Junior articles, the time of overclocking history, the CPU uses unified main design, the frequency of the central processor is the frequency, chipset, memory, and cache of the motherboard, so there is no multiplication on the motherboard. Line, each motherboard is only suitable for a CPU. Increasing the frequency of the crystal oscillator on the motherboard to achieve overclocking, the earliest overclock record is the Motorola chip of Amiga 500 from 9MHz to 12 MHz, Intel 80286 from 8MHz to 12MHz. Later, Intel launched a multiplier CPU. Some motherboards started compatible with more than one chip. At that time, there was no formal documentation. We can only rely on experience to determine which needle is multiplier jumper, welding method. To overclock, just like the most recent K7. The first leap forward in the history of overclocking is the emergence of Pentium chips, almost all Pentium 75 can exceed 90MHz, and the overclocking revolution began to fully carry out around the world. Subsequent 133 super 166,166 super 200, 233mmmx super 266 can only increase one to two, but the highest is not over 4. When the history of the wheel moves into the Celeri era, the most brilliant overclocking era is finally coming, the first is the cached Covington Celeron 266 super 400, then the biggest breakthrough in overclocking history - 300A over 450MHz, it increases the performance of the CPU 50% of the whole! Moreover, the probability of overclocking is very high, and the average of two CPUs can be super. In the race, the last one capable of stabilizing overclocking is 366, 366 over 550MHz performance gains by 183MHz, which is the highest realm that is air cooled, and does not add voltage overclocking. Today, the most potential overclocking chip is PII-500E and PII-600E, which can increase the frequency of 200MHz. Second, the bus speed and frequency frequency
486DX2 is the first multiplier CPU, 2 indicates two times, the frequency = outer frequency × frequency, the fields of 486 DX2 / 50, 66, 8 are 25, 33, 40 MHz, chipset, and memory Work at the 1/2 clock of the frequency. Affected by sound card, graphics card, hard disk, optical drive, memory, etc., we can't increase influence in infinity, only multiplier. Today's CPU has reached the frequency of 10X, meaning that the speed of the processor is ten times that of the bus. When the overclocking technique is widely circulated, the random transformation of the outer frequency and the clogging has become the root cause of bad dealers to make counterfeit goods, and they use the low frequency CPU to high frequency and modify the identification code (REM-ARK) of the chip surface. In order to prevent such problems, Intel started from Pentium MMX, adding multiplier lock in the CPU, we can no longer change the multiplier, can only work hard from the outer frequency. Since the frequency of foreign frequency setting is limited, overclocking is difficult, such as: 300A standard overclocking is 100 × 4.5 = 450, you can't overcome it into 150 × 3 = 450. Recently, motherboard manufacturers have launched a linear overclocking product, effectively ease of contradictions, such as 83 options between 100MHz and 183MHz, ie, 1 MHz as a frequency transform reference. Ultra-frequency most dangerous is affecting the frequency of PCI / AGP, and the standard frequency of PCI is 33MHz. The standard frequency of AGP is 66MHz. If the standard frequency is exceeded, it is likely to cause hardware damage. For overclocker, I offer three most secure extraordale: 66, 100, 133MHz, and their PCI / AGP frequencies are 33 / 66MHz, which will never have hardware problems. Recently, some new motherboards use 2/3 division and 1/4 frequency division design, as long as PCI and AGP are running at standard frequencies, how many foreign frequencies have nothing to do. Third, the basic principle of overclocking CPU is the core of computer hardware, and its speed is doubled every 18 months. In this endless upgrade, we have to exhaust your own precious energy and money, but buy it back is Soon, it is necessary to turn around, and overclocking, how many of this abnormal reincarnation process, let us have more time to use the computer to do what you want to do. At the same time, through overclocking, you can understand the hardware common sense of the computer, master the inherent law of computer hardware. The CPU comes out of the production line and must be tested to determine its limit frequency, then determine the nominal frequency of its normal operation, and will enter the market after the flag is hit. For safety, the limit frequency must be highly nominal frequencies and maintain a certain amount of space for no test. What we have to do is in terms of stability, creative conditions try to make CPUs under its ultimate frequency and let it play the greatest effect. The CPU is a central processor integrated with a large number of transistors, which integrates such a large amount of components in a small range, will have a huge amount of heat during operation, and the high-calorie generation of high calories makes the CPU itself Increase, the generated interference signal will seriously affect the quality of normal signal transmission. On the other hand, high-calorie is also the main factor in generating electron migration, affecting the life of the CPU. Therefore, you must solve the cooling problem of the CPU. In addition, individual differences are also the main factors that affect CPU limit operating frequency. Individual differences are caused by materials, process and production lines in the process of production, and some CPUs have special superclocity. Therefore, we must obtain the ideal over-frequency frequency, the selected block, and reduce the temperature of the CPU work is the main route of our overclocking success.
