Improved SENSE Calibration and Benefits Guide About this document Scope and purpose This application note illustrates the achievable current sense accuracy of Infineon’s high-side power switch family Power PROFET. It provides details to the current sense “default” accuracy explaining relevant fundamentals and explains thee effects on quantitative examples. The application note illustrates possible measures to improve system accuracy by means of calibration including recommendations regarding software implementation. Intended audience This application note is targeted for all design engineers, which need to understand and possibly further improve the system accuracy performance of the Power PROFET current sense functionality. Application Note www.infineon.com Please read the Important Notice and Warnings at the end of this document V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Abstract Table of Contents About this document....................................................................................................................... 1 Table of Contents ........................................................................................................................... 2 1 Abstract ........................................................................................................................ 3 2 2.1 2.2 2.3 2.4 2.5 Introduction.................................................................................................................. 4 Pin Names and Functions ....................................................................................................................... 4 Voltages and Currents ............................................................................................................................. 5 Flowchart Nomenclature ........................................................................................................................ 6 Example Circuit Board Scenario ............................................................................................................. 7 Fundamental Concepts ........................................................................................................................... 7 3 3.1 3.2 “Default” current sense performance ............................................................................. 10 Measureable load current range .......................................................................................................... 11 “Default” current sense accuracy ......................................................................................................... 12 4 4.1 4.2 4.3 Calibration Techniques ................................................................................................. 15 Variation effects .................................................................................................................................... 15 1-point calibration................................................................................................................................. 16 2-point calibration................................................................................................................................. 21 5 Important considerations .............................................................................................. 25 6 6.1 6.2 6.2.1 6.2.2 6.2.3 6.3 Calibrating Power PROFET............................................................................................. 26 Calibration Nomenclature and Equations ........................................................................................... 26 Application Software Implementation ................................................................................................. 27 No Calibration (No Cal) .................................................................................................................... 27 1-point Calibration ........................................................................................................................... 28 2-Point Calibration ........................................................................................................................... 30 Accuracy of Different Calibration Options............................................................................................ 32 7 Conclusion ................................................................................................................... 33 Revision History ............................................................................................................................ 33 Application Note 2 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Abstract 1 Abstract Smart, high-side power switches from Infineon® are designed to control all types of resistive, inductive, and capacitive loads. These devices provide protection and diagnostic functions and are specially designed to drive loads in harsh automotive environments. The diagnostic feature analog current sense is often used to diagnose, control and protect the load as well as to protect and diagnose the overall system including wire harness. Ideally the analog current sense diagnostic should reflect the load current without any additional error contribution. However in reality analog current sense diagnostics does always have an inherent inaccuracy associated. Infineon offers multiple high-side power switches families with different accuracy performance. For specific high-side power switches additional calibration techniques are supported, which achieve increased levels of accuracy compared to the specified, “default” overall sense performance. This application note explains the calibration techniques for the high-side power switch family Power PROFET. The application note first introduces some fundamental concepts. It explains the dominating sources of inaccuracy, their observed behavior and how to reduce these by means of calibration. Beside the fundamental explanation quantitive examples are given for the Power PROFET BTS50015-1TAD. The application note also details the sense performance constraints which remain after calibration. Note: The following information is given as an implementation suggestion only, and shall not be regarded as a description or warranty of a certain functionality, condition, or quality of any device. Application Note 3 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Introduction 2 Introduction Current sensing is implemented within high-side switches to diagnose systems and to protect them in the event of failures. High-side current sensing is used to protect both the load and the wiring harness, to diagnose the load so as to ensure proper operation, and to measure the output current for the purpose of controlling the output power. Note: Further generic information on high-side switches with diagnostics and protection can be found in the Application Note: What the designer should know: Short introduction to PROFETTM +12V. There are two main dominating effects with conventional high-side current sensing solutions which have an impact on the overall accuracy performance. The first is the inaccuracy which is resulting from an internal amplifier offset voltage. This offset voltage deteriorates the current sense accuracy especially at lower load currents. In addition it suppresses the current sense functionality below certain load current thresholds. The second is the slope (steepness) inaccuracy, which becomes more significant at higher load currents. The overall current sense performance specified in the Power PROFET datasheet has to cover all possible combinations of both, offset voltage and slope (steepness) inaccuracy including their variation over production, life time and specified operating conditions. Whenever the overall specified sense performance does not meet the accuracy specified, additional calibration techniques can be introduced. In case of Power PROFET the supported option to further improve sense accuracy is to perform either a 1-point or preferably a 2-point calibration utilizing end-of-line measurement and low application software overhead. 