the frequency doubling technology appeared, which can make the internal working frequency of CPU become a multiple of the external frequency, thus achieving the purpose of improving the main frequency by increasing the frequency doubling. So after 486, we came into contact with two new concepts-external frequency and frequency doubling. Their relationship with the main frequency is external frequency x frequency doubling = main frequency. The external frequency of a CPU is the same as FSB (Front Side Bus) that we often say today (note that the frequency is the same). At present, the CPU external frequencies on the market mainly include 66MHz (Celeron series), 100MHz (part PIII and part Thunderbird and all P4 and duron) and 133MHz (part PIII and part Thunderbird). It is worth mentioning that there are media reports that the external frequency of some CPU has reached 200MHz (Duron), 266MHz (Thunderbird) or even 400MHz(P4), which actually confuses the external frequency with the front-end bus. In fact, their external frequencies are still 100MHz and 133MHz, but due to the special process, the front-end bus can be used. However, from the definition of external frequency and frequency doubling, their external frequency has not changed. I hope everyone will pay attention to this. Today, the external frequency has not been greatly improved, but the frequency doubling technology has developed to a very high stage today. In the past, the frequency doubling could only reach 2-3 times, but now P4 and Thunderbird have reached more than 10 times. I really don't know if it will be higher in the future. At present, CPU frequency multipliers are usually locked before leaving the factory (except for some engineering samples), while external frequencies are not locked. Some CPUs, such as AMD's Duron and Thunderbird, can unlock their frequency multiplication by special means, but the CPU produced in "Serious Warning: It is strictly forbidden to publish other people's contact information in this forum, and offenders will be given ID titles" cannot.

Due to the continuous improvement of external frequency, it has gradually increased to the point where other equipment can't bear it, so the frequency division technology has appeared (in fact, this is the function of the motherboard North Bridge chip). Frequency division technology is to reduce the external frequency of CPU through the north bridge chip of the motherboard, and then provide it to cards, hard disks and other devices. In the early era of 66MHz external frequency, PCI equipment divided by 2 and AGP equipment did not divide by 2. Later, in the era of 100MHz external frequency, PCI devices divided by 3 and AGP devices by 2/3 (some 100MHz north bridge chips also supported PCI devices divided by 4); At present, Northbridge chips generally support 133MHz external frequency, that is, PCI devices divide by 4 and AGP devices divide by 2. In short, under the standard external frequency (66MHz, 100MHz, 133MHz), the Northbridge chip must make the PCI device work at 33MHz and the AGP device work at 66MHz, so that the chip can officially support this external frequency.

Finally, talk about the problem of CPU overclocking. In fact, CPU overclocking is to increase the main frequency of CPU by increasing the external frequency or frequency doubling, thus improving the performance of the whole system. Overclocking has a long history (in fact, it is only a few years), but it has really been loved by everyone since the production of Celeron series, of which Celeron is over 300 a 450 and over 366 550, which is still talked about by people. And they just increased the external frequency of Celeron CPU to 100 66MHz, thus increasing the main frequency of CPU. The early Duron overclocking is different from Celeron, which is to increase the frequency by breaking the frequency doubling lock and then increasing the frequency doubling. Generally speaking, overclocking is more stable than overclocking, because overclocking does not change the external frequency, so it will not affect the normal work of other equipment; However, if it exceeds the external frequency, it may encounter non-standard external frequencies such as 75MHz, 83MHz, 1 12MHz. In these cases, due to the limitation of frequency division technology, other devices can't work at normal frequency, which may lead to system instability, even hard disk data loss and serious possible damage. Therefore, the author warns everyone here: overclocking is good, but it is also very dangerous. Please overclocking carefully!