Fourth, the CPU selection Celeron CPU adopts. 25 production process, built-in L1 and synchronous L2 cache, has similar integers and floating point calculations with Pentium II, it is the most practical Pentium II CPU, and the price is low. Due to the use of 66MHz, there is excellent overclocking ability, it is easy to reach or exceed 100MHz, and the protagonist of this experiment is well-deserved. V. Primary overclocking primary articles will introduce several of the most commonly used and simplest CPU overclocking methods (frequency conversion method, selection method, cooling method, fan method, heat sink method and thermally conductive silicone]. No newcomers are complex and dangerous, even if they are not overclocking, these methods can improve the stability of the CPU. 1. The operating frequency within the frequency conversion method CPU is working according to the external frequency by multiplier. For example, Celeron 300 uses 66MHz out-of-frequency, multiplying 4.5 multiplier, due to Celeron CPU The multiplier cannot be changed (locked by Intel), so the most basic overclocking method is to improve the outer frequency of the CPU to improve the internal operating frequency. There are three specific methods: BIOS settings, motherboards (or transfer cards) DIP jumpers and stickers. BIOS setting: Now there is a set of CPU parameters in the BIOS. When startup, hold down the DEL key, enter the "CPU setting" in the BIOS, change the external frequency frequency of the CPU, from 66MHz Set to 75, 83 or 100MHz, restart after saving. If the computer can display new frequencies and stabilize, then overclocking is successful. For the initial setting of Celeron 300a, it shows that the race 300a is changed to 100MHz. If the machine is unstable or cannot be started after the overclocking, when it is necessary to restore the original state, you can enter the BIOS setting to the original setting. If you can't succeed, please find the motherboard to clear the CMOS jumper, insert it, start, start, You can clear the original setting (remember to restore the position of the jumper when using it). If the above method does not care, the CPU will be unplugged and then plugged into it. Motherboard (or transfer card) DIP setting: Find the DIP jumper position corresponding to different outer frequencies in the motherboard (or transfer card) manual, step-by-adjustment of its frequency, if it is unstable or unable to start, turn the DIP after shutting down Jumper restore. Sticker method: Although the sticker method is exactly the same as the above two methods, it will be described in the following intermediate articles due to the actual operation. The general motherboard provides an outer frequency of 66, 75, 83, 100, 112, 124 and 133MHz, etc. It also has linear foreign versus from 66 to 200 MHz, which is more convenient for overclocking testing. It should be noted that when the external frequency change of the motherboard, the operating frequency of the motherboard PCI and AGP is also changing, so it is necessary to consider whether other components such as hard drives, graphics cards, and sound cards can work over higher frequencies, when foreign frequency exceeds When 100 MHz, the PCI can be selected in quadritical state, and the AGP is selected in the triple frequency state. 2. The selection method is strictly in strict, and the choice method itself is not overclocking. As can be seen from the above, the CPU works is ideal when working at 100MHz, so it is necessary to pay attention to "excellent varieties" that can go up to 100MHz. The same is the CPU, the 370 interface of the 370 interface is better than Slot 1, and Slot 1's 300A can only run to 100MHz's outer frequency, that is, 100 × 4.5 = 450MHz, and 370 interface 300A can run Higher. Similarly, the 333, 366 of the 370 interface is also a good choice.
