L9352B Application Note

June 2008 Rev 1 1/18 AN2791 Application note L9352B coil driver for ABS/ESP applications: current regulated channel analysis Introduction This document describes a detailed analysis on the current regulated channels of the ST coil driver L9352B. This intelligent quad-low side switch is typically used to drive inductive loads such as on-off valves of the hydraulic modulator of ABS(a)/ESP(b) control unit. a. Antilock Brake System (ABS). b. Electronic Stability Program (ESP). www.st.com zhou Line zhou Text Box 这里说是开关阀 Contents AN2791 2/18 Contents 1 L9352B overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2 Test bench layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3 Linearity relationship test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 4 Opening/closing time of the INLET valves versus duty-cycle of the hold-phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 5 Virtual current control loop on the Q1, Q2 channels of L9352B . . . . . 13 6 Revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 AN2791 List of tables 3/18 List of tables Table 1. INLET valve opening/closing time versus duty-cycle strategy . . . . . . . . . . . . . . . . . . . . . . 12 Table 2. Results of the first operative condition under test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Table 3. Results of the second operative condition under test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Table 4. Comparison between the current on loadsdrien by reg. channels and unreg. channels before and after the VCCL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Table 5. Results of the last operative condition under test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Table 6. Document revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 List of figures AN2791 4/18 List of figures Figure 1. L9352B application diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Figure 2. Comparison between the "ideal" linear relationship and the experimental data . . . . . . . . . . 6 Figure 3. Test bench layout used to characterize current control channels of L9352B. . . . . . . . . . . . 7 Figure 4. Current waveform produced on the load by the test coil energizing strategy . . . . . . . . . . . . 8 Figure 5. Comparison between the "ideal" linear relationship and the experimental data for two different load conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Figure 6. Coil current driven by means of L9352B in a typical ABS mission profile: coils on valves . . 9 Figure 7. Coil current driven by means of L9352B in a typical ABS mission profile: stand-alone coils. 10 Figure 8. "Pull-in"- "hold" phase duty-cycle strategy traditionally adopted to drive on-off valves. . . . 11 Figure 9. Evaluation of the VCCL on the L9352B unregulated channels: first test condition block scheme. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Figure 10. Evaluation of the VCCL on the L9352B unregulated channels: last test condition block scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 AN2791 L9352B overview 5/18 1 L9352B overview The L9352B (see Figure 1) is designed to drive inductive loads (e.g. relays, electromagnetic valves, etc.) in low side configuration. Integrated active Zener-clamp, for channels 1 and 2, or free wheeling diodes, for channels 3 and 4, allow the recirculation of the current of the inductive loads during the off-state of the DMOS. All four channels are monitored with a status output. All wiring to the loads and supply pins of the device are controlled. The device is self-protected against short circuit at the outputs and over-temperature. Channels 3 and 4 work as current regulator. A PWM signal, with a 2 kHz frequency, on the input defines the target for the output current, in particular, there is a linear relationship between the duty-cycle of the PWM input signal and the target value of the current (see Figure 2). The current is measured during recirculation phase of the load, that is, during the off-state of the DMOS. A sensing resistor, integrated in the IC and placed on the drain of the DMOS and of the free-wheeling diode, is devoted to measure the current. The benefit of the current regulation is an optimization of the PWM duty-cycle strategy against changes in the load conditions (e.g. temperature gradient and as a consequence coil resistor increases). Moreover, a test mode compares the differences between the two regulators. This "drift" test compares the output PWM of the regulators. Using this feature a drift of the load during lifetime can be detected. Figure 1. L9352B application diagram LOGIC Overload Open Load GND-det. Overtemperature Channel 1 Overtemperature Channel 4 LOGIC & DA Overload Open Load GND-det. IPD IPD LOGIC Overload Open Load GND-det. Overtemperature Channel 2 Overtemperature Channel 3 LOGIC & DA Overload Open Load GND-det. IPD IPD drift-det. VS VCC VDD Q1 Free wheeling valve Free wheeling valve Regulated valve Regulated valve D4 Q4 Q2 D3 Q3 GND TEST ST3 IN3 ST2 IN2 ST4 IN4 ST1 IN1 CLK EN Internal Supply Microcontroller 5 V logic supply Vbatt zhou Line zhou Line zhou Line zhou Line zhou Line zhou Text Box free wheeling diodes为与负载并联的蓄流二极管 zhou Text Box 在ABS 设计 领导形象设计圆作业设计ao工艺污水处理厂设计附属工程施工组织设计清扫机器人结构设计 中，阀高端都是有二极管的，这个与蓄流二极管应当如何搭配使用？ zhou Line zhou Text Box 这里没有明确说明是哪个通路可以用线性阀驱动 zhou Line zhou Line zhou Line zhou Line zhou Line zhou Text Box 与以前的用的低端驱动芯片不同，这里需要接到vs，这个vs应当是高端的输出，而不应当是Vbat，是不是蓄流用的？需要较线画粗？ zhou Line zhou Text Box 上面说Q3和Q4具有Free wheeling diodes，而这个图中画出来Q1和Q2的线圈为Free wheeling valves，是不是Q1和Q2在使用时需要加装蓄流二极管，而Q3和Q4则不用 L9352B overview AN2791 6/18 Figure 2. Comparison between the "ideal" linear relationship and the experimental data AN2791 Test bench layout 7/18 2 Test bench layout As shown in the Figure 2, the accuracy in the current control of the L9352B depends on the range of values of the PWM input signal duty-cycle. Basically, for duty-cycle greater than 16% it is possible to consider a current control accuracy of 6 %. The experimental data shown in the Figure 2 (i.e. red point) are related to the following test layout (see Figure 3): ● dSPACE Microautobox ● LEM Sensor LAH 25-NP ● INLET valves of the 8.0 ABS/ESP Bosch control unit ● coils with a resistor of 4.6 Ohm ● coil energizing frequency (i.e. valve opening/closing frequency) of 5 Hz ● coil energizing strategy: – for the first 20 ms, duty-cycle = 0.1 % – for the next 30 ms, duty-cycle = 90 % – for the last 150 ms, duty-cycle = [5:5:90] % The current waveform produced by this coil energizing strategy is shown in the Figure 4. Figure 3. Test bench layout used to characterize current control channels of L9352B. zhou Text Box 测试周期为200ms，分为三段，各段有不同的占空比，但是没有说明PWM的频率，频率是否为2kHz zhou Line zhou Line zhou Text Box 上面说Q3与Q4是调节电流的 zhou Text Box 从下面的测试中，没有看到测试周期为200ms Linearity relationship test AN2791 8/18 3 Linearity relationship test The measurements of the values of the mean current to compare with the "ideal" linear relationship of the L9352B current control channels have been carried out on the "hold- phase" of the current waveform. The Figure 5 describes a comparison between the results obtained on two different loads: ● stand-alone coils (i.e. blue stars, crosses and balls); ● coils on the valves (i.e. red stars, crosses and balls); The main difference between the two considered different load conditions is that for the coils stand alone you have an equivalent R-L circuit with a fixed inductance. On the other hand, when as load you consider a coil on a valve, from the point of view of the equivalent R-L circuit there is an inductance changing with the opening/closing dynamics of the valve. As the results of our analysis show in the Figure 5, 6 and 7 the current control loop of the Q3 and Q4 channels has been conceived in order to drive variable inductance loads, in fact, the spread between maximum and minimum values of the current for a fixed duty-cycle value of the PWM input signal is minimum in the case of a variable inductance load. Figure 4. Current waveform produced on the load by the test coil energizing strategy zhou Line zhou Line zhou Line zhou Line zhou Line zhou Line AN2791 Linearity relationship test 9/18 Figure 5. Comparison between the "ideal" linear relationship and the experimental data for two different load conditions Figure 6. Coil current driven by means of L9352B in a typical ABS mission profile: coils on valves Blue stars, crosses and balls indicate minimum, maximum and mean value of the target current for stand-alone coils. Red stars, crosses and balls indicate minimum, maximum and mean value of the target current for coils on valves. zhou Line zhou Line zhou Line Linearity relationship test AN2791 10/18 Figure 7. Coil current driven by means of L9352B in a typical ABS mission profile: stand-alone coils AN2791 Opening/closing time of the INLET valves versus duty-cycle of the hold-phase 11/18 4 Opening/closing time of the INLET(a) valves versus duty-cycle of the hold-phase The conventional strategy adopted to drive on-off valves used in the hydraulic modulator of ABS/ESP control unit is described in the Figure 8. The "pull-in" phase corresponds to the maximum values of duty-cycle applied for the first part of the valve opening/closing time. This phase guarantees the opening/closing of the valve against stiction phenomena due, for example, to the aging of the valve, to the dirt into the brake fluid, to the stiffness change of the valve spring and so on. The "hold" phase corresponds