2.1 Pin Names and Functions Single-channel, high-side power switches of the general type considered in this paper have five pins (GND, IN, OUT, IS, and VS) as illustrated in Figure 1. Figure 1 Control input from microcontroller IN Current sense output to microcontroller IS VS OUT Main output to drive the load GND Pin names The functions of these pins are detailed in Table 1. Table 1 Pin functions Pin Name Pin Function GND Ground: Ground connection IN Input: Digital 3.3V and 5V compatible logic input; activates the power switch if set to HIGH level (definitions for HIGH and LOW can be found in the parameter tables of the Application Note 4 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Introduction respective device datasheet) Please Note that PowerPROFET 500xx-1TAD & TMD family offers a VS capable Input pin which supports in addition control with Voltages above 5V up to VS OUT Output: Protected high-side power output IS Sense: Analog sense current signal VS Supply Voltage: Positive supply voltage for both the logic and power stages 2.2 Voltages and Currents Figure 2 illustrates the voltages and currents referenced in this application note. The load current IL and the sense current IIS will be the focus of the following discussions. VS VS IS IIN IN IL VS OUT VIN VOUT IIS IS VIS GND IGND GND Figure 2 Definition of currents and voltages These abbreviations are defined in Table 2. Table 2 Voltage and current abbreviations Abbreviation Meaning VS Supply voltage GND Ground VIN Control input voltage VOUT Output voltage driving the load Application Note 5 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Introduction VIS Sense voltage IL Load current IIS Sense current IS Supply current IGND Ground current 2.3 Flowchart Nomenclature With regard to flowcharts used in this application note, the representation of the five main symbols is illustrated in Figure 3. Document / Note Process / Action Decision Pre-defined Process (Subroutine) Figure 3 Application Note Internal Storage Flowchart nomenclature 6 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Introduction 2.4 Example Circuit Board Scenario For the purposes of this application note, it is assumed that a circuit board containing a microcontroller and some number of single-channel, high-side power switches as illustrated in Figure 4. Circuit board High-side power switches Microcontroller Figure 4 Example circuit board scenario The microcontroller is used to turn the high-side power switches ON and OFF, and also to measure the value of the sense current (IIS) outputs from the switches. 2.5 Fundamental Concepts In order to understand the problems associated with conventional current sense functionality as also implemented in Power PROFET, it is first necessary to be familiar with some fundamentals. The effects which dominate the resulting current sense performance as well as the measures to improve the overall current sense accuracy can be explained in the following simplified manner utilizing a straight line in "slope-intercept" form. The general formula for a straight line in "slope-intercept" form, is presented in Equation (1). Equation (1) y = m×x + b In this case, y is the value on the vertical axis (Y), x is the value on the horizontal axis (X), m is the slope of the line, and b – which is known as the y-intercept – is the point at which the line intersects the Y-axis. For the purposes of this application note, both, positive and negative y-intercept values need to be discussed as illustrated in Figure 5. Application Note 7 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Introduction y y y b x x (a) b = 0 Figure 5 (b) Positive b value b x (c) Negative b value Generic lines with identical positive slopes All three lines in Figure 5 have the same m (slope) value. The difference between the lines is the b (y-intercept) value. The line in Figure 5(a) has a b value of zero; the line in Figure 5(b) has a positive value for b; and the line in Figure 5(c) has a negative value for b. Let’s take now a look to the high-level block diagram and the specified diagnosis performance for a conventional high-side power switch as illustrated in Figure 6. VS IN IIS IS Input circuit Sense output circuit Power Switch OUT IL Diagnosis and Protection GND Figure 6 High-level block diagram for a conventional high-side power switch The ideal relationship between the sense current IIS and the load current IL is shown in Figure 7(a). Ideally the sense current should show a replica of the load current across a load current range. Ideally the sense current should always be a fixed ratio (or portion) of the load current. In reality, however, there may be errors involved. These errors will be a slope (steepness) error as illustrated in Figure 7(b). The slope error is mainly dependent on part-to-part production variation. The effects of slope error are more pronounced at higher load currents. Application Note 8 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Introduction In addition there may also be a sense offset error. Resulting from an internal amplifier offset voltage. The sense offset error is strongly dependent on production variation and the operating temperature of the device. The effects of the offset are more pronounced at lower load currents. The sense offset error for a Power PROFET may be both, positive and negative. Positive offset errors will result in a remaining current at the IS pin, even though no load current is flowing. Negative offset error will “disabled” the sense functionality below a certain load current threshold, which would cause a theoretical negative sense current. Load currents at this threshold or below will result in no sense current at the IS pin as illustrated by the horizontal portion of the solid green line in Figure 7(c). typ IIS IIS Slope (steepness) error IIS typ IIS typ IIS IIS max IIS Offset error min IIS IL (a) Ideal curve IL (b) With slope error IL (c) With offset error (Offset = 0, Slope = typical) (Offset = 0) (Blue = positive; Green = negative) Figure 7 Relationship between IIS and IL in conventional devices Figure 8 shows the enhanced illustration of Figure 5 considering both, offset error and slope error. The offset error varies between a negative and positive offset bMIN and bMAX. The slope error will decrease/increase the typical slope m towards a minimum slope mMIN and a maximum slope mMAX. y m mMAX yMAX=mMAX*x+bMAX y y y=mMAX*x-bMIN mMIN y=mMIN*x+bMAX bMAX yMIN=mMIN*x-bMIN x (a) b = 0 Figure 8 Application Note x (b) Positive bMAX value bMIN x (c) Negative bMIN value Generic lines with varying positive slopes 9 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide “Default” current sense performance 3 “Default” current sense performance For a Power PROFET the relationship between the sense current IIS and the load current IL is expressed using Equation (2). Equation (2) IIS = ( 1 d𝑘ILIS × IL ) + IIS0 Comparing Equation (2) and Equation (1) it can be seen, that The sense current IIS in Equation (2) corresponds to y in Equation (1) the load current IL in Equation (2) corresponds to x in Equation (1) the slope defined by 1/dkILIS in Equation (2) corresponds to m in Equation (1). If dkILIS varies between a maximum limit and a minimum limit, o the maximum limit of dkILIS(MAX) will result in a minimum slope steepness mMIN and o the minimum limit of dkILIS(MIN) will result in a maximum slope steepness mMAX. the sense offset current IIS0 in Equation (2) corresponds to the y-intercept b in Equation (1). o A maximum limit of IIS0(MAX) represents a positive y-intercept bMAX o A minimum limit of IIS0(MIN) represents a negative y-intercept bMIN. The “electrical characterstics” datasheet section “Diagnostic Function: Current Sense Ratio Signal in the Nominal Area, Stable Current Load Condition” outlines the “default” diagnosis performance of every Power PROFET device. All shipped Power PROFET devices will show a sense performance within the specified MIN-MAX-range of the slope defined by 1/dkILIS. The slope (steepness) error varies therefore from a minimum slope