2. About overclocking If AMD's CPU wants to overclock, find out its frequency limit.

AMD recently released their brand-new Athlon XP processors with frequencies of XP 1500+, 1600+, 1700+ and 1800+. In order to counter the Pentium 4 processor in "Serious Warning: It is forbidden to publish other people's contact information in this forum, and offenders will be given ID titles", Athlon XP once again uses PR value (performance index) to name the processor, while Ahlon XP 1600+ means the same performance as Pentium 4 1600MHz.

Athlon XP adopts a new Palonmino core based on 0. 18 micron process, and the core area is increased from Thunderbird's 120mm2 to 128mm2. This packaging method has also become OPGA packaging similar to FC-PGA PentiumIII. AMD claims that after adopting the new core, the calorific value of Athlon XP will be 20% lower than that of Thunderbird with the same frequency. And lower heat dissipation naturally means stronger overclocking performance.

So, we decided to test the overclocking ability of Athlon XP. We choose Athlon XP 1600+ with better cost performance. It is much cheaper than 1800+, but the overclocking ability seems to be above 1900Mhz.

Athlon XP also has an L 1 bridge similar to Thunderbird, but it was cut off by the laser. In order to overclock, the L 1 bridge must be reconnected first. The specific bridge connection method can be found in the related articles of this site. Because the default voltage of the processor is 1.75v, a motherboard with voltage regulation function is needed to better exert the overclocking limit of the processor. We use Pan Ying 8K7A and 8KHA+ for comparison. Although 8K7A is inconvenient to adjust, its overclocking performance is better than the new 8KHA+.

After de-frequency, we first set the frequency multiplier to 6, and then set the external frequency to the highest. At 8K7A, we raised the maximum external frequency of the processor to 200MHz(400MHz DDR). Through the memory performance test under 200MHz external frequency, we can see that the memory bandwidth after overclocking has surpassed AMD760 chip by about 40%.

The test just now was only the result of air cooling, which was just the beginning. Next, we will test the overclocking limit of the processor in extreme cooling environment. After installing the water cooler. We adjust the voltage to 2.1v. And VDDR is adjusted to 2.9v

The test results are amazing. We finally stabilized the processor at the external frequency of 178MHz, and the frequency has reached 1873.89MHz.

Although we hoped to break through the mark of 1900MHz, we failed. At the same time, we also found that the motherboard is also very important for overclocking Athlon XP. Although 8KHA+ uses an updated chipset with better performance, it is not as good as its predecessor 8K7A in overclocking ability. The overclocking ability of the new core Athlon XP has also been verified.

======================================================

internal storage

======================================================

1. Basic knowledge of technical vocabulary of memory RAM

Cdram- cache D RAM- cache memory CVRAM- cache VRAM- cache video memory DRAM- dynamic RAM- dynamic memory edram- enhanced DRAM- enhanced dynamic memory EDRAM- extended date output RAM-extended data mode memory Edo SRAM- extended date output SRAM- external extended data mode static memory Edo VRAM- extended date output VRAM- External Extended Data Mode Video Memory FPM- Fast Page Mode-Fast Page Mode Fram- Ferroelectric RAM- Ferroelectric Memory S DRAM- Synchronous DRAM- Synchronous Dynamic Memory SRAM- Static RAM- Static Memory SVRAM- Synchronous VRAM- Synchronous

======================================================

display card

======================================================

For every DIY who pursues computer performance, the graphics card is undoubtedly the most important accessory. At the present stage of the rapid development of graphics card technology, although the number of optional graphics card chip manufacturers has decreased, the models of graphics cards based on the same manufacturer are very detailed and their performance is not the same. The complexity may be difficult for even professionals to describe. Users are more exposed to some professional data when choosing a graphics card, such as simple points such as chip core frequency and video memory frequency, and complex points such as pixel filling rate and video memory bandwidth. Various graphics card brands have also mentioned this point in their respective graphics card descriptions, but for some aspects, they may deliberately ignore some details and only provide those dazzling superior data, which is unfair to users who do not fully understand it. Here I mainly introduce the performance parameters of the graphics card and how to determine the performance of the graphics card according to these parameters. I hope you can better choose the products you need next time you buy a graphics card.