to the duty-cycle value that is necessary to maintain the valve opened/closed. Clearly this value is less than that used for the "pull-in" phase, because the force required to overcome the static friction is greater than the force required to overcome the dynamic one. Obviously, this kind of duty-cycle strategy is power saving too. Figure 8. "Pull-in"- "hold" phase duty-cycle strategy traditionally adopted to drive on-off valves Several tests have been carried out fixing the coil energizing frequency at 5 Hz and the time strategy at: ● for the first 20 ms, duty-cycle = 0.1 % ● for the next 30 ms, duty-cycle = 90 % ● for the last 150 ms, duty-cycle = [5:5:90] % The different duty-cycle configurations considered are summarized on the first column of the Table 1. In these tests, we measured the opening and closing time of the INLET valve, and, in addition, the time in which armature-piston of the valve starts its motion. An interesting result comes out. While the armature-piston motion and the closing time of the INLET valve are not affected by the duty-cycle configurations, the opening time is affected. In particular, this increases of 0.5 ms for each 5 % of duty-cycle increase of the "hold" phase. a. Take into account that the INLET valves of an ABS/ESP hydraulic modulator are on-off valves normally opened and normally controlled by a current control loop. On the hydraulic modulator there are also OUTLET valves. These valves normally closed do not require a current control loop but conventional low-side switch. zhou Line zhou Line zhou Line zhou Line zhou Line zhou Line zhou Text Box 带有蓄流二极管的Q3和Q4 zhou Text Box 这里的Pull-in时间是否为一个固定值呢？因为它是克服弹簧刚度用的，如果期望开度要是小的话，这个Pull-in时间是否相同呢？ zhou Text Box armature 衔铁 zhou Text Box 开启的时间受到PWM占空比的影响，而并断时间则不受影响。 Opening/closing time of the INLET valves versus duty-cycle of the hold-phase AN2791 12/18 Table 1. INLET valve opening/closing time versus duty-cycle strategy Duty-cycle strategy [“pull-in”_”hold”] Armature-piston motion [ms] Closing time [ms] Opening time [ms] 20_75 1.5 7 5 20_80 1.5 6.5 5 20_85 1.5 6 5 20_90 1.5 6 5 25_75 1.5 7 5.5 25_80 1.5 6.5 5.5 25_85 1.5 6 5.5 25_90 1.5 6 5.5 30_75 1.5 7 6 30_80 1.5 6.5 6 30_85 1.5 6 6 30_90 1.5 6 6 35_75 1.5 7 6.5 35_80 1.5 6.5 6.5 35_85 1.5 6 6.5 35_90 1.5 6 6.5 zhou Text Box 这个时间是如何测试出来的？它的时间这么长，如何用2KHz控制它呢？ zhou Text Box 这里开启与关断的时间是如何测试的？ zhou Text Box 这里Pull in的占空比反而小于Hold的占空比，这是为什么呢？ AN2791 Virtual current control loop on the Q1, Q2 channels of L9352B 13/18 5 Virtual current control loop on the Q1, Q2 channels of L9352B In this section we describe an analysis, done on the unregulated channels Q1, Q2 of the L9352B, aimed to understand the limits of a virtual current control loop on the same channels. The idea is to tune the duty-cycle of the Q1, Q2 channels on a measurement of the duty-cycle observed on the regulated channels Q3, Q4. Clearly, we considered same load conditions, that is, for both the regulated and unregulated channels of the L9352B, we considered same coils, same INLET valves. Furthermore, to balance the difference of PWM signal frequency on the L9352B channels, the unregulated ones (i.e. Q1, Q2) have been driven with a frequency of 3.9 kHz(b). In order to allow the current recirculation during the off-state of the Q1, Q2 channels, external free-wheeling diodes have been used to link the channel output and Vbat. As free-wheeling diode we have considered the ST power Shottky diodes 1N5817. Figure 9. Evaluation of the VCCL on the L9352B unregulated channels: first test condition block scheme As first operative condition for our tests (see Figure 9) we can refer to the following data: ● INLET valve opening/closing frequency of 1Hz ● for the first 450 ms, duty-cycle = 0.1 % ● for the next 50 ms, duty-cycle = 90 % ● for the last 500 ms, duty-cycle =[ 15:10:75] % The Table 2 shows the results related to this first operative condition that we considered for our tests. As we can see in the last four columns, the difference between the mean current on the load driven by the unregulated channel and the mean current on the load driven by the regulated channel reduces as the set-point, that is, the duty of the hold phase increases. b. Take into account that the ideal frequency of the output PWM signal of the current regulated channels (i.e. Q3, Q4) is the (clock frequency)/64, that is, 3.9 kHz for a clock frequency of 250 kHz. zhou Line zhou Line zhou Line zhou Line zhou Line zhou Line zhou Line zhou Text Box 1、Q3与Q4在测试时可以不只是2Khz，与时钟频率有关 2、这个时钟频率是什么时钟频率，SPI的，这里并没有说SPI的频率是多少。 