Buck converters are the most common power topology, and power engineers know their strengths and weaknesses. One of the challenges of power system design is current detection. In buck converters, a popular and convenient method is DCR current detection. It's easy because it doesn't add extra cost or power to the power supply design, but what is known is that this circuit is very accurate, especially when using small, low-ESR inductors. This is the case.
Figure 1: DCR current detection circuit
The calculation of the RC value is simple enough. 
,among them:
L = L1 inductance value;
DCR = DC resistance of inductor L1;
R = R2 in the schematic of Figure 1 (or, in the presence of R3, the parallel connection of R2 and R3);
C = C1 in the schematic of Figure 1.
Note that in Figure 1, if the magnitude of the ISENSE peak signal saturates the differential amplifier, then increase R3 to reduce the peak signal amplitude to within the specified range of the differential amplifier.
Simple is always welcome, but as the common sense puts it, "cheap is not good, good is not cheap." The accuracy of this circuit is very poor.
First, the DCR of the inductor has a wide tolerance range, and ±7% or even ±10% is common.
Figure 2: Typical Specifications of Inductor DCR
If the initial tolerance is 10%, the DCR of the 180nH inductor shown in Figure 1 may be as low as 261mΩ or as high as 319mΩ. To make matters worse, the inductor will heat up and the temperature coefficient of the copper wire winding is 3930PPM/oC or 0.393%/oC. If the application's temperature rises to 35oC above ambient and the inductor itself heats up to raise the temperature by 35oC, then the nominal DCR may rise to:
Increased by 35% from the nominal value.
The worst case limit is:
Increase the nominal value by 40%.
The worst case lower limit is:
An increase of 15% from the nominal value (the lower total error is because the positive coefficient of the copper compensates for the low initial value of the inductor).
From an engineering design point of view, this is really bad because over-current flags and over-current shutdowns are based on these resistance settings. If the circuit is too sensitive, it will shut down if it does not reach the point where it needs to be shut down. This is not the result we want. If the circuit is insensitive, there is a risk that the inductor and power FETs are overstressed. This is not the result we want.
How bad can the situation be?
Assume that you are designing a circuit that can provide up to 35A at 1V (this value is now reasonable for a practical single-phase buck converter). If the inductor DCR is at the low end of the tolerance, then when the output gets 35A, the controller thinks that 40A is provided. This means that OCP cannot be set below 40A, otherwise the power supply will stop at nominal load.
Conversely, when the OCP is set to 40A and the inductor DCR increases by 10%, how bad is the situation?
In this case, the actual load current is 40A, but the DCR is 407μΩ, so the controller thinks that the output current is 65A. This means that the OCP needs to be set to 65A. If it is not set to this value, there is a risk that the OCP will stop at less than 40A. This seems unacceptable, but once the OCP is set to 65A, the circuit must be designed to provide such large currents continuously, with occasional accurate reporting of current. This means that the output inductor and power FET are seriously over-designed, and the power supply must provide 35A, but it must be designed to provide continuous 65A. Moreover, to make things even worse, the current in the inductor has a peak to peak ripple in addition to the DC component. How large is this ripple? For ripple current, the commonly used design principle is 20%. This means that the cycle-by-cycle current limit must be set above 65A, so the ability to protect the output FET becomes very problematic.
Guess what will happen if you design for 30% ripple current?
Then, you will realize that the typical current detection voltage range is 10mV to 20mV. If in a power supply, a switching node rings, there is a stray magnetic field generated by the output inductor, and current flows in the bypass capacitor and the output capacitor, it is difficult to obtain an acceptable signal-to-noise ratio (SNR). To have any hope of signal quality, current sense connection lines must be carefully arranged as differential pairs (thus, any noise picked up is common mode) and placed away from inductors, switching nodes and high current/high frequency Current loop. This is difficult in space-constrained designs, as everything in space-constrained designs now looks very difficult.
Figure 3: Kelvin inductor current sense wiring
what can we do?
First, the temperature of the inductor can be estimated based on experience by using a thermistor or a temperature detection diode (usually a forward-biased PNP base-nose in a small transistor). In this way, the thermal response of the copper wire winding resistance can be adjusted. This is too helpful. The engineers are really amazing. If we do really careful, then the best result may reach ±10%.
What else can we do?
We can ignore the simple DCR circuit and connect the output inductor in series with an expensive, temperature-stable current-sense resistor. This adds cost and impairs converter efficiency, but with good differential signal routing, we can detect output current with much higher accuracy. As tolerances accumulate, we can get an overall current detection performance of ±5% or better. Engineers in the design review both proved the rationality of this program, but also avoided criticism of its design impact on efficiency and cost, their courage I admire.
What about using an inductor made of a temperature-stable alloy winding? As soon as this idea emerged, my heart was scared to jump.
Is there any other way?
There is something better than a current sense resistor. Let the power link device report its current. This method uses a well-designed Smart Power State (SPS) that, while increasing the current detection cost, can provide peak power capabilities that are very close to the nominal output requirements, both of which are offset. As a result, waste caused by overdesigning power link elements is greatly reduced. How much expectation can we place on this current detection method? For a reasonable operating area (do not expect a miracle when the output current is near zero), we can get an initial accuracy of ± 1%, with a worst tolerance of ± 2% with aging and temperature changes. Friends, this is the Christmas gift I want.
Technology advances year after year, providing engineers with better and better basic components. The simple DCR current detection circuit still goes with the wind.
Multi-core cpu is more and more popular, people are used to take dual core processors laptop before. However, Quad Core Laptop is becoming selling like a hot cake nowadays. You can find Quad Core Processor Laptop at our store, like I3 Quad Core Laptop,I7 Quad Core Laptop, I5 Quad Core Laptop, intel celeron J4125 or N5095 15.6 Inch Laptop, even N4120 quad core 14 Inch Laptop, etc. Therefore, you can just share the parameters prefer, like size, cpu, ram, rom, gpu, application scenarios, thus save much time to get a win-win solution.
Do you know the reason why more people choose quad core device? The core reason is that heavier tasks people need to finish at a higher speed than before. Nowadays quick rhythm is becoming the main style in city even everyone is eager to downshifting. So more powerful laptop, computer, mobile phone is a trend, though most functions never are used in lifetime.
15.6 inch Gaming Laptop or 14 Inch Gaming Laptop is becoming the most popular level at the market.
Quad Core Laptop,I3 Quad Core Laptop,I7 Quad Core Laptop,I5 Quad Core Laptop,Quad Core Processor Laptop
Henan Shuyi Electronics Co., Ltd. , https://www.shuyicustomtablet.com