TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 1 TMC603A – DATASHEET Three phase motor driver with BLDC back EMF commutation hallFX™ and current sensing ® TRINAMIC Motion Control GmbH & Co. KG Sternstraße 67 D – 20357 Hamburg GERMANY www.trinamic.com 1 Features The TMC603 is a three phase motor driver for highly compact and energy efficient drive solutions. It contains all power and analog circuitry required for a high performance BLDC motor system. The TMC603 is designed to provide the frontend for a microcontroller doing motor commutation and control algorithms. It directly drives 6 external N-channel MOSFETs for motor currents up to 30A and up to 50V and integrates shunt less current measurement, by using the MOSFETs channel resistance for sensing. Integrated hallFX™ (pat.) allows for sensorless commutation. Protection and diagnostic features as well as a step down switching regulator further reduce system cost and increase reliability. Highlights Up to 30A motor current, up to 50V operating voltage 3.3V or 5V interface 8mm x 8mm QFN package Integrated dual range high precision current measurement amplifiers Supports shunt less current measurement using power MOS transistor RDSon hallFX™ sensorless back EMF commutation emulates hall sensors Integrated break-before-make logic: No special microcontroller PWM hardware required EMV optimized current controlled gate drivers – up to 150mA possible Overcurrent / short to GND and undervoltage protection and diagnostics integrated Internal QGD protection: Supports latest generation of power MOSFETs Integrated supply concept: Step down switching regulator up to 500mA / 300kHz Common rail charge pump allows for 100% PWM duty cycle Applications Motor driver for industrial applications Integrated miniaturized drives Robotics High-reliability drives (dual position sensor possible) Pump and blower applications with sensorless commutation Motor type 3 phase BLDC, stepper, DC motor Sine or block commutation Rotor position feedback: Sensorless, encoder or hall sensor, or any mix *) note: The term TMC603 in this datasheet refers to the TMC603A and TMC603 Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Life support policy TRINAMIC Motion Control GmbH & Co. KG does not authorize or warrant any of its products for use in life support systems, without the specific written consent of TRINAMIC Motion Control GmbH & Co. KG. Life support systems are equipment intended to support or sustain life, and whose failure to perform, when properly used in accordance with instructions provided, can be reasonably expected to result in personal injury or death. © TRINAMIC Motion Control GmbH & Co. KG 2009 Information given in this data sheet is believed to be accurate and reliable. However no responsibility is assumed for the consequences of its use nor for any infringement of patents or other rights of third parties which may result from its use. Specifications subject to change without notice Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG 2 TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 3 2 Table of contents 1 FEATURES .......................................................................................................................................... 1 2 TABLE OF CONTENTS ........................................................................................................................ 3 3 SYSTEM ARCHITECTURE USING THE TMC603 ................................................................................... 5 4 PINOUT ............................................................................................................................................. 6 4.1 4.2 5 PACKAGE CODES ............................................................................................................................. 6 PACKAGE DIMENSIONS QFN52........................................................................................................... 7 TMC603 FUNCTIONAL BLOCKS .......................................................................................................... 8 5.1 BLOCK DIAGRAM AND PIN DESCRIPTION................................................................................................ 8 5.2 MOSFET DRIVER STAGE ................................................................................................................ 10 5.2.1 Principle of operation ....................................................................................................... 10 5.2.2 Break-before-make logic ................................................................................................... 11 5.2.3 PWM control via microcontroller ...................................................................................... 12 5.2.4 Slope control .................................................................................................................... 