Celeron 400, 433, 466 These grades of CPU multiplier have reached 6, 6.5, 7, which can be almost no chances of 100MHz outgoing, so they are not recommended. This experiment is conducted for classifier 300a for the 370 interface, while other CPUs can also refer to such overclocking methods. 3. The radiator method has now been 100MHz. Then the heat dissipation and cooling problems must be seriously considered. This is the key to normal overclocking. Usually, most of our overclocking failure is no treatment of CPU high temperature issues. The CPU of the retail 370 interface is usually not equipped with a fixing device, so after purchasing the CPU, it is best to carefully select a highly efficient heat dissipation device. It is judged whether or not the heat dissipation device is excellent, the simplest method is to replace different heat sinks while testing the temperature of the CPU kernel (not a heat sink), the lower the temperature, the heat dissipation device. This measurement method is equally applicable to other heat dissipation methods later. The shape and material of the heat sink have a great influence on the heat dissipation effect, the larger the surface area, the higher the thermal conductivity, the better the heat dissipation, so it is necessary to use the fork to refer to a large radiator. Although copper is a great heat dissipation material (the thermal conductivity of copper is better than aluminum), it is easy to oxidize and deformed, so it is rarely seen in the market that most of the heat sink is aluminum material. The temperature of the test kernel is more troublesome. The temperature measurement heads provided on the motherboard is usually used to measure the temperature of the heat sink, and only the temperature measurement head is buried next to the metal sheet in the middle of the class, and the metal sheet can accurately display the kernel. temperature. The most common heat sink is actually used, but it is not possible to apply to a CPU that has a greater heat after overclocking. If such a heat sink is selected, the Celeron 300A is overclocked to 450MHz (the outer frequency is up to 66MHz to 100MHz, CPU voltage 2V), when the ambient temperature is 14 degrees, the temperature of the core reached 35 degrees, to the summer, the kernel temperature will More than 60 degrees, inevitably causing a deadline to fail. If you want to overclock, this heat sink is definitely not. A hot pin-shaped fins called the Arctic wind After using a copper sheet, folded in the original heat sink to increase the surface area, the weight of the finished heat sink is light, and the effect is good. In addition, there is a high-grade heat sink, its unique turbukin structure with the ball bearing fan is simply an artwork! A slightly high price cannot stop its charm. Using radiator heat is a way to be the most straightforward and safest, most friends will not let their beloved CPUs run in a high fever. Now just move the finger, take the heat sink on the original CPU, spend more than dozens of dollars to install a beautiful high-efficiency radiator, maybe your CPU will work immediately on a higher frequency! 4. The fan method is usually divided into two types of axial fan and turbine fan. Most of the computer use is axial fan. The fan is an indispensable component of the radiator. Because the intramids inside the computer chassis are relatively closed, the natural cooling mode of the cooling heat sink cannot meet the requirements, and the heat sink is highly efficient and simple. The fan is different, the wind speed and the air volume are also different. The effect of heat dissipation with the heat sink is also very different, and at the same time, the working noise of various fans is different. Let's take a look at the actual effects of different fans.