mMIN=1/dkILIS(MAX) to a maximum slope mMAX=1/dkILIS(MIN) the specified MIN-MAX-range of the calculated sense offset current IIS0. The calculated sense offset error at zero load current varies therefore between the limits of IIS0(MIN) and IIS0(MAX) the specified minimum and maximum sense currents IIS1, IIS2, IIS3 and IIS4 for given load currents IIL1, IIL2, IIL3 and IIL4. Note: Due to the nature of the sense circuitry the calculated sense offset current IIS0 may vary for a given device over temperature (IIS0=f(TJ)). All devices with a negative calculated sense offset current (IIS0<0) will only provide a current at the IS pin, whenever the load current IL exceeds a threshold IL0, where the resulting sense current is larger than the sense offset (IL>IL0=-(dkILIS(MAX) x -IIS0 ). Note: Since the calculated sense offset current IIS0 may vary for a given device over temperature (IIS0=f(TJ)) the load current threshold IL0 may vary over temperature also (IL0=f(TJ)). Figure 9 shows the current sense transfer function of the Power PROFET BTS50015-1TAD. Application Note 10 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide “Default” current sense performance 3.5 dkILIS(min) 3 dkILIS(typ) 2.5 dkILIS(max) IIS [mA] 2 1.5 1 0.5 IIS0(max) 0 0 20 40 IL1 IL2 60 80 IL3 100 120 140 160 IL4 IL[A] Figure 9 Power PROFET BTS50015-1TAD current sense Comparing Figure 9 with Figure 8 it can be seen that the current sense performance of this specific Power PROFET varies in a range which is limited by an upper line resulting from a device which has a combination of maximum positive calculated sense offset current IIS0(MAX) and maximum 1/dkILIS slope (or vice versa a minimum dkILIS(MIN)). This line can be represented by the general formula yMAX=mMAX*x+bMAX. Applying Power PROFET parameters the formula changes to IIS(MAX)=1/ dkILIS(MIN)* IL + IIS0(MAX) a lower line resulting from a device which has a combination of maximum negative calculated internal sense offset current IIS0(MIN) and minimum 1/dkILIS slope (or vice versa a maximum dkILIS(MAX)). This line can be represented by the general formula yMIN=mMIN*x-bMIN. . Applying Power PROFET parameters the formula changes to IIS(MIN)=1/ dkILIS(MAX)* IL + IIS0(MIN) (with IIS0(MIN) being negative). Note: Connecting the points of the maximum sense current limits for given load currents ([IISi;IILi] with i=1..4) will form the upper limiting line (blue line according Figure 7 and Figure 8). Connecting the points of the minimum sense current limits ([IISi;IILi] with i=1..4) will form the lower limiting line (green line according Figure 7 and Figure 8). Note: Any negative offset error IIS0 will not result in sense current that is sinked by the IS pin. The IS pin can only source a sense current. Any negative offset error will disable the load current sense function once the load current is less than a certain threshold. The variation of this “default” current sense performance brings some limitations in terms of measurable load current range and accuracy. 3.1 Measureable load current range In general the measurable load current range is limited towards a lower load current threshold and towards an upper load current threshold. Application Note 11 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide “Default” current sense performance The lower load current threshold results from the limiting lower line (yMIN=mMIN*x-bMIN) where the line intercepts the x-Axis. In order to guarantee any variation in x will result in a variation of y, a lower threshold of x>-(-bMIN/mMIN) needs to be exceeded. Applying Power PROFET parameters the formula changes to the lower load current threshold IL> IL0=-(IIS0(MIN)*dkILIS(MAX)) (with IIS0(MIN) being negative). Note: Looking at Figure 9 bottom solid dILIS(MAX) line it can also be seen, that for the example of a Power PROFET BTS50015-1TAD only load currents of IL0>~9A will result in a sense current at the IS pin. The upper load current threshold results from the limiting upper line (yMAX=mMAX*x+bMAX) which can be reliably provided to peripheral readout ciruit. For Power PROFET in general the upper load current threshold is the specified maximum load current IL4. Load currents above IL4 may already trigger the activation of protection mechanisms or the resulting maximum sense current IIS4(MAX) may start to saturate. Note: Looking at Figure 9 it can be seen that for the example of a Power PROFET BTS50015-1TAD only load currents up to IL4=135A are specified. Checking BTS50015-1TAD datasheet limits of the “Current Trip Detection Level” (P_6.1.35) and and “Sense Signal Saturation Current” (P_6.1.75) will show, that at higher load currents the device may already switch off from over current or that the sense current may already saturate. 3.2 “Default” current sense accuracy The “default” accuracy depends on the load current. For the “default” accuracy is of general interest, in which error limits the x value may vary, assuming a certain y value is read and processed. Figure 10 illustrates the xi error for a given yi value yMAX=mMAX*x+bMAX y y=m*x MinError MaxError yi bMAX bMIN Figure 10 yMIN=mMIN*x-bMIN xi(MIN) xi xi(MAX) x X-Axis error for generic lines with varying positive slopes The general absolute error for any xi value will range between a minimum value xi(MIN) and a maximum xi(MAX). The minimum error results from the limiting upper line (yMAX=f(x)=mMAX*x+bMAX) in relation to the typical line (y=f(x)=m*x). The maxmum error results from the limiting lower line (yMIN=f(x)=mMIN*x-bMIN) in relation to the typical line (y=f(x)=m*x). Application Note 12 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide “Default” current sense performance In general the absolute minimum x- error-value can be calculated from solving the equation (x=f(yMAX))-(x=f(y)). The absolute minimum x- error-value (MinError=xi(MIN)- xi) can be calculated according Equation (3) Equation (3) 𝑥×(𝑚−𝑚𝑀𝐴𝑋 )−𝑏𝑀𝐴𝑋 MinAbsError = 𝑚𝑀𝐴𝑋 The relative minimum x-error (MinRelError=( xi(MIN)- xi)/xi) can be calculated according Equation (4) Equation (4) MinRelError = 𝑚×𝑥−𝑏𝑀𝐴𝑋 𝑚𝑀𝐴𝑋 ×𝑥 𝑚 −1= 𝑚𝑀𝐴𝑋 𝑏𝑀𝐴𝑋 − 𝑚𝑀𝐴𝑋 ×𝑥 −1 The absolute maximum x- error-value (MaxError=xi(MAX)- xi) can be calculated according Equation (5) Equation (5) MaxAbsError = 𝑥×(𝑚−𝑚𝑀𝐼𝑁 )+𝑏𝑀𝐼𝑁 with bMIN being an absolute number. 𝑚𝑀𝐼𝑁 The relative maximum x-error (MaxRelError=( xi(MAX)- xi)/xi) can be calculated according Equation (6) Equation (6) MaxRelError = 𝑚×𝑥+𝑏𝑀𝐼𝑁 𝑚𝑀𝐼𝑁 ×𝑥 −1= 𝑚 𝑚𝑀𝐼𝑁 + 𝑏𝑀𝐼𝑁 −1 𝑚𝑀𝐼𝑁 ×𝑥 with bMIN being an absolute number Applying Power PROFET parameters to Equation (3) to Equation (6) results in Equation (7) to Equation (10). Equation (7) 1 MinAbsError = (𝐼𝐿 × ( 𝑑𝐾𝐼𝐿𝐼𝑆(𝑇𝑌𝑃) − 1 𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐼𝑁) )− 𝐼𝐼𝑆0(𝑀𝐴𝑋) 106 ) × 𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐼𝑁) with IIS0(MAX) as µA value 𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐼𝑁) Equation (8) MinRelError = Equation (9) MaxAbsError = (𝐼𝐿 × ( 𝑑𝐾𝐼𝐿𝐼𝑆(𝑇𝑌𝑃) − 𝐼𝐼𝑆0(𝑀𝐴𝑋) ×𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐼𝑁) 𝐼𝐿×106 1 𝑑𝐾𝐼𝐿𝐼𝑆(𝑇𝑌𝑃) − 1 𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐴𝑋) −1 )− 𝐼𝐼𝑆0(𝑀𝐼𝑁) 106 ) × 𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐴𝑋) with IIS0(MIN) as µA value being negative Equation (10) MaxRelError = 𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐴𝑋) 𝑑𝐾𝐼𝐿𝐼𝑆(𝑇𝑌𝑃) − 𝐼𝐼𝑆0(𝑀𝐼𝑁) ×𝑑𝐾𝐼𝐿𝐼𝑆(𝑀𝐴𝑋) 𝐼𝐿×106 − 1 with IIS0(MIN) as µA value being negative Table 3 shows the calculated absolute and relative errors for certain load currents for the example of BTS50015-1TAD. These values can alternatively also be graphically derived from Figure 9, estimating the load current difference for any, fixed sense current between the typical, dashed line in comparison to the limiting upper and lower solid line. Application Note 13 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide “Default” current sense performance Table 3 ILoad minimum absolute minimum relative ILoad error ILoad error Maximum maximum relative absolute ILoad error ILoad error 10A -10A -98% +11A +107% 20A -11A -55% +12A +60% 40A -13A -34% +14A +36% 60A -16A -26% +17A +28% 80A -18A -23% +19A +24% 100A -21A -21% +22A +22% In terms of accuracy it can be seen, that the “default” accuracy reaches at nominal current just ~35..40%. This accuracy deteriorates further towards lower load current. At high load a “default” accuracy of about 22% can be achieved. Note: The above example outlines approximated values to show the general accuracy trend. For full system performance additional contributors like board leakage currents, variation of sense resistor RIS, error contribution of uC for analog digital conversion etc. need to be considered. Application Note 14 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques 4 Calibration Techniques Whenever the “default” sense performance does not meet the system accuracy targets, additional calibration techniques can be introduced. 