13 5.2.5 Reverse capacity (QGD) protection .................................................................................... 14 5.2.6 Considerations for QGD protection ................................................................................... 15 5.2.7 Effects of the MOSFET bulk diode ..................................................................................... 16 5.2.8 Adding Schottky diodes across the MOSFET bulk diodes ................................................. 16 5.2.9 Short to GND detection .................................................................................................... 17 5.2.10 Error logic ......................................................................................................................... 17 5.2.11 Paralleling gate drivers for higher gate current ............................................................... 18 5.3 CURRENT MEASUREMENT AMPLIFIERS .................................................................................................. 19 5.3.1 Current measurement timing............................................................................................ 20 5.3.2 Auto zero cycle ................................................................................................................. 20 5.3.3 Measurement depending on chopper cycle ...................................................................... 21 5.3.4 Compensating for offset voltages .................................................................................... 21 5.3.5 Getting a precise current value using MOSFET on-resistance ........................................... 21 5.4 HALLFX™ SENSORLESS COMMUTATION ............................................................................................... 22 5.4.1 Adjusting the hallFX™ spike suppression time ................................................................ 22 5.4.2 Adjusting the hallFX™ filter frequency ............................................................................. 23 5.4.3 Block commutation chopper scheme for hallFX™ ............................................................ 23 5.4.4 Start-up sequence for the motor with forced commutation ............................................ 24 5.5 POWER SUPPLY ............................................................................................................................. 26 5.5.1 Switching regulator .......................................................................................................... 26 5.5.2 Charge pump .................................................................................................................... 28 5.5.3 Filter capacitors for switching regulator and charge pump ............................................. 28 5.5.4 Supply voltage filtering and layout considerations ......................................................... 28 5.5.5 Reverse polarity protection ............................................................................................... 29 5.5.6 Standby with zero power consumption ........................................................................... 29 5.5.7 Low voltage operation down to 9V ................................................................................. 29 5.6 TEST OUTPUT................................................................................................................................ 30 5.7 ESD SENSITIVE DEVICE ................................................................................................................... 30 6 ABSOLUTE MAXIMUM RATINGS ..................................................................................................... 31 7 ELECTRICAL CHARACTERISTICS ...................................................................................................... 31 7.1 7.2 8 OPERATIONAL RANGE ..................................................................................................................... 31 DC CHARACTERISTICS AND TIMING CHARACTERISTICS ........................................................................... 32 DESIGNING THE APPLICATION ...................................................................................................... 39 8.1 CHOOSING THE BEST FITTING POWER MOSFET .................................................................................... 39 8.1.1 Calculating the MOSFET power dissipation ...................................................................... 40 Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 8.2 8.3 8.4 9 4 MOSFET EXAMPLES....................................................................................................................... 41 PROGRAMMING A BLOCK COMMUTATION FOR HALLFX™ ......................................................................... 42 DRIVING A DC MOTOR WITH THE TMC603 ......................................................................................... 