It is a large fan, thick body, high speed and low noise in the market. It is its unique feature, and even more delighted. In exchange for such a fan, you can use the CPU kernel. Temperature decreases! Usually the intermediate portion of the axial fan is not blown down, so that the cooling heat sink is the highest heat, the effect is not good, if it can be installed on the ordinary heat sink to the ordinary heat sink, each The biggest wind of the fan is just in the center of the heat sink, the effect is better! The heat sink using a double fan can reduce the temperature! It's a great idea. If the heat sink is used directly, the effect is better. Many overclocking enthusiasts like to increase the speed of the speed by increasing the fan voltage. After the fan is high, it is indeed the effect of further cooling. However, since the fan itself is an inductive load, the voltage increase is not more than the wind speed, and the power consumption is increased, so the wind is not suitable for overclocking. 5. The thermally conductive silicone fat can not guarantee complete contact even if it looks flat, and therefore affects thermal conductivity. The thermally conductive silicone is a white or gray insulated viscous object. It has good thermal conductivity that applies it on two contact surfaces, which can play a good thermal conductivity, greatly reduce heat, so It is widely used in various fields that need heat. On the computer market, a small box is bought back, and it is thin and uniformly applied to the metal plate of the Celeron CPU, and it also applies a layer in place that is in contact with the CPU. No need, Then, the heat sink is buckled on the CPU, press two down, allowing it to sufficiently, and finally buckle the clamp. The preliminary method of overclocking is approximately these kinds. It usually uses these methods to solve the CPU cooling problem well, and 450MHz is easy to take. In addition, the choice of motherboard and transfer card also pays attention to that overclocking stability of the brand-name motherboard may not be consistent with its famous gas, sometimes a very ordinary motherboard can be super, but not on the famous brand motherboard, good in such The difference is not obvious. If I use the promotion motherboard, although the CPU overclocking is not behind, the cooperation with the memory is very bad, only Kingmax and ordinary LG's memory 133MHz, and Hy and gold bars are connected to 112MHz, if not Note that such motherboard and memory configuration are selected, it is easy to cause the CPU that cannot be overclocking, which is an aspect that everyone must pay attention to. In addition, the quality of the ATX power supply is also critical, and inferior power is unable to provide pure 3.3V power, which will also affect the work of memory and hard drives. Intermediate articles in this chapter, we need some courage and hands-on ability, but please feel free, just do it according to reasonable ways, things are not so terrible. 1. Pipeline method is not so easy to change the external frequency for some of the main boards that do not provide external frequency jumpers or set, usually in the early Slot 1 race and the motherboard will encounter Such a problem, at this time, as long as the CPU gold inserted the B21 feet of the CPU, it will become a 100MHz of the outer frequency of the transparent film. However, the current transfer card has a conversion of 66 / 100MHz frequency. 2. Adjust the CPU voltage method to improve the working voltage of the CPU to make the signal level of the transmission than the noise of itself, which is conducive to the stability of the work, and in practice, the increase in CPU voltage is almost overclocking. There are three ways to adjust the CPU voltage: motherboard BIOS adjustment, transfer card jumper and patchwork. Motherboard BIOS adjustment: The BIOS of some motherboards contains the adjustment of the CPU voltage. As long as the CPU parameter setting in the BIOS in the startup, the CPU voltage can be easily changed, and the range of general adjustments is between 1.5V ~ 2.3V. .
When the CPU processes heat dissipation in the 2V voltage, it is still unable to stabilize, and it is necessary to improve the voltage of the CPU. Normally increase the voltage to 2.2V, many CPUs that cannot be stabilized at 100 MHz will be tamed. Card Jumper Method: The level of VID0, 1, 2, 3, 4 of the CPU can control the size of the motherboard to provide the voltage of the CPU. When using the transfer card, you can lead these five The foot is in the form of jumper, and it is set according to the needs of the user. The specific voltage / jumper is related to the instructions of the transfer card. Piping method: The feet method is the same as changing the Frequency adjustment of the CPU (using the race of Slot 1), as long as the corresponding gold finger on the CPU is attached, the motherboard will provide the corresponding voltage for the CPU. The correspondence is as follows. After adjusting the CPU voltage, the working temperature of the CPU has also increased, so it is necessary to repeatedly test, find the most stable overclocking method. When it is usually used, the temperature of the CPU has also increased after increasing the working voltage, while the pressurization is often more effective when using the water cooling described later. It should be noted that the working voltage of the CPU cannot be too high. When the air-cooled heat is hot, the voltage should not exceed 2.4V, and the voltage should not exceed 2.8V when using other cooling methods. 