First of all, let's understand what is the most important indicator for a graphics card. Excluding the graphics card plays a decisive role in the display performance of the whole system including CPU, memory, motherboard and driver software. Such a platform must deal with a large number of geometric operations, such as T &；; L that is, the light source and deformation processing technology need strong floating-point operation and occupy the bandwidth of main memory. If the graphics card has no hardware T&; L function, this part of the task will all fall on the working group composed of CPU, memory and motherboard. When calculating the frame size of graphics, vertices and textures are transmitted to 3D cards through buses (PCI or AGP 1x, 2x, 4x).

At this time, the faster the platform, the more frames will be transmitted. These external factors that affect the performance of graphics cards are not what I want to talk about today. The most important thing about the graphics card itself is the pixel filling rate provided by its chip and its memory bandwidth. Let's look at them:

The maximum pixel filling rate is the 3D clock multiplied by the number of rendering processes. For example, NVIDIA's GeForce 2 GTS chip has a core frequency of 200 MHz and has four rendering pipelines, each of which contains two texture units. Then its filling rate is 4x2 pixels x2 billion/sec =16 billion pixels/sec. The pixels here make up the picture we see on the display screen. At the resolution of 800x600, a single * * has 800x600 = 480,000 pixels, and so on, with the resolution of 1024x768 and 1024 x768 = 786432 pixels. We often set the resolution when playing games and using some graphics software. The higher the resolution, the more pixels the display chip renders, so the filling rate is of great significance to measure the performance of a graphics card. Just now, we calculated that the filling rate of GTS is 65.438+0.6 billion pixels per second. Let's look at MX200 again. Its standard core frequency is 175, and there are only two rendering pipelines, so its filling rate is 2x2 pixels x 65438+7500 million pixels/sec = 700 million pixels/sec, which is an important reason why its performance is half that of GTS. As we all know, the filling rate depends on the display chip. At present, as long as you buy a graphics card from a regular manufacturer, there will be no mechanism on the chip, and you can get what you want for a penny, but the memory I will focus on below is not so transparent.

When we buy graphics cards, we can often see the parameters of memory, mainly the speed of memory, in nanoseconds; The working frequency of the video memory is the data bit width of the MHz memory, in bits. The speed of video memory here determines its working frequency. For example, the standard frequency of -7.5ns video memory can reach 133MHz, and the standard frequency of -5ns video memory can reach 200MHz. However, sometimes the working frequency of the video memory on the graphics card is not proportional to its speed. For example, Geforce3 generally uses 3.8ns DDR memory, and the standard should be 263MHz. Because it is DDRAM, the standard frequency is 526MHz, and we know that the standard frequency of Geforce3 memory is 460MHz, which reserves a lot of overclocking space for users. And those whose memory speed is marked as -7ns should be 143MHz, but the default working frequency is166 MHz; Some video memory speed is marked as -4.5ns, but it can't reach 222MHz. So when you buy a graphics card, you can't just look at the speed value marked on the memory stick. Be sure to ask the default working frequency of memory.