4.1 Variation effects During the development of Power PROFET Infineon performed intense stress and characterization efforts to understand the overall sense performance after calibration. The investigations revealed the following findings: the 1/dkILIS(i) slope of any calibrated Power Profet varies after calibration for each individual Power PROFET (i) over temperature and stress within certain limits. These limits vary depending on the Power PROFET product type / respective family member. The maximum variation per calibrated Power PROFET product type will remain within the limits of the Parameter (dKILIS(CAL)) (see datasheet parameter P_6.1.47). devices, which show after calibration an individual dkILIS(i) value, which is close to the absolute limits of the Current Sense Differential Ratio dKILIS (Parameter 6.4.41), will not violate these limits over temperature and stress The calculated sense offset current IIS0(i) of any calibrated Power Profet varies after calibration for each individual Power PROFET (i) over temperature and stress within certain limits. These limits vary depending on the Power PROFET product type / respective family member. The maximum variation per Power PROFET product type will remain within the specified temperature dependant calculated sense offset current limits (see datasheet limits of Parameter P_6.1.42). devices, which show after calibration an individual calculated sense offset current IIS0(i) which is close to the absolute limits of the calculated sense offset current IIS0 (Parameter 6.4.42), will not violate these limits over temperature and stress Based on the example of Power PROFET BTS50015-1TAD this means: According to datasheet parameter P_6.1.47 the individual 1/dkILIS(i) slope of any calibrated BTS500151TAD varies after calibration for each individual BTS50015-1TAD (i) over temperature and stress a maximum of +/-5%. To give an example, assuming an individual dkILIS(i)=50000 (or vice versa 1/dkILIS(i)=2E-5) has been derived by means of calibration at Tj=25°C, the device specific dkILIS(i) will maximum vary over life time and temperature (-40°C<=Tj<=+150°C) between 47500<=dkILIS<=52500 (or vice versa 2.105E-5>=1/dkILIS>=1.9E-5). BTS50015-1TAD (i), which have an individual dkILIS(i) value close to the minimum limit of parameter 6.1.41 of 45300, will not violate this limit over temperature and stress. BTS50015-1TAD (i), which have an individual dkILIS(i) value close to the maximum limit of parameter 6.1.41 of 57700, will not violate this limit over temperature and stress. In case a calibrated BTS50015-1TAD (i) shows a positive calculated current sense offset, IIS0(cal)>0, then this individual sense offset will vary over temperature and stress towards the differences between the room temperature maximum limit and the respective maximum limits at “cold” (Tj=-40°C) and “hot” (Tj=+150°C). To give an example, assuming an individual positive calculated current sense offset, IIS0(cal)=50µA, has been derived by means of calibration at Tj=25°C, the device specific IIS0(cal) will vary over life time and temperature (-40°C<=Tj<=+150°C) between the MAX datasheet limits of IIS0. This results in a lower difference of IIS0(MAX)(@TJ=150°C)−IIS0(MAX)(@TJ=25°C)=60µA-125µA=-65µA and an Application Note 15 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques upper difference of IIS0(MAX) (@TJ=-40°C)−IIS0(MAX)(@TJ=25°C)=190µA-125µA=+65µA. So the individual calculated current sense offset will vary over life time and temperature from typically 50µA down towards -15µA (50µA-65µA) and up to +115µA (50µA+65µA). In case a calibrated BTS50015-1TAD (i) shows a negative calculated current sense offset, IIS0(cal)<0, then this individual sense offset will vary over temperature and stress towards the differences between the room temperature minimum limit and the respective minimum limits at “cold” (Tj=-40°C) and “hot” (Tj=+150°C).. To give an example, assuming an individual negative calculated current sense offset, IIS0(cal)=-100µA, has been derived by means of calibration at Tj=25°C, the device specific IIS0(cal) will vary over life time and temperature (-40°C<=Tj<=+150°C) between the MIN datasheet limits of IIS0. This results in an upper difference of IIS0(MIN)(@TJ=150°C)−IIS0(MIN)(@TJ=25°C)=-65µA-(-115µA)=+50µA and a lower difference of IIS0(MIN)(@TJ=-40°C)−IIS0(MIN)(@TJ=25°C)=-165µA-(-115µA)=-50µA. So the individual calculated current sense offset will vary over life time and temperature from typically -100µA down towards -150µA (100µA-50µA) and up to -50µA (-100µA+50µA). 4.2 1-point calibration One option to improve the current sense performance is to perform a 1-point calibration. The idea of 1-point calibration is to perform a manufacturing test that measures the sense current IIS(x1) at a defined load current IL(x1), where ideally IL(x1) is in the load current range where the highest accuracy needs to be achieved. Usually this manufacturing test is performed at an ambient temperature of 25°C. The measured values will be stored in the microcontroller’s non-volatile memory to be used by the application software. To state it upfront, although the 1-point calibration offers the lowest measurement effort during manufacturing of all possible calibration options, the accuracy improvements of a 1-point calibration remain in the special case of Power PROFET moderate. This results from the circumstance, that the device specific slope and offset remains unknown. Therefore certain assumptions have to be made which will under worst case conditions contribute to a remaining error. Nevertheless 1-point calibration achieves an improved current sense function compared to “default” sense accuracy. In case of Power PROFET, it specifically helps to reduce the offset error as it will be shown in the later BTS50015-1TAD example. The fundamental aspects of the 1-point calibration will be explained in the remaining portion of this section. However, if a significant accuracy improvement is required, the 2-poin calibration should be applied (and the reader should directly jump to the next chapter). In general mathematical terms 1-point calibration means, that first the point x1,y1 needs to be identified. As stated in chapter 3.1, x needs to be chosen in a way that x>-(-bMIN/mMIN) will be fulfilled. Since the slope m can not be derived from 1-point calibration and hence remains unknown, it has to be assumed that the slope m will be typical mTYP. The resulting y-intercept can then be calculated by changing Equation (1) to Equation (11) Equation (11) b = y − mTYP x Although calculated, b remains in the end an estimated value only. With the assumed values of slope m=m TYP and offset b (y-intercept) these values can be substituted into the generic equation for a line as defined in Equation (1), and then this equation can be used to determine the (x, y) values of any other point on the line. To derive the achievable accuracy of 1-point calibtarion the following aspects have to be considered. Since the typical assumed slope m can vary between certain limits, mMIN1 and mMAX1, two extreme y intercepts, b1 and b2, need to be derived for the accuracy investigation. Assuming, that the possible, individual slopes m MIN1 and mMAX1 will further vary over lifetime stress and temperaure, mMIN1 between a lower slope mMIN12 and an