3. Surface Polishing The above has been introduced to replace the heat sink and the fan to reach the kernel temperature, and now take a look at the metal cover on the Celeron CPU. It is a piece of thick copper plate plated with nickel, CPU works. It is transmitted from here. The surface of the copper plate always distributes some small pits, not flat, now doing is re-grinding. First, according to the situation, the number of sandpapers used, the higher the number of sandpaper, the finer, and 3 of the 600, 800 and 1000 can be selected to throw off the outer layer, and then use 1, 3, 5 and 7 gold. 5 sheets of sandpaper were polished, put on the copper, and the reason why it bought so much was because the utilization rate of fine sandpaper was very low, and it was useless. Then, the CCI CPU is loaded with an anti-static film (plastic bag of the card, the hard disk), and the plastic film of the copper plate is excavated, see. When polished, keep a uniform speed and strength, not fast, because the copper is very soft, so as not to throw the copper plate into a spherical surface. In order to avoid polishing uneven problems, I recommend a way: put the glass pad on the table, lay the sandpaper, put the CPU head down on the sandpaper, and the other hand slowly takes away the sandpaper, and the CPU does not move The direction of each extraction should be consistent. When the first fine sandpaper is changed, the CPU is rotated 90 degrees until it completely removes the friction marks left in the previous process, so that the surface of the CPU can be polished to a mirror. Apply silicone stickers on the heat sink, now look at the situation of cooling, actually falling so much. The methods described above are all kinds of air-cooled cooling, and there are several completely different heat dissipation methods. 4. Water-cooled water is a thermal capacity of fluid substances, low cost, and after the heat-dissipated object is put into water, heat will be rapidly absorbed. What we have to do is to dissipate the CPU using the cyclic water substitute. The advantage of using water-cooled is to reduce the core temperature of the CPU to approaching the room temperature! Water-cooled systems include several parts: heat exchanger, circulatory system (inlet, water pipe), water tank, water pump and water, can also increase the heat dissipation structure as needed. The heat exchanger is the core of the entire water-cooled system. The quality of the water-cooled system is also completely determined by it, which is also the most clever part of the whole system. The circulation system feeds water into and discharges the heat exchanger, and the other end of the water pipe is connected to the water pump. The water pump is placed in a water storage tank or other structure of the water tank, and the water pipe will be researched into the water tank. If necessary, the hot water in the water supply is lowered by the heat dissipation system to room temperature and then discharge the back water tank.
Another key to success or failure is how to prevent leakage, so the heat exchanger should be completely sealed, and each joint cannot leak. The following integrated water cooler can effectively prevent leakage problems. The entire heat exchanger and circulation system use a completely sealed air-conditioned copper tube, and the copper pipe plate is wound around the ordinary heat sink, and the gap is filled with tin, and the copper pipe is reached out and then the plastic pipe is applied to the water pump or water back water tank. . The advantages of this method are safe, but the pipeline is fine, and the heat exchange is not easy to overcome. A large amount of ice is put into a large amount of ice to reduce the temperature of the water (the lowest is also near zero), so that the temperature of the CPU can be greatly reduced, but it should be noted that the aggregate phenomenon occurs when the temperature is below room temperature, if the aggregate dropped to the computer The circuit board means destruction. Another trouble of the water cooler is the problem of water, and the water will deteriorate, so the appropriate amount of preservative is to be added. Since the water cooling system is used, the temperature of the core is only about 3 degrees of water temperature, and the effect is very obvious. Now your CPU has taken 124 × 4.5 = 558MHz. 5. The cooling film method is powerful, but still cannot reduce the core temperature of the CPU to zero level, and cooling is the best way to improve the CPU operating frequency. The emergence of refrigerant will change this. The refrigeration sheet is an array of crystal diode structures, and there is a ceramic sheet thereof to conduct heat, and the porcelain surface is between 30 × 30 mm to 40 × 40 mm. After the power is applied, one of the porcelain is quickly refined, and the other side is quickly heated, and when the electrode is reversed, cold, hot face is adjusted. The cold surface is attached to the copper or kernel of the CPU, and then the heat sink or the water cooler is attached to the heat end. The cooling sheet can reduce the temperature of the CPU to room temperature, which is a comparative implied method. However, due to lower than room temperature, CPU and rear adapter card, the motherboard will generate aggregate phenomena. The water vapor in the air will condense on the surface of the CPU, and the time will flow on the circuit board to form a short circuit, so it is strict Anticitation treatment. Let's introduce a relatively simple and