zhou Text Box 整个的测试周期为1s Virtual current control loop on the Q1, Q2 channels of L9352B AN2791 14/18 Similar results have been observed considering another operative condition, characterized by a different duty-cycle strategy (see Table 3). Clearly, the duty-cycle applied on the unregulated channels in both the operative conditions under test is the same measured on the regulated channel. See columns 2, 3 of the following tables to understand the difference between the two duty-cycles applied on the unregulated and regulated channels of the L9352B on the same load conditions. As second operative condition for our tests (see Figure 9) we can refer to the following data: ● INLET valve opening/closing frequency of 1 Hz ● for the first 500 ms, duty-cycle = 0.1 % ● for the last 500 ms, duty-cycle = [15:10:75] % The Table 2 and 3 show the results of a comparison between the mean current on the loads driven by the regulated channels of L9352B and the unregulated ones. These last have been trained on the output duty-cycle of the regulated channels. Just to highlight the results obtained by this analysis, it is important to summarize the difference in the coil current of the unregulated channels of L9352B before and after the Table 2. Results of the first operative condition under test Input duty (hold phase) Unreg. ch. output duty (before the manual regulation) Reg. ch. output duty Unreg. ch. mean current [mA] Unreg. ch. pk-to-pk current [mA] Reg. ch. mean current [mA] Reg. ch. pk-to-pk current [mA] 0.15 0.83 0.75 520 280 390 360 0.25 0.73 0.65 749 360 660 360 0.35 0.64 0.53 990 400 920 480 0.45 0.55 0.41 1200 480 1200 480 0.55 0.44 0.26 1480 480 1480 480 0.65 0.35 0.1 1700 480 1740 440 0.75 0.23 0.09 1780 520 1760 480 Table 3. Results of the second operative condition under test Input duty (hold phase) Unreg. ch. output duty (before the manual regulation) Reg. ch. output duty Unreg. ch. mean current [mA] Unreg. ch. pk-to-pk current [mA] Reg. ch. mean current [mA] Reg. ch. pk-to-pk current [mA] 0.15 0.83 0.75 530 320 390 360 0.25 0.73 0.65 780 400 660 360 0.35 0.64 0.53 980 440 920 440 0.45 0.53 0.39 1240 480 1200 480 0.55 0.44 0.25 1460 560 1480 520 0.65 0.34 0.09 1630 440 1740 480 0.75 0.25 0.09 1630 440 1820 480 zhou Text Box 这里控制与输出的电压是取反的，与输出的电流是正比的，这是对的。 AN2791 Virtual current control loop on the Q1, Q2 channels of L9352B 15/18 regulation inspired to the duty-cycle value carried out by the regulated channels of the same device. In Table 4 we can see the results of a comparison between the mean current on the loads, driven by the L9352B regulated channels and L9352B unregulated channels before and after the VCCL regulation. As last operative condition considered in our tests we can refer to the following data and the Figure 10: ● INLET valve opening/closing frequency of 1 Hz ● for the first 450 ms, duty-cycle = 0.1 % ● for the next 50 ms, duty-cycle = 90 % ● for the last 500 ms, duty-cycle = [15:10:75] % The main idea is to increase the resistance of the loads of about the 15 %. The initial value of 4.8 Ohm, that is, 4.6 Ohm of the coil resistor plus 0.2 Ohm of Rds-ON of the DMOS has been increased of 0.6 Ohm. So doing, we simulated a gradient temperature of about 35°. For this calculation we referred to the formula of the resistance of the chopper versus the temperature: Rlast = Rinitial(1 + 0.004ΔT) From the results shown in the table 7-4, it comes out that the unregulated channels maintains a satisfactory tracking capability of the current values driven on the loads also in simulated conditions of temperature gradient. Figure 10. Evaluation of the VCCL on the L9352B unregulated channels: last test condition block scheme Table 4. Comparison between the current on loadsdrien by reg. channels and unreg. channels before and after the VCCL Hold phase duty Mean current on the cycle-time for the reg. ch. [mA] Mean current on the cycle- time for the unreg. ch. –- before the regulation -- [mA] Mean current on the cycle- time for the unreg. ch. –- after the regulation -- [mA] 0.35 530 450 560 0.45 660 550 670 zhou Line zhou Line zhou Line zhou Line zhou Text Box VCCL是如何调节的？ Virtual current control loop on the Q1, Q2 channels of L9352B AN2791 16/18 In Table 5 comparison between the mean current on the loads driven by the regulated channels of L9352B and the unregulated ones. These last have been trained on the output duty-cycle of the regulated channels. Table 5. Results of the last operative condition under test Input duty (hold phase) Unreg. ch. output duty (before the manual regulation) Reg. ch. output duty Unreg. ch. mean current [mA] Unreg. ch. pk-to-pk current [mA] Reg. ch. mean current [mA] Reg. ch. pk-to-pk current [mA] 0.15 0.83 0.73 500 220 400 260 0.25 0.73 0.6 730 280 670 300 0.35 0.64 0.46 1000 380 945 360 0.45 0.54 0.32 1180 440 1220 400 0.55 0.44 0.09 1400 480 1480 400 0.65 0.35 0.09 1400 480 1620 440 0.75 0.23 0.09 1400 480 1600 400 AN2791 Revision histo