upper slope mMIN11, and mMAX1 between a lower slope mMAX12 and an upper slope mMAX11, certain error ranges already appear in case the extreme y-intercepts b1 and b2 would remain constant. Assuming that in addition the y-Intercepts b1 Application Note 16 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques and b2 will additionally vary within certain limits over lifetimes stress and temperature , b1 between a lower offset b1-b1 and upper offset b1+b1, and b2 between a lower offset b2-b2 and upper offset b2+b2, these error ranges further increase defining the limiting conditions for the final achievable accuracy of 1-point calibration. Figure 11 illustrates these effects. m y y Slope MAX1 (steepness) error mMIN1 y m (x1, y1) (x1, y1) (a) Identify one point Figure 11 (x1, y1) b1 b2 b x mMIN11 mMAX11 mMAX1 mMAX12 mMIN1 mMIN12 x (b) Consider worst case slope x (c) Determine y-intercept y1=mMIN11*x+b1+b1 y2=mMAX11*x+b2+b2 y (2) (4) (1) b1+b1 y4=mMIN12*x+b1-b1 b2-b2 (3) y3=mMAX12*x+b2-b2 x (d) Derive worst case xaxis error slopes X-Axis error for generic lines with varying positive slopes and offsets As it can be seen Figure 11 (d), the xi error for a given yi value will be determined by the four limiting functions y1=f(x)=mMIN11*x+b1+b1 y2=f(x)=mMAX11*x+b2+b2 y3=f(x)=mMAX12*x+b2-b2 and y4=f(x)=mMIN12*x+b1+b1. The typical function, assuming a typical slope mTYP will follow the formula y=mTYP*x+b. For low x and low y values the error of the typical function y=mTYP*x+b will vary towards a lower error. This error can be calculated solving the equation MinAbsError=(x=f(y1))-(x=f(y)). The error will also vary towards an upper error. This error can be calculated solving the equation MaxAbsError=(x=f(y3))-(x=f(y)). Whenever formula y4=f(x) results in a lower y value for a certain x value compared to y3=f(x), the maximum error will vary to an upper error, which can be calculated solving the equation MaxAbsError=(x=f(y4))-(x=f(y)). For high x value and high y values / whenever formula y2=f(x) results in a higher y value for a certain x value compared to y1=f(x), the lower error can be calculated solving the equation MinAbsError=(x=f(y2))-(x=f(y)). The absolute minimum x- error-value (MinAbsError=xi(MIN)- xi) can be calculated according Equation (12) and Equation (13). Equation (12) MinAbsError𝐶𝐴𝑆𝐸(1) = 𝑚𝑇𝑌𝑃 𝑥×+𝑏−𝑏1 −∆𝑏1 Equation (13) MinAbsError𝐶𝐴𝑆𝐸(2) = 𝑚𝑇𝑌𝑃 𝑥×+𝑏−𝑏2 −∆𝑏2 Application Note 𝑚𝑀𝐼𝑁11 𝑚𝑀𝐴𝑋11 − 𝑥 with b1 being an absolute number − 𝑥 with b2 being an absolute number 17 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques The absolute maximum x- error-value (MaxAbsError=xi(MAX)- xi) can be calculated according Equation (14) and Equation (15). Equation (14) MaxAbsError𝐶𝐴𝑆𝐸(3) = 𝑚𝑇𝑌𝑃 𝑥×+𝑏−𝑏2 +∆𝑏2 Equation (15) MaxAbsError𝐶𝐴𝑆𝐸(4) = 𝑚𝑇𝑌𝑃 𝑥×+𝑏−𝑏1 +∆𝑏1 𝑚𝑀𝐴𝑋12 𝑚𝑀𝐼𝑁12 − 𝑥 with b2 being an absolute number − 𝑥 with b1 being an absolute number In addition it has to be considered, that there are enveloping conditions which will never be exceeded. In reference to Figure 10 positive values resulting from y1=f(x) and y2=f(x) will never exceed yMAX=f(x)=mMAX*x+bMAX. So whenever y1=f(x) and y2=f(x) would cause a higher y-value for a given-x value, y will be limited to yMAX=mMAX*x+bMAX. In reference to Figure 10 negative values resulting from y3=f(x) or y4=f(x) will never “exceed” yMIN=f(x)=mMIN*x+bMIN. (with bMIN being negative). So whenever y3=f(x) or y4=f(x) would cause a lower y-value for a given-x value, y will be limited to yMIN=mMIN*x+bMIN. These beneficial circumstances and high mMAX and bMAX as well as low mMIN and bMIN values will however have no effect on the maximum achievable accuracy since the overall accuracy will be defined by possible devices that will neither violate yMAX=f(x)=mMAX*x+bMAX nor yMIN=f(x)=mMIN*x+bMIN. To show the possible accuracy improvement another BTS50015-1TAD example is outlined. Although the example is based on a very specific sense current measurement, the resulting load current errors are valid also for other sense current measurements as long as the load current IL(cal) , the IIS0(cal) and (dkKILIS(cal)) values remain. Assuming an individual sense current of a BTS50015-1TAD was measured at Tj=25°C and IL=20A and showed a value of IIS=0.5mA. Considering that according to the datasheet the current sense differential ratio dkILIS (Parameter P_6.1.41) varies in a range of MIN 45300, TYP51500 and MAX57700 the current sense ratio spread of (dkKILIS(CAL)) varies between MIN-5% and MAX+5% and that IIS0(CAL) for any positive calculated sense offset current varies according to the differences of the temperature dependent limits of Parameter P_6.1.42 between -65µA up to +65µA The following values can be derived: The assumed typical current sense function follows the formula: IIS=IL/dkILIS(TYP)+IIS(0)=IL/51500+1.117E-4 with IIS(0)= IIS(x1)- IL(x1)/dkILIS(TYP)=0.0005-20/51500=1.117E-4 To derive the achievable accuracy of 1-point calibration mMAX1, mMAX11 and mMAX12 can be derived from dkILIS(MIN) and (dkKILIS(CAL)). mMAX11=1/dkILIS(MIN)=1/45300=2.20751E-5, mMAX1=(1-(dkKILIS(cal)))/dkILIS(MIN)=0.95*mMAX11=2.097E-5, mMAX12=(1-(dkKILIS(cal)))/dkILIS(MIN)/(1+(dkKILIS(CAL)))=mMAX1/1.05=1.9973E-5 Application Note 18 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques mMIN1, mMIN11 and mMIN12 can be derived from dkILIS(MAX) and (dkKILIS(CAL)). mMIN12=1/dkILIS(MAX)=1/57700=1.7331E-5 mMIN1=(1+(dkKILIS(cal)))/dkILIS(MAX)=1.05*mMIN12=1.82E-5 mMIN11=(1+(dkKILIS(cal)))/dkILIS(MAX)/(1-(dkKILIS(cal))=mMIN1/0.95=1.91553E-5 the offset calculation will result in b=1.117E-4, b1=1.36E-4 and b2=8.057E-5 Table 4 shows the calculated absolute and relative errors for certain load currents for the example of BTS50015-1TAD with a calibration at Tj=25°C and IL=20A Application Note 19 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques Table 4 ILoad minimum absolute minimum relative ILoad error ILoad error Maximum maximum relative absolute ILoad error ILoad error 10A -4.5A -45% +4.5A +45% 20A -4.4A -22% +4.8A +24% 40A -6.4A -16% +7.2A +18% 60A -8.8A -15% +9.6A +16% 80A -11.2A -14% +12A +15% 100A -13.6A -14% +14.1A +14% Table 5 shows the calculated absolute and relative errors for certain load currents for an additional example of BTS50015-1TAD with a calibration at Tj=25°C and IL=40A. Table 5 ILoad minimum absolute minimum relative ILoad error ILoad error Maximum maximum relative absolute ILoad error ILoad error 10A -5.8A -58% +6.1A +61% 20A -5.7A -28% +5.8A +29% 40A -5.4A -13% +5.8A +14% 60A -7.4A -12% +8.2A +14% 80A -9.8A -12% +10.6A +13% 100A -12.2A -12% +13A +13% Table 6 shows the calculated absolute and relative errors for certain load currents for an additional example of BTS50015-1TAD with a calibration at Tj=25°C and IL=10A. Table 6 ILoad minimum absolute minimum relative ILoad error ILoad error Maximum maximum relative absolute ILoad error ILoad error 10A -3.9A -39% +4.3A +43% 20A -4.6A -23% +5.5A +27% 40A -7.1A -18% +7.9A +20% 60A -9.5A -16% +10.3A +17% 80A -11.9A -15% +12.7A +16% 100A -14.3A -14% +15.1A +15% Application Note 20 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques Comparing the content of 0 to Table 6 with Table 3 it can be seen that: 1-point calibration does improve the accuracy compared to “default” performance accuracy improves compared to “default” performance mostly near the load current, where the calibration was performed a calibration at a load current just above IL0 (see chapter 3.1) is recommended, since the accuracy improvements at medium and high currents (IL>=40) will reach similar accuracy improvements compared to higher load current calibrations and the manufacturing test can be performed at more moderate load currents in relation to the test equipment capability. 