practical anti-integrated method: in the back of the transfer card (white rod solid, melting after heating, used for bonding, fixing wire, no conductivity) to bring the route near CPU All the plates are applied, and the thickness is about 5 mm. See that the part of the transfer card CPU zero pressure socket is carefully filled with glue (reducing the space to prevent aggregate), and pay attention to the socket also stick, see. On the wiring board around the CPU socket, install the CPU, and attach the double-sided adhesive on the CPU substrate, and then cover the entire adapter card and CPU with the fresh film for the refrigerator. The portion of the double-sided adhesive is tightened, the excess portion and the film on the CPU copper plate are cut off, and the purpose is to isolate the CPU, the transfer card with the air. The heat-conductive silicone is applied to the CPU copper plate and the surrounding substrate, and the purpose is also isolated from silica. The heat gum is also applied to the cooling tab (frequently aggregated here) and placed on the CPU, and the heat-conductive silicone is applied to the water cooler. If there is a place to penetrate the film, it is necessary to fill it with hot glue. After the multi-layer treatment can isolate the CPU with air, no problem for a long time. The working voltage of the cooling piece is usually 12 to 17V, the larger the rated current, the larger the cooling amount, but it cannot exceed its rated power. Depending on the size of the power, the price is usually 40 to 100 yuan. The hot end of the refrigeration piece should be connected to the heat sink, otherwise there is a risk of burning after power.
The temperature difference of the two surfaces in the most ideal environment can reach 70 degrees, but because the efficiency of the refrigeration tablets is not high, the single-chip refrigeration tablets can't work with the CPU, CPU work, the temperature difference is only 3 ~ 4 Degree, the effect is not ideal, so it can be considered in the form of 1 2, i.e., on the CPU, the hot end is attached to a copper plate, and two pieces of refrigeration sheets are added side by side in the copper plate, and their hot ends are then refined. This can generate low temperatures of zero 20 to 30 degrees! The power consumption of the refrigeration piece is not small, and one AT or ATX power supply is generally required. After using the refrigerated film method, this race has been able to run in 133MHz for a long time. The effect is very obvious! 6. The practice of memory voltage adjustment is to reduce the core of CPU temperature. There is also a lot of factors that affect CPU overclocking. When the outer frequency increases to 133MHz, the components in the computer will have a severe test, of which Memory requirements are higher, and some memory cannot be stabilized in the state of CAS3, and the appropriate method of properly increasing the operating voltage is a common method of stabilizing memory. The memory of the memory is directly taken from 3.3V of the power supply, so it is possible to increase the power supply 3.3V voltage output. ATX's power supply, 3.3V voltage is adjusted by KA431, and the resistance of the two 5-ring metal film resistance around KA431 changes the reference determination voltage, and the voltage output of 3.3V is changed. Figure 9. Histed 3.3V voltage has a limit, usually between 3.3 to 3.9V. The content of advanced advanced presentation usually has certain dangers or difficulty, requiring courage, money and experience, is not going to tiger points, it is rare to the tiger (giving a child, don't live a wolf). 1. The copper plate on the top of the Cai Yang CPU is used to protect the kernel, but the calorie of the kernel must also pass through the copper plate to the heat sink. Although the nickel on the copper plate is throwing off the polishing method, the thermal conduction is more smooth, but the total heat transfer efficiency will be reduced for each additional contact. For higher frequencies, it is now a time to show it. The copper plate on the CPU is soldered on the substrate. It is very effective with the hot air gun, but I don't know if I will burn the core, so I have to change the method. That is the same as the polished method, this time you have to pack the CPU, use a paper cutter to cut along the CPU surface and slowly cut along the edge of the copper plate, and then cross the surface of the substrate, the solder is treated, and the soldering is cleaned. Returning with a small pen slam along the edge of the copper plate, it should be able to take a copper plate, which is handled about 30 minutes. Now let's see the inside of Celean CPU! The kernel of the class includes two parts: silicon wafers and ceramic composite substrates. We now see just electroplated copper ceramic substrates, and true silicon wafers have much fragile, so don't want to move the idea of ceramic slices. Now that we have to do is to install the heat sink directly on the kernel! Since the core and CPU substrate do not focus on the same plane, it is directly desirable to use the heat sink to the side, so the part around the heat sink bottom should be detached to retain the portion of the central and kernel contact. This is a hard work, which is necessary to check the location of the kernel, but also fill the extra part, but these are nothing more than 3 degrees. Some CPUs can work with wind cooling and heat. The CPU copper plate is not guaranteed, and it is also possible to scratch the CPU kernel, so it is necessary to force.