The bit width of video memory is a parameter that is often ignored by users, but its importance even exceeds the working frequency of video memory, because the bit width determines the bandwidth of video memory, and the bandwidth of video memory has become a bottleneck restricting the performance of graphics cards. The data exchange speed between display chip and memory is the bandwidth of memory. The processing power of a single chip is very strong, but if the bandwidth of the memory is not high, the memory will limit the chip to achieve the designed processing power. We exceed the memory frequency of Geforce3 to 500MHz, and the bandwidth is as high as 8GB/s, but in some complex graphics environments, the processing speed will be affected by the lack of memory bandwidth. During the work of graphics card, Z buffer, frame buffer and texture buffer will occupy a lot of memory bandwidth resources. Bandwidth is the standard of data transmission between 3D chip and local storage. At this time, the memory capacity is not important, and it will not affect the bandwidth. For graphics cards with the same memory bandwidth, there is little difference in memory performance between 64MB and 32MB. Because the bottleneck of the system at this time is the video memory bandwidth, when a large number of pixels are rendered, the insufficient video memory bandwidth will cause data transmission congestion, which will lead to the display chip waiting and affect the speed. At present, the main memory is divided into 64 bits and 128 bits. At the same operating frequency, the bandwidth of 64-bit memory is only half that of 128 bits. The calculation method of video memory bandwidth is bandwidth = working frequency x data bit width /8. That's why the performance of Geforce2 MX200(64-bit SDR) is far less than that of Geforce2 MX400( 128-bit SDR). 64-bit video memory is avoided in many video card advertisements. Users should ask this question clearly when purchasing video cards. At the same frequency, the performance of 16M 128bit may be better than that of 32M 64bit, because the bandwidth of video memory is too important for the performance of graphics cards. In order to improve the performance of graphics cards in the future, it is imperative to solve the bandwidth problem of video memory.

Due to the extremely low price of memory chips at this stage, many manufacturers began to make a fuss about memory capacity. There are more and more graphics cards with 64MB memory. However, it seems that a Geforce2 MX400 uses 64MB of memory, but it uses 64-bit memory instead of the standard 128-bit memory of MX400, and its performance will not be improved. Personally, I think this method has the component of deceiving users, attracting users with memory capacity, but not telling users the truth about performance. Users have to spend more money than ordinary 32MB graphics cards to buy products that he deliberately reduces performance to cater to the market. But for this manufacturer, the cost is indeed higher, which is not cost-effective for both parties. This kind of market means is too failure, mainly because the planners don't put users first for their sake, they only play with market means, and finally they suffer.

======================================================

Integrated sound card

======================================================

Integration technology is the development trend of PC. At present, some motherboards on the market have fully demonstrated this feature, and those motherboards that integrate graphics cards and sound cards are more popular (among which integrated sound cards are the most common). However, due to misunderstanding, many DIYer are not interested in integrated sound cards, and even equate "integrated sound cards" with "inferior sound cards" or simply call them "garbage". Is that really the case?

I. What is AC'97?

Since VIA put forward the concept of "AC'97 sound card" in its MVP3 motherboard chip, we have often seen it in various motherboard manuals, and finally got the term "AC'97 soft sound card". Later, "AC'97" simply became synonymous with soft sound card. But if you look at the technical data of some high-end sound cards, you will be surprised to find that "the card adopts AC'97 standard". Is the high-end sound card also a soft sound card? To know this secret, we must first understand the AC'97 specification (or standard).

1.AC'97 1996. In June, five well-known authoritative software and hardware companies in the PC field put forward a new idea of the sound source structure of chip-level PC, which is the "AC' 97" standard (Audio Codec97) that we now see.

2. What is AC'97 specification? In the early ISA sound card, due to low integration, a large number of components were scattered on the sound card. Later, with the development of technology and technology, a single sound card appeared, and all the functions of the sound card can be completed with only one chip. However, because the digital part and analog part of the sound card are integrated together, it is difficult to reduce the influence of electromagnetic interference on the analog part, which makes the signal-to-noise ratio of ISA sound card not ideal.

The AC'97 standard puts forward a "double chip" structure, that is, the digital part and the analog part of the sound card are separated, and each part uses a single chip. AC'97 standard combines the advantages of digital processing and analog processing. On the one hand, it reduces the noise that may occur when analog lines are converted to digital lines, and creates a purer sound quality. On the other hand, the cost can be further reduced after the sound processing is integrated into the chipset.

3. After AC' 97 was applied to 1997, most PCI sound cards appeared in the market have begun to conform to AC' 97 specification, separating analog circuits from sound card chips and becoming a small chip called "audio codec", while the main chip of sound card, that is, the digital part, has become a large chip called "digital control" (digital signal controller).