4.3 2-point calibration The recommended option to optimally improve the current sense performance is to perform a 2-point calibration. The idea of 2-point calibration is to perform a manufacturing test that measures two sense currents IIS(x1) and IIS(x2) at two different, defined load currents IL(x1) and IL(x1). Ideally IL(x1) is just above IL0 (see chapter 3.1). IL(x2) should be chosen in a way, that IIS(x2) will sufficiently differ from IIS(x1). Usually this manufacturing test is performed at an ambient temperature of 25°C. The measured values will be stored in the microcontroller’s non-volatile memory to be used by the application software. The fundamental aspects of the 2-point calibration will be explained in the following. In general mathematical terms 2-point calibration means, that first the point x1,y1 needs to be identified followed by the second point x2, y2. Note: As stated in chapter 3.1, x1 needs to be chosen in a way that x>-(-bMIN/mMIN) will be fullfilled. With the values of x1, y1, x2 and y2 the individual slope m2 can be derived as illustrated in Figure 12 and Equation (16). y (x2, y2) (x1, y1) y y (x2, y2) m2 (x2, y2) m2 (x1, y1) (x1, y1) b2 x (a) Identify two points Figure 12 Equation (16) Application Note x (b) Determine slope x (c) Determine y-intercept Determining the characteristics of a straight line m2 = y2 −y1 x2 −x1 21 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques Once the slope has been determined as illustrated in Figure 12(b), this value can be used to calculate the yintercept (b2) as illustrated in Figure 12(c). This can be accomplished by picking the (x,y) values for any point and solving for the y-intercept using Equation (17). Equation (17) b2 = y − m2 × x After m2 (slope) and b2 (y-intercept) have been determined, these values can be substituted into the generic equation for a line as defined in Equation (1), and then this equation can be used to determine the (x, y) values of any other point on the line. To derive the achievable accuracy of 2-point calibration the following aspects have to be considered. According to chapter 4.1 the derived slope m2 may vary between certain limits mMIN2 and mMAX2. Also the derived yIntercepts b2 may vary within certain limits towards a lower offset b2-b and an upper offset b2+b. Special attention has to be paid here that the variation of b may vary depending on the case, whether b2 is positive or negative. Figure 13 illustrates these effects for an example where b2 is positive. y Slope mMAX2 (steepness) y error (x2, y2) (x2, y2) mMIN2 m2 (x1, y1) (x1, y1) b2 x x (b) Derive offset and consider worst case slope (a) Identify two points Figure 13 y b2+b b2 b2-b y1=mMAX2*x+b2+b y=m2*x+b2 y2=mMIN2*x+b2-b x (c) Derive worst case xaxis error slopes X-Axis error for generic lines with varying positive slopes and offsets As it can be seen, the xi error for a given yi value will be determined by two limiting functions y1=f(x) and y2=f(x). y1=f(x)=mMAX2*x+b2+b y2=f(x)=mMIN2*x+b2-b The typical function, assuming a typical slope m2 will follow the formula y=m2*x+b2. Comparing Figure 13 with Figure 10 it can be seen, that the general absolute error for any xi value will again range between a minimum value xi(MIN) and a maximum xi(MAX). Compared to “default” current sense performance however the error variation will reduce as following: The absolute minimum x- error-value (MinError=xi(MIN)- xi) can be calculated according Equation (18) Equation (18) Application Note MinAbsError = 𝑥×(𝑚2 −𝑚𝑀𝐴𝑋2 )−∆𝑏 𝑚𝑀𝐴𝑋2 22 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques The relative minimum x-error (MinRelError=( xi(MIN)- xi)/xi) can be calculated according Equation (19) Equation (19) MinRelError = 𝑚2 𝑥×−∆𝑏 𝑚𝑀𝐴𝑋2 ×𝑥 −1 = 𝑚2 𝑚𝑀𝐴𝑋2 − ∆𝑏 𝑚𝑀𝐴𝑋2 ×𝑥 −1 The absolute maximum x- error-value (MaxError=xi(MAX)- xi) can be calculated according Equation (20) Equation (20) MaxAbsError = 𝑥×(𝑚2 −𝑚𝑀𝐼𝑁2 )+∆𝑏 with b being an absolute number. 𝑚𝑀𝐼𝑁2 The relative maximum x-error (MaxRelError=( xi(MAX)- xi)/xi) can be calculated according Equation (21) Equation (21) MaxRelError = 𝑚2 𝑥×+∆𝑏 𝑚𝑀𝐼𝑁2 ×𝑥 −1= 𝑚2 𝑚𝑀𝐼𝑁2 + ∆𝑏 𝑚𝑀𝐼𝑁2 ×𝑥 − 1 with b being an absolute number Applying Power PROFET parameters to Equation (18) to Equation (21) results in Equation (22) to Equation (25). Equation (22) (1 + MinAbsError = (𝐼𝐿 × ( 1 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) 1 − 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ×(1+ 𝑀𝑖𝑛.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 100% )− ∆𝐼𝐼𝑆0(𝑐𝑎𝑙) 106 ) ) × 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) × 𝑀𝑖𝑛.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 100% ) with Min.(dKIILIS(cal)) being a negative % value and with IIS0(cal) as µA value. 𝑀𝑖𝑛.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) Equation (23) MinRelError = 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ×(1+ 100% ) − 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ∆𝐼𝐼𝑆0(𝑐𝑎𝑙) ×𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ×(1+ 𝑀𝑖𝑛.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 100% ) 𝐼𝐿 ×106 −1 with Min.(dKIILIS(cal)) being a negative % value and with IIS0(cal) as µA value. Equation (24) (1 + MaxAbsError = (𝐼𝐿 × ( 1 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) 1 − 𝑀𝑎𝑥.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ×(1+ 100% )− ) ∆𝐼𝐼𝑆0(𝑐𝑎𝑙) 106 ) × 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) × 𝑀𝑎𝑥.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 100% ) with Max.(dKIILIS(cal)) being a positive % value and with IIS0(cal) as µA value being negative. Equation (25) MaxRelError = 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ×(1+ 𝑀𝑎𝑥.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 100% ) − 𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ∆𝐼𝐼𝑆0(𝑐𝑎𝑙) ×𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ×(1+ 𝑀𝑎𝑥.∆(𝑑𝐾𝐼𝐿𝐼𝑆(𝑐𝑎𝑙) ) 100% 𝐼𝐿×106 ) −1 with Max.(dKIILIS(cal)) being a positive % value and with IIS0(cal) as µA value being negative. To show the possible accuracy improvement another BTS50015-1TAD example is outlined. Although the example is based on a very specific sense current measurement, the resulting load current errors are valid also for other sense current measurements as long as the IIS0(cal) and (dkKILIS(cal)) values remain. Application Note 23 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibration Techniques Assuming an individual sense current of a BTS50015-1TAD was measured at Tj=25°C and IL1=10A and IL2=20A showed a values of IIS1=0.3mA and IIS2=0.51mA. Considering that according the datasheet the current sense ratio spread of (dkKILIS(cal)) varies between MIN-5% and MAX+5% and that IIS0(cal) for any positive calculated sense offset current varies according to the differences of the temperature dependant limits of Parameter P_6.1.42 between -65µA<=IIS0(cal)<=+65µA (see also chapter 4.1) The following values can be drived: The assumed typical current sense function follows the formula: IIS=IL/dkILIS(cal)+IIS(0)=IL/47620+9E-5 with m=(y2-y1)/(x2-x1)= =(IIS(2)- IIS(1))/( IL(2)- IL(1))=( (5.1E-4-3E-4)/(20-10)=2.1E-5 or dkILIS(cal)=1/m=47620 And with IIS(0)= IIS(x1)- IL(x1)/dkILIS=0.0003-10/47620=9E-5 To derive the achievable accuracy of 1-point calibtarion mMAX2 can be derived from m2 and (dkKILIS(cal)). mMAX2=m2/(1-(dkKILIS(cal)))=2.1E-5/0.95=2.211E-5 or dkILIS(MIN)=1/mMAX2=45238 mMIN2 can be derived from m2 and (dkKILIS(CAL)). mMIN2 m2/(1+(dkKILIS(cal)))= 2.1E-5/1.05=2E-5 or dkILIS(MIN)=1/mMAX2=50000 the offset calculation will result in b2+b=9E-5+6.5E-5=1.55E-4 and b2-b=9E-5-6.5E-5=-2.5E-5 Table 7shows the calculated absolute and relative errors for certain load currents for the example of BTS500151TAD with a calibration at Tj=25°C and at IL1=10A and IL2=20A. Table 7 ILoad minimum absolute minimum relative ILoad error ILoad error Maximum maximum relative absolute ILoad error ILoad error 10A -3.4A -34% +3.8A +38% 20A -3.9A -20% +4.3A +21% 40A -4.9A -12% +5.3A +13% 60A -5.9A -10% +6.3A +10% 80A -6.9A -9% +7.3A +9% 100A -7.9A -8% +8.3A +8% Comparing the content of Table 7 with 0 to Table 6 and Table 3 it can be seen that: 2-point calibration does achieve the best accuracy Also at 2-point calibration the accuracy does depend on the load current. The higher the load current is, the smaller the relative error and the better the overall accuracy will be. At very high load currents an accuracy of ~8% can be achieved. Application Note 24 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Important considerations 