2. This method is very similar to water cooling. Different is that it uses a refrigerator refrigeration technology to increase the heat exchanger (evaporator) of the refrigerant-ammonia in the CPU. Convert to a gas, absorb the heat of the CPU, and then re-coagulate the heat on the condenser to release heat. The compressor refrigeration method can reduce the core temperature of the CPU to 30 to 40 degrees. This kind of equipment is currently sold in the form of finished products, and the difficulty of self-acting is quite large. The temperature of the CPU can be controlled at zero by using a cooling box method. 3. The soaking method is similar to the cooling box method, just putting the main board, CPU and graphics, etc., in an insulated coolant, cooling through the refrigerator or air conditioner. Since the CPU is directly distributed by the coolant in the coolant, the efficiency is high. The coolant usually uses liquids such as colorless and transparent mineral oil or anti-frozen, requiring the entire environment to prevent water vapor in the air to condense the surface of the low temperature coolant. Obviously, few people will use such simple and horrible cooling methods, but in a sense, such ideas are still very unique. When this method is cooled, the components such as the motherboard and graphics card cannot withstand the low temperature, so the CPU is generally cooled to -10 degrees. 4. Liquid nitrogen frozen method This is currently capable of minimizing CPU temperature. Under ideal conditions, CPU can cool to an absolute temperature of 20 degrees! The method is to seal the main board, CPU and other components in one space with a glass container, and the CPU is sealed by another glass container, which is filled with liquid nitrogen, and the circulating device is as water. This method is currently only realized in the laboratory, which enhances the outer frequency of the CPU to 170 MHz. 5. The PLL modification can also be seen from the various overclocking methods above. A CPU may work at 150MHz frequencies, but the motherboard does not provide such a high frequency, we can't know the limit of the CPU overclock. Even if some of the motherboard provides an outer frequency of up to 200MHz, and can be incremented by sight, the PCI can only use quadrestriction, and the AGP frequency can only be three-point frequency. When the frequency is 150m, the frequency of PCI and AGP Far far exceeds the specified range, the performance of other components will affect the overclocking of the CPU. Using an external frequency approach to generate a fixed PCI, AGP frequency can be adjusted separately by frequency synthesizer. Making such a frequency synthesizer is not complicated, but there is a certain understanding of the lines of the motherboard, and the various frequencies needed to work correctly to the corresponding position. The PLL modification law itself cannot overclock CPU, but it can fit other overclocking methods to find the true limit of the CPU. 6. Whether the wind-cooled design is the easiest, most reliable and practical, and the key to air-cooled heat dissipation is not the design of the heat sink and fan! The reason is very simple, only reasonable wind path design can play the effect of heat sink and fan, and is a trial of intelligence, which is also the purpose of being in the final way in the advanced article. Feng Road design requires rich knowledge, so I can only say some fur here, everyone can play freely. According to the type of fan (axial fan and turbine fan), the principle is different. Our most commonly used axial axis of the axial fan in the center, the blade is outside, and the center is a blind spot in the center of the wind. If the fan is close to the heat sink, the intermediate zone of the reinforcing wind speed is in the blind zone and cannot be dissipated in time, and when the fan is close to the heat sink, partially transfusion gas stream can also affect the fan of the fan. Therefore, the most reasonable fan design should take into account the distance between the heat sink, the closer the better. When using a double fan, one set is used at all, it is not possible to increase the wind speed, and it should be arranged in parallel, and the two winds need to have some distances, avoiding the airflow disturbances that just opposite the two directions.