42 TABLE OF FIGURES ......................................................................................................................... 43 10 10.1 REVISION HISTORY .................................................................................................................... 44 DOCUMENTATION REVISION ......................................................................................................... 44 Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 5 3 System architecture using the TMC603 POWER TMC603A NFET power MOS half bridges slope control BUS / IO slope HS 12V step down regulator slope LS 1 of 3 shown 5V linear regulator +VM DRIVER SECTION HS HS-drive N S break before make logic micro controller gate off detection BLDC motor LS LS-drive RS bridge current measurement short to GND detection HallFXTM for sensorless commutation position sensor RS1 short to GND 1,2,3 RS2 RS3 optional shunt resistors error logic figure 1: application block diagram The TMC603 is a BLDC driver IC using external power MOS transistors. Its unique feature set allows equipping inexpensive and small drive systems with a maximum of intelligence, protection and diagnostic features. Control algorithms previously only found in much more complex servo drives can now be realized with a minimum of external components. Depending on the desired commutation scheme and the bus interface requirements, the TMC603 forms a complete motor driver system in combination with an external 8 bit processor or with a more powerful 32 bit processor. A simple system can work with three standard PWM outputs even for sine commutation! The complete analog amplification and filtering frontend as well as the power driver controller are realized in the TMC603. Its integrated support for sine commutation as well as for back EMF sensing saves cost and allows for maximum drive efficiency. The external microcontroller realizes commutation and control algorithms. Based on the position information from an encoder or hall sensors, the microcontroller can do block commutation or sine commutation with or without space vector modulation and realizes control algorithms like a PID regulator for velocity or position or field oriented control based on the current signals from the TMC603. For sensorless commutation, the microcontroller needs to do a forward controlled motor start without feedback. This can be realized either using block commutation or sine commutation. A sine commutated start-up minimizes motor vibrations during start up. As soon as the minimum velocity for hallFX™ is reached, it can switch to block commutation and drive the motor based on the hallFX™ signals. The TMC603 also supports control of three phase stepper motors as well as two phase stepper motors using two devices. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 6 GNDP HS1 BM1 LS1 VCP HS2 BM2 LS2 VLS HS3 BM3 LS3 GNDP 52 51 50 49 48 47 46 45 44 43 42 41 40 4 Pinout VLS 1 39 VCP GNDP 2 38 ENRS_TEST VM 3 37 SWOUT GND 4 36 GND RS2G 5 35 RSLP H1 6 34 CLR_ERR H2 7 33 /ERR_OUT H3 8 32 ENABLE FILT1_RS1 9 31 INV_BL FILT2_RS2 10 30 BBM_EN FILT3_RS3 11 29 SENSE_HI COSC 12 28 VCC SCCLK 13 27 SP_SUP 14 15 16 17 18 19 20 21 22 23 24 25 26 BH1 BL1 SAMPLE1 CUR1 BH2 BL2 SAMPLE2 CUR2 BH3 BL3 SAMPLE3 CUR3 5VOUT TMC 603A-LA QFN52 8mm x 8mm 0.5 pitch figure 2: pinning / QFN52 package (top view) 4.1 Package codes Type TMC603A Package QFN52 (ROHS) Temperature range -40°C ... +125°C Code/marking TMC603A-LA Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 4.2 Package dimensions QFN52 REF MIN NOM MAX A 0.80 0.85 0.90 A1 0.00 0.035 0.05 A2 - 0.65 0.67 A3 b 0.203 0.2 0.25 D 8.0 E 8.0 e 0.5 0.3 J 6.1 6.2 6.3 K 6.1 6.2 6.3 L 0.35 0.4 0.45 All dimensions are in mm. Attention: Drawing not to scale. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG 7 TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 8 5 TMC603 functional blocks 5.1 Block diagram and pin description +VM 220n 220n BAV99 (70V) BAS40-04W (40V) 4µ7 Tantal 25V 16V 100n (2x) 12V supply LSW SS16 TP0610K or BSS84 (opt. BC857) 100n (2x) 100µ (150mA with sel. transistor) VM+10V charge pump LSW: 220µH for 100kHz VM COSC RSLP ENABLE BH1 BL1 slope HS slope control RSLP: 100k -> 100mA BBM_EN VLS VCP TMC603A COSC: 470p ->100kHz INV_BL SWOUT 12V step down regulator 5V linear regulator 5VOUT VCC slope LS 100nF +VM 1 of 3 shown DRIVER SECTION D D VCP HS1 HS-drive D BM1 break before make logic D D Zener 12V BZT52B12-V/ BZV55C12 220R Gate off detection BM1 RS2G short to GND detection LS1 LS-drive short to GND 1 GNDP RS2G: 470k -> 1000ns amplification 4.5x or 18x RDS current sense LS D BM1 RS1 VCC 1 of 3 power MOS half bridges track & hold stage signed current, centered at 1/3 VCC ENRS A SAMPLE1 automatic sample point delay D D hall sensor emulation BM2 D D BM3 switched capacitor filter A D A A test logic CLR_ERR CUR1 BRIDGE CURRENT MEASUREMENT BM1 SCCLK Motor coil output opt. for high QGD FETs : MSS1P3 / ZHCS1000 VLS SENSE_HI Provide sufficient filtering capacity near bridge transistors (electrolyt 5V supply capacitors and ceramic capacitors). SENSORLESS COMMUTATION D GND DIE PAD GND ENRS_TEST undervoltage VLS, VCP short to GND 1,2,3 D H1 H2 H3 FILT1_RS1 FILT2_RS2 FILT3_RS3 /ERR_OUT error logic set reset SP_SUP CSUP: 1n -> 90µs RS2G, RSLP and BMx: Use short trace and avoid stray capacitance to switching signals. Place resistors near pin. figure 3: application diagram The application diagram shows the basic building blocks of the IC and the connections to the power bridge transistors, as well as the power supply. The connection of the digital and analog I/O lines to the microcontroller is highly specific to the microcontroller model used. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Type 9 Pin Number Function VLS 1, 44 Low side driver supply voltage for driving low side gates GNDP 2, 40, 52 Power GND for MOSFET drivers, connect directly to GND VM 3 Motor and MOSFET bridge supply voltage GND 4, 36 Digital and analog low power GND, connect directly to GND RS2G 5 AI 5V Short to GND control resistor. Controls delay time for short to GND test Hx 6, 7, 8 DO 5V hallFX™ outputs for back EMF based hall sensor emulation FILTx_ RSx 9, 10, 11 AI 5V AO 5V Output of internal switched capacitor filter or input for external sense resistor (select using pin ENRS_TEST) COSC 12 A 5V Oscillator capacitor for step down regulator SCCLK 13 DI 5V Switched capacitor filter clock input for hallFX™ filters. BHx 14, 18, 22 DI 5V High side driver control signal: A positive level switches on the high side BLx 15, 19, 23 DI 5V Low side driver control signal: Polarity can be reversed via INV_BL SAMPLEx 16, 20, 24 DI 5V Optional external control for current measurement sample/hold stage. Set to positive level, if unused CURx 17, 21, 25 AO 5V Output of current measurement amplifier 5VOUT 26 SP_SUP 27 VCC 28 Output of internal 5V linear regulator. Provided for VCC supply A 5V An external capacitor on this pin controls the commutation spike suppression time for hallFX™. +5V supply input for digital I/Os and analog circuitry SENSE_HI 29 DI 5V Switches current amplifiers to high sensitivity BBM_EN 30 DI 5V Enables internal break-before-make circuitry INV_BL 31 DI 5V Allows inversion of BLx input active level (low: BLx is active high) ENABLE 32 DI 5V Enables the power drivers (low: all MOSFETs become actively switched off) /ERR_OUT 33 DO 5V Error output (open drain). Signals undervoltage or overcurrent. Tie to ENABLE for direct self protection of the driver CLR_ERR 34 DI 5V Reset of error flip-flop (active high). Clears error condition RSLP 35 AI 5V Slope control resistor. Sets output current for MOSFET drivers SWOUT 37 O Switch regulator transistor output ENRS_ TEST 38 DI 5V O 12V Enables sense resistor inputs rather than RDSON measurement. Test multiplexer output VCP 39 LSx 41, 45, 49 O 12V Low side MOSFET driver output BMx 42, 46, 50 I (VM) Sensing input for bridge outputs. Used for MOSFET control and current measurement. HSx 43, 47, 51 O (VCP) High side MOSFET driver output Exposed die pad - GND Connect the exposed die pad to a GND plane. It is used for cooling of the IC and may either be left open or be connected to GND. Charge pump supply voltage. Provides high side driver supply Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 5.2 10 MOSFET Driver Stage The TMC603 provides three half bridge drivers, each capable of driving two MOSFET transistors, one for the high-side and one for the low-side. In order to provide a low onresistance, the MOSFET gate driving voltage is about 10V to 12V. VLS VCP TMC603 HS-DRV +VM HS1 Z 12V BM1 220R LS-DRV LS1 GNDP The TMC603 bridge drivers provide a number of unique features for simple operation, explained in the following chapters: An integrated automatic break-beforemake logic safely switches off one transistor before its counterpart can be switched on. Slope controlled operation allows adaptation of the driver strength to the desired slope and to the chosen transistors. The drivers protect the bridge actively against cross conduction (QGD protection) The bridge is protected against a short to GND +VM HS-DRV HS2 Z 12V 3 phase BLDC motor BM2 220R LS-DRV LS2 +VM HS-DRV HS3 Z 12V BM3 220R LS-DRV LS3 figure 4: three phase BLDC driver 5.2.1 Principle of operation The low side gate driver voltage is supplied by the VLS pins. The low side driver supplies 0V to the MOSFET gate to close the MOSFET, and VLS to open it. The TMC603 uses the following driver principle for driving of the high side (pat. fil.): The high-side MOSFET gate voltage is referenced to its source at the center of the half bridge. Due to this, the TMC603 references the gate drive to the bridge center (BM) and has to be able to drive it to a voltage lying above the positive bridge power supply voltage VM. This is realized by a charge pump voltage generated from the switching regulator via a Villard circuit. When closing the high-side MOSFET, the