5 Important considerations In order to utilize the sense performance of Power PROFET, independent whether calibration is applied or not, the following conditions have to be maintained: voltage conditions: in order to ensure that the sense circuitry works in the specified range a voltage drop between the VS pin and the IS pin of minimum 5V is required (VS-VIS>=5V). This condition can have an impact on the selection of the external Sense Resistor RIS especially if the Power PROFET has to operate at low supply voltages (VS<=10V provided the IS signal is read and processed by a 5V micro controller). measurement range: the load current range, in which the current sense function can be used – independent whether or not calibration is applied - is limited towards a lower load current IL0 and towards an upper load current IL4 (see chapter 3.1). timings: o whenever the Power Profet is commanded on by applying a positive IN signal, a certain time has to be considered to allow the device to switch on and to allow the sense current to provide a stable sense signal. According to the datasheet parameter P_6.1.48 the sense pin will provide a 90% value of the final steady state value within the time tpIS(ON)_90. According to the datasheet parameter P_6.1.49 the sense pin will provide the steady state value latest after tpIS(ON). o whenever the Power Profet is already in on-state and the load current changes, a certain time has to be considered to allow the sense circuit to adjust the sense current to the new, steady state sense signal. According to the datasheet parameter P_6.1.51 the sense pin will provide the new, steady state value within the time tpIS(LC). Application Note 25 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET 6 Calibrating Power PROFET 6.1 Calibration Nomenclature and Equations The nomenclature used in the high-side power switch datasheets and the information presented earlier in this application note references calibration information in terms of current. However, the analog-to-digital converter (ADC) in the microcontroller that is used to monitor the IS (sense current) output from the high-side switch reads voltages, not currents. Thus, the calibration techniques discussed below are presented in terms of voltages because these are what the manufacturing test and application software read. Consider the reference circuit illustrated in Figure 14 (the resistors RINPUT and RSENSE are for protection and have no or minimal effect on the calibration calculations). VS +5V RINPUT VS IN µC (e.g. XC866) IL OUT RSENSE IS GND IIS + VIS – IL RL RIS GND Figure 14 Reference circuit for calibration nomenclature The analog sense current signal IIS flows through resistor RIS. The corresponding voltage potential VIS, which is developed across this resistor, and which is seen by the microcontroller’s ADC input, is determined by Ohm’s law as shown in Equation (26). Equation (26) VIS = IIS x R IS With the exception of the No Calibration scenario discussed later in this application note, the initial values for dkILIS(cal) and VIS0(cal) will be determined by manufacturing test and stored in the microcontroller’s non-volatile memory for use by the application software. Note: This application note assumes that manufacturing test will store VIS0(cal) (the voltage value in ADC counts) in the microcontroller’s non-volatile memory; that is, it is assumed that manufacturing test will NOT store IIS0(cal) (the current value). Application Note 26 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET In case of a 1-point calibration dkILIS(cal) is assumed to be dkILIS(TYP), i.e. the typical value of datasheet parameter P_6.1.41. Equation (27) d𝑘ILIS(cal) = d𝑘ILIS(TYP) The individual offset is calculated Equation (28) VIS0(cal) = IIS0(cal) ∗ R IS = (IIS(x1) − IL(x1) dkILIS(TYP) ) ∗ R IS with VIS0(cal) being positive or negative In case of a 2-point calibration dkILIS(cal) is calculated as shown in Equation (29). Equation (29) d𝑘ILIS = IL1 − IL2 V (I ) V (I ) ( IS L1 ⁄R )− ( IS L2 ⁄R ) IS IS The load current IL can be calculated changing Equation (2) as following Equation (30) IL = d𝑘ILIS(cal) x (IIS − IIS0(cal) ) Considering and Equation (26) and Equation (28) the load current can be calculated by Equation (31) IL = d𝑘ILIS(cal) x ( VIS RIS − VIS0(cal) RIS ) Factoring Equation (31) allows the application software to calculate the load current IL as shown in Equation (32). Equation (32) IL = 𝑑𝑘KLIS(cal) RIS x (VIS − VIS0(cal) ) with VIS0(cal) being positive or negative 6.2 Application Software Implementation 6.2.1 No Calibration (No Cal) With this calibration option, no calibration is performed by manufacturing test; thus, no individual varying device values for VIS0(cal) and dkILIS(cal) are stored in the microcontroller’s non-volatile memory. Instead, the application developer simply sets VIS0(cal) to zero, dkILIS(TYP) as specified in the datasheet and RIS according the Application Note 27 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET used value. This scheme is the least expensive in terms of time and manufacturing cost, but it also yields the least accuracy. The term DUT (Device Under Test) refers to the high-side power switch that is being calibrated by manufacturing test or measured by the application software. The flowchart in Figure 15 summarizes the process used by the application software when the No Calibration option is being used. During normal output turn ON Use dkILIS(TYP) from datasheet Use VIS0(cal)=0 from NVM Use RIS from NVM Delay for t > tsIS(ON)_90 (ideally t > tsIS(ON)) Convert/Read DUT (VIS) with µC ADC Calculate IL using Equation (32) Figure 15 Use result IL for diagnostics and protection Application software procedure for No Calibration option During normal device/load turn-on cycles, the software reads the IS pin from the ADC after delaying for the current sense settling time. It then uses the datasheet values for dkILIS(TYP) to calculate the load current IL using Equation (32). The application software would then compare the calculated load current value to diagnostic threshold limits stored in the microcontroller’s non-volatile memory to determine the load condition (normal, short-to-battery, short-to-ground, etc.) 6.2.2 1-point Calibration With this calibration option, manufacturing test measures the value of VIS(x1) at load current IL(x1) at an ambient temperature of 25°C. This measured value will be further processed into “calculated”, calibrated sense offset VIS0(cal) which will be stored in the microcontroller’s non-volatile memory along with the typical datasheet value of dkILIS(cal)=dkILIS(TYP) and RIS. These are the values that will be used by the application software. Single-point calibration involves switching a known load at a known temperature (typically 25°C) and then measuring the analog sense current. With conventional high-side switches, the polarity of the offset must be determined and tracked such that the software can add or subtract the offset value from the measured values. Figure 16 summarizes the process used by manufacturing test when the 1-point calibration option is being used. Application Note 28 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET Switch in Load #1 (IL1) to DUT output to calibrate Figure 16 Turn DUT output ON to calibrate Calculate VIS0(cal) using Equation (28) Delay for t > tsIS(ON) Store VIS0(cal) for future calculations Convert/Read DUT ( VIS(IL1) ) With µC ADC Store RIS & dKILIS(TYP) in NVM Manufacturing test procedure for the 1-point calibration option The manufacturing test turns the device input ON with a known load connected to the device, delays for the current sense settling time and then reads and stores the corresponding VIS(IL1) value. Next the manufacturing test software calculates the offset using Equation (28) and stores this value plus the datasheet values of dkIILIS(typ) and the RIS value in the microcontroller’s non-volatile memory (NVM). Figure 17 summarizes the process used by the application software when the 1-point calibration option is being used. During normal output turn ON Use dkILIS(TYP) from