Therefore, AC'97 is not a sound card, but a standard.

Second, the mainstream integrated sound card-soft sound card

Through the above introduction, we know that a sound card conforming to AC'97 standard has two chips: "audio codec" and "digital control". So what does the so-called "AC'97 soft sound card" mean? It turns out that VIA and IN's "serious warning: it is forbidden to publish other people's contact information in this forum, and offenders will be sealed with ID" have successively added the function of sound card to the South Bridge chip of the motherboard chipset. The software simulates the sound card to complete the function of the main chip on the general sound card, and the audio output is given to the "audio codec" chip. So there is no such big "digital control" chip on this kind of motherboard, only a small "audio codec" chip. Let's take an innovative soundblaster PCI128 Digital and an i8 15E motherboard as examples to see the difference between ordinary sound cards and AC'97 soft sound cards.

We can easily find the larger main chip-"digital control" and the smaller "audio codec" on the sound card. The model of "digital control" chip of Sound Blaster PCI 128 Digital (marked as 1 in Figure) is "CT5880". As the core processing chip on the sound card, "numerical control" is equivalent to the CPU in the computer, which needs to complete most sound card functions, such as WAV playback, MIDI synthesis, sound processing and so on. The main technical parameters of the sound card depend on it, which is an important basis for determining the grade of the sound card. Not far from the "digital control" is the "audio codec" chip. Although small, it can accomplish more functions than ordinary DAC (digital-to-analog conversion) chips, including ADC (analog-to-digital conversion) that converts analog signals into digital signals, mixed input and output of multiple analog signals and other functions, similar to the functions of digital encoder/decoder and preamplifier in audio. The "audio codec" here is the STAC9708 chip of Sigma's "serious warning: it is forbidden to publish other people's contact information in this forum, and offenders will be sealed with ID". According to AC'97 standard, the pins of different "audio codec" chips are compatible and can be replaced in principle.

Because the soft sound card does not have a "numerical control" chip, but uses software simulation, the CPU occupancy rate is higher than that of the general sound card. If the CPU speed can't meet the requirements or the driver software has problems, it is easy to produce sonic booms, which will affect the sound quality.

Third, the "substitute" in the integrated sound card-hard sound card

Because the soft sound card has many shortcomings, some motherboard manufacturers have thought of another method of integrating sound cards-moving the "digital control" chip on the ordinary sound card to the motherboard, that is, integrating the chip and auxiliary circuits onto the motherboard (this "integrated sound card" is actually a sound card in the traditional sense). Compared with the separate motherboard and sound card, the cost is reduced a lot, and the sound effect is similar to that of an independent sound card in theory. You can find a large "digital control" chip near the PCI slot of this integrated hard sound card motherboard.

At present, there are more and more motherboards integrating hard sound cards. The common chips are:

1.CT5880

CT5880 is a main product of innovative company for low-end market. The sound card made of this chip is "Sound Blaster PCI128 Digital". Support 128 polyphony and multi-timbre, 16 MIDI channel, support 4 channels; Support Microsoft DirectSound, DirectSound 3D and their derived standards. As far as the performance of CT5880 is concerned, it can meet the needs of most users who are not very demanding on sound. At present, CT5880 is the most widely used sound chip integrated on the motherboard.

2.CMI8738

CMI8738 is a product of Hua Xun Electronics (C-Media) in Taiwan Province Province. 1999 developed a 4-channel sound chip CMI8738/4CH, which not only has 3D positioning function, but also provides digital optical fiber interface to support home theater system. On the basis of CMI8738/4CH, Hua Xun introduced a 6-channel CMI8738/6CH sound chip. This chip not only has all the functions of CMI8738/4CH, but also adds six output functions. It can be matched with 6 channels of 5. 1 or 4. 1.

References:

/question/ 17 178896 . html？ fr=qrl3