high-side driver drives it down to the actual BM potential, since an external induction current from the motor coil could force the output to stay at high potential. This is accomplished by a feedback loop and transistor TG1 (see figure). In order to avoid floating of the output BM, a low current is still fed into the HS output via transistor TG1a. The input BM helps the high side driver to track the bridge voltage. Since input pins of the TMC603 must not go below -0.7V, the input BM needs to be protected by an external resistor. The resistor limits the current into BM to a level, the ESD protection input diodes can accept. High side driver VCP +VM Ion HS On HS T1 Z 12V BM one coil of motor 220R HS Off LS TG1 TG1a Ioff Iholdoff one NMOS halfbridge figure 5: principle of high-side driver (pat. fil.) Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 11 A zener diode at the gate (range 12V to 15V) protects the high-side MOSFET in case of a short to GND event: Should the bridge be shorted, the gate driver output is forced to stay at a maximum of the zener voltage above the source of the transistor. Further it prevents the gate voltage from dropping below source level. The maximum permissible MOSFET driver current depends on the motor supply voltage: Parameter Symbol Max Unit MOSFET driver current with VVM < 30V IHSX, ILSX 150 mA 150-2.5*(VVM-30V) mA 100 mA MOSFET driver current with 30V < VVM < 50V MOSFET driver current with VVM = 50V IHSX, ILSX Pin Comments LSx Low side MOSFET driver output. The driver current is set by resistor RSLP. A Schottky protection diode to GND may be required for MOSFETs, where QGD is larger than QGS. Check that LSx voltage does not drop below GND by more than 0.5V. HSx High side MOSFET driver output. The driver current is set by resistor RSLP BMx Bridge center used for current sensing and for control of the high side driver. For unused bridges, connect BMx pin to GND and leave the driver outputs unconnected. Place the external protection resistor near the IC pin. RSLP The resistor connected to this pin controls the MOSFET gate driver current. A 40µA current out of this pin (resistor value of 100k to GND) results in the nominal maximum current at full supply range. Keep interconnection between IC and resistor short, to avoid stray capacitance to adjacent signal traces of modulating the set current. Resistor range: 60 k to 500 k VLS Low side driver supply voltage for driving low side gates VCP Charge pump supply voltage. Provides high side driver supply GNDP Power GND for MOSFET drivers, connect directly to GND BHx High side driver control signal: A positive level switches on the high side. For unused bridges, tie to GND. BLx Low side driver control signal: Polarity can be reversed via INV_BL INV_BL Allows inversion of BLx input active level (low: BLx is active high). When high, each BLx and BHx can be connected in parallel in order to use only 3 PWM outputs for bridge control. Be sure to switch on internal break-before-make logic (BBM_EN = Vcc) to avoid bridge short circuits in this case. 5.2.2 Break-before-make logic Each half-bridge has to be protected against cross conduction during switching events. When switching off the low-side MOSFET, its gate first needs to be discharged, before the high side MOSFET is allowed to be switched on. The same goes when switching off the high-side MOSFET and switching on the low-side MOSFET. The time for charging and discharging of the MOSFET gates depends on the MOSFET gate charge and the driver current set by RSLP. When the BBM logic is enabled, the TMC603 measures the gate voltage and automatically delays switching on of the opposite bridge transistor, until its counterpart is discharged. The BBM logic also prevents unintentional bridge short circuits, in case both, LSx and HSx, become switched on. The first active signal has priority. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 12 Alternatively, the required time can be calculated and pre-compensated in the PWM block of the microcontroller driving the TMC603 (external BBM control). Control signals Internal BBM control BLx 0V BHx 0V External BBM control VVLS Miller plateau LSx MOSFET drivers 0V VVM BMx 0V VVCP VVM HSx 0V HSxBMx Miller plateau VVCP VVM 0V tLSON tLSOFF tBBMHL tHSON tBBMLH Load pulling BMx down Load pulling BMx up to +VM figure 6: bridge driver timing Pin Comments BBM_EN Enables internal break-before-make circuitry (high = enable) 5.2.3 PWM control via microcontroller There are a number of different microcontrollers available, which provide specific BLDC commutation units. However, the TMC603 is designed in a way in order to allow BLDC control via standard microcontrollers, which have only a limited number of (free) PWM units. The following figure shows several possibilities to control the BLDC motor with different types of microcontrollers, and shall help to optimally adapt the TMC603 control interface to the features of your microcontroller. The hall signals and further signals, like CURx interconnection to an ADC input, are not shown. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Microcontroller with BLDC PWM unit PWM1 OUT1 PWM1 OUT2 BH1 TMC603 BL1 PWM1 OUT BH1 13 Block (Hall or hallFX) or sine commutated BLDC motor TMC603 BL1 Microcontroller with 3 PWM outputs Sine commutated BLDC motor INV_BL +VCC BBM_EN Microcontroller with 3 PWM outputs PWM1 OUT DIG OUT BH1 TMC603 BL1 PWM1 OUT DIG OUT / HI-Z BH1 2k2 Microcontroller with 3 PWM outputs Block (Hall) commutated BLDC motor TMC603 BL1 Block (Hall) or sine commutated BLDC motor INV_BL +VCC BBM_EN Microcontroller with 1 PWM output DIG OUT BH1 DIG OUT PWM1 OUT BL1 TMC603 ENABLE Block (hallFX) commutated BLDC motor 2k2 DIG IN /ERR_OUT +VCC BBM_EN figure 7: examples for microcontroller PWM control 5.2.4 Slope control The TMC603 driver stage provides a constant current output stage slope control. This allows to adapt driver strength to the drive requirements of the power MOSFET and to adjust the output slope by providing for a controlled gate charge and discharge. A slower slope causes less electromagnetic emission, but at the same time power dissipation of the power transistors rises. The duration of the complete switching event depends on the total gate charge. The voltage transition of the output takes place during the so called miller plateau (see figure 6). The miller plateau results from the gate to drain capacity of the MOSFET charging / discharging during the switching. From the datasheet of the transistor (see example in figure 8) it can be seen, that the miller plateau typically covers only a part (e.g. one quarter) of the complete charging event. The gate voltage level, where the miller plateau starts, depends on the gate threshold voltage of the transistor and on the actual load current. 10 25 VM 8 20 6 15 4 10 2 5 0 0 5 10 15 20 VDS – Drain to source voltage (V) VGS – Gate to source voltage (V) MOSFET gate charge vs. switching event 0 25 QMILLER QG – Total gate charge (nC) figure 8: MOSFET gate charge as available in device data sheet vs. switching event Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 14 The slope time tSLOPE can be calculated as follows: Whereas QMILLER is the charge the power transistor needs for the switching event, and I GATE is the driver current setting of the TMC603. Taking into account, that a slow switching event means high power dissipation during switching, and, on the other side a fast switching event can cause EMV problems, the desired slope will be in some ratio to the switching (chopper) frequency of the system. The chopper frequency is typically slightly outside the audible range, i.e. 18 kHz to 40 kHz. The lower limit for the slope is dictated by the reverse recovery time of the MOSFET internal diodes, unless additional Schottky diodes are used in parallel to the MOSFETs source-drain diode. Thus, for most applications a switching time between 100ns and 750ns is chosen. The required slope control resistor RSLP can be calculated as follows: Example: A circuit using the transistor from the diagram above shall be designed for a slope time of 200ns. The miller charge of the transistor is about 6nC. The nearest available resistor value is 330 k . It sets the gate driver current to roughly 30mA. This is well within the minimum and maximum RSLP resistor limits. 5.2.5 Reverse capacity (QGD) protection The principle of slope control often is realized by gate series resistors with competitor’s products, but, as latest MOSFET generations have a fairly high gate-drain charge (QGD), this approach is critical for safe bridge operation. If the gate is not held in the off state with a low resistance, a sudden raise of the voltage at the drain (e.g. when switching on the complementary transistor) could cause the gate to be pulled high via the MOSFETs gate drain capacitance. This would switch on the transistor and lead to a bridge short circuit. The TMC603 provides for safe and reliable slope controlled operation by switching on a low resistance gate protection transistor (see figure). Vgate Ion on QGD D Slope controlled G S off QGS Ioff full, safe off TMC603 QGD protected driver stage External MOSFET figure 9: QGD protected driver stage Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 15 5.2.6 Considerations for QGD protection This chapter gives the background understanding to ensure a safe operation for MOSFETs with a gate-drain (Miller) charge QGD substantially larger than the gate-source charge QGS. In order to guarantee a safe operation of the Q GD protection, it is important to spend a few thoughts on the slope control setting. Please check your transistors’ data sheet for the gate-source charge QGS and the gate-drain charge QGD (Miller charge). In order to turn on the MOSFET, first the gate-source charge needs to be charged to the transistor’s gate. Now, the transistor conducts and switching starts. During the switching event, the additional QGD needs to be charged to the gate in order to complete the switching event. Wherever QGD is larger than QGS, a switching event of the complementary MOSFET may force the gate of the switched off MOSFET to a voltage above the gate threshold voltage. For these MOSFETs the QGD protection ensures a reliable operation, as long as the slopes are not set too fast. Calculating the maximum slope setting for high QGD MOSFETs: Taking into account effects of the MOSFET bulk diode (compare chapter 5.2.7), the maximum slope of a MOSFET bridge will be around the double slope as calculated from the Miller charge and the gate current. Based on this, we can estimate the current required to hold the MOSFET safely switched off: During the bridge switching period, the driver must be able to discharge the difference of Q GD and QGS while maintaining a gate voltage below the threshold voltage. Therefore Thus the minimum value required for IOFFQGD can be calculated: Where ION is the gate current set via RSLP, and IOFFQGD is the QGD protection gate current. The low side driver has a lower QGD protection current capability than the high side driver, thus we need to check the low side. With its RLSOFFQGD of roughly 15 Ohm, the TMC603 can keep the gate voltage to a level of: Now we just need to check UGOFF against the MOSFETs output characteristics, to make sure, that no significant amount of drain current can flow. Example: A MOSFET, where QGD is 3 times larger than QGS is driven with 100mA gate current. The TMC603 thus can keep the gate voltage level to a maximum voltage of UGOFF = 133mA * 15Ω = 2V This is sufficient to keep the MOSFET safely off. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 16 Note: Do not add gate series resistors to your MOSFETs! This would eliminate the effect of the Q GD protection. Gate series resistors of a few Ohms only may make sense, when paralleling multiple MOSFETs in order to avoid parasitic oscillations due to interconnection inductivities. 5.2.7 Effects of the MOSFET bulk diode Whenever inductive loads are driven, the inductivity will try to sustain current when current becomes switched off. During bridge switching events, it is important to ensure break-before-make operation, e.g. one MOSFET becomes switches off, before the opposite MOSFET is switched on. Depending on the actual direction of the current, this results in a short moment of a few 100 nanoseconds, where the current flowing through the inductive load forces the bridge output below the lower supply rail or above the upper supply rail. The respective MOSFET bulk diode in this case takes over the current. The diode saturates at about -1.2V. But the bulk diode is not an optimum device. It typically has reverse recovery time of a few ten to several 100ns and a reverse recovery charge in the range of some 100nC or more. Assuming, that the bulk diode of the switching off MOSFET takes over the current, the complementary MOSFET sees the sum of the coil current and the instantaneous current needed to recover the bulk diode when trying to switch on. The reverse recovery current may even be higher than the coil current itself! As a result, a number of very quick oscillations on the output appear, whenever the bulk diode leaves the reverse recovery area, because up to the half load current becomes switched off in a short moment. The effect becomes visible as an oscillation due to the parasitic inductivities of the PCB traces and interconnections. While this is normal, it adds high current spikes, some amount of dynamic power dissipation and high frequency electromagnetic emission. Due to its high frequency, the ringing of this current can also be seen on the gate drives and thus can be easily mistaken as a gate driving problem. In order to reduce overshoot and ringing, a snubber element can be used, e.g. a capacitor with some nano Farad in series with a resistor in the range some 100mΩ on each motor output. VVM HS takes over output current UBMX IHS ILSBULK 0V -1.2V IOUT 0A 0A -IOUT Phase of switching event LS bulk diode conducting IOUT HS curr. rise up to IOUT LS bulk reverse recovery overshoot + ringing normal slope switching complete HS starts conducting figure 10: effect of bulk diode recovery A further conclusion from this discussion: Do not set the bridge slope time higher than or near to the reverse recovery time of the MOSFETs, as the parasitic current spikes will multiply the instantaneous current across the bridge. A plausible time is a factor of three or more for the slope time. If this cannot be tolerated please see the discussion on adding Schottky diodes. 5.2.8 Adding Schottky diodes across the MOSFET bulk diodes In order to avoid effects of bulk diode reverse recovery, choose a fast recovery switching MOSFET. The MOSFET transistors can also be bridged by a Schottky diode, which has a substantially faster reverse recovery time. This Schottky diode needs