NVM Use VIS0(cal)) from NVM Use RIS from NVM Delay for t > tsIS(ON)_90 (ideally t > tsIS(ON)) Convert/Read DUT (VIS) with µC ADC Calculate IL using Equation (32) Figure 17 Application Note Use result IL for diagnostics and protection Application software procedure for the 1-point calibration option 29 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET During normal device/load turn-on cycles, the software reads the IS pin from the ADC after delaying for the current sense settling time. It then uses the values for dkILIS(TYP), VIS0(cal) and RIS stored in the microcontroller’s non-volatile memory to calculate the load current IL using Equation (32). This load current is then compared to normal or faulted threshold limits to determine the condition of the load. 6.2.3 2-Point Calibration With this calibration option, manufacturing test measures two values of VIS(x) at two different load currents IL(x) at an ambient temperature of 25°C. Both of these measured values, VIS(x1) at low load current IL(x1) and VIS(x2) at higher load current IL(x2) will be stored in the microcontroller’s non-volatile memory to be used by the application software. Figure 18 summarizes the process used by manufacturing test when the 2-Point calibration option is being used. Figure 18 Switch in Load #1 (IL1) to DUT output to calibrate Switch in Load #2 (IL2) to DUT output to calibrate Turn DUT output ON to calibrate Delay for t > tsIS(LC) Delay for t > tsIS(ON) Convert/Read DUT ( VIS(IL2) ) With µC ADC Convert/Read DUT ( VIS(IL1) ) With µC ADC Store VIS(IL2) for future calculations Calculate VIS(cal) using Equation (28) Store VIS(IL1) for future calculations Calculate dkILIS(cal) using Equation (29) Store dkILIS0(cal), VIS0(cal), RIS in NVM Manufacturing test procedure for the 2-Point calibration option The manufacturing test turns the device input ON with a known load connected to the device, delays for the current sense settling time, and then reads and stores the corresponding VIS(IL1) value. Then manufacturing test changes to a higher current rated load, delays for the current sense settling time, and then reads and stores the corresponding VIS(IL2) value again. Next the manufacturing test software calculates the slope using Equation (29) and the offset using Equation (28) and stores these value plus the RIS value in the microcontroller’s nonvolatile memory (NVM). Application Note 30 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET Figure 19 summarizes the process used by the application software when the 2-Point calibration option is being used. During normal output turn ON Use dkILIS(cal) from NVM Use VIS0(cal)) from NVM Use RIS from NVM Delay for t > tsIS(ON)_90 (ideally t > tsIS(ON)) Convert/Read DUT (VIS) with µC ADC Calculate IL using Equation (32) Figure 19 Use result IL for diagnostics and protection Application software procedure for the 2-Point calibration option During normal device/load turn-on cycles, the software reads the IS pin from the ADC after delaying for the current sense settling time. It then uses the values for dkILIS(cal), VIS0(cal) and RIS stored in the microcontroller’s nonvolatile memory to calculate the load current IL using Equation (32). This load current is then compared to normal or faulted threshold limits to determine the condition of the load. Application Note 31 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Calibrating Power PROFET 6.3 Accuracy of Different Calibration Options Figure 20 illustrates the accuracy provided by the various calibration options discussed above. In the case of the sense current graphs, the red lines represent the typical slopes, the blue lines represent the maximum deviation from typical, and the green lines represent the minimum deviation from typical. Sense current Sense current percentage error +150% +100% +50% 0% -50% -100% -150% IIS (a) No calibration IL IL +150% +100% +50% 0% -50% -100% -150% IIS (b) 1-Point IL (c) 2-Point IL Figure 20 ~+15% ~-15% IL +150% +100% +50% 0% -50% -100% -150% IIS ~+22% ~-22% ~+8% ~-8% IL Accuracy of different calibration options The sense current percentage error graphs clearly shows that 2-point calibration, although it requires the highest effort during manufacturing test, offers the best accuracy performance. Application Note 32 V1.0 2016-04-07 Improved SENSE Calibration and Benefits Guide Conclusion 7 Conclusion Current sensing is a well-accepted feature in high-side power switches. Devices with a traditional concept, to which also Power PROFET belongs, have an offset current that deteriorates the current sense accuracy, especially at lower load currents, and that may disable the current sense functionality below certain load current thresholds IL(0). In case of Power PROFET the “default” current sense performance offers a moderate accuracy. Whenever this accuracy needs to be improved, 1-point or 2-point calibration can bring a significant accuracy improvement especially at higher load currents thanks to the nature of Power PROFET sense variations. Nevertheless, even with calibration the measureable load current range remains in the exact same range of IL = IL(0) to IL(4) as in the case of the “default” sense performance. Revision History V1.0, 2016-04-07 (Major changes since the last revision) Page or Reference Application Note Description of change 33 V1.0 2016-04-07 Trademarks of Infineon Technologies AG µHVIC™, µIPM™, µPFC™, AU-ConvertIR™, AURIX™, C166™, CanPAK™, CIPOS™, CIPURSE™, CoolDP™, CoolGaN™, COOLiR™, CoolMOS™, CoolSET™, CoolSiC™, DAVE™, DI-POL™, DirectFET™, DrBlade™, EasyPIM™, EconoBRIDGE™, EconoDUAL™, EconoPACK™, EconoPIM™, EiceDRIVER™, eupec™, FCOS™, GaNpowIR™, HEXFET™, HITFET™, HybridPACK™, iMOTION™, IRAM™, ISOFACE™, IsoPACK™, LEDrivIR™, LITIX™, MIPAQ™, ModSTACK™, my-d™, NovalithIC™, OPTIGA™, OptiMOS™, ORIGA™, PowIRaudio™, PowIRStage™, PrimePACK™, PrimeSTACK™, PROFET™, PRO-SIL™, RASIC™, REAL3™, SmartLEWIS™, SOLID FLASH™, SPOC™, StrongIRFET™, SupIRBuck™, TEMPFET™, TRENCHSTOP™, TriCore™, UHVIC™, XHP™, XMC™ Trademarks updated November 2015 Other Trademarks All referenced product or service names and trademarks are the property of their respective owners. Edition 2016-04-07 Published by Infineon Technologies AG 81726 Munich, Germany ©ifx1owners. 2016 Infineon Technologies AG. All Rights Reserved. Do you have a question about this document? Email: [email protected] Document reference IMPORTANT NOTICE The information contained in this application note is given as a hint for the implementation of the product only and shall in no event be regarded as a description or warranty of a certain functionality, condition or quality of the product. Before implementation of the product, the recipient of this application note must verify any function and other technical information given herein in the real application. Infineon Technologies hereby disclaims any and all warranties and liabilities of any kind (including without limitation warranties of non-infringement of intellectual property rights of any third party) with respect to any and all information given in this application note. The data contained in this document is exclusively intended for technically trained staff. It is the responsibility of customer’s technical departments to evaluate the suitability of the product for the intended application and the completeness of the product information given in this document with respect to such application. For further information on the product, technology, delivery terms and conditions and prices please contact your nearest Infineon Technologies office (www.infineon.com). WARNINGS Due to technical requirements products may contain dangerous substances. For information on the types in question please contact your nearest Infineon Technologies office. Except as otherwise explicitly approved by Infineon Technologies in a written document signed by authorized representatives of Infineon Technologies, Infineon Technologies’ products may not be used in any applications where a failure of the product or any consequences of the use thereof can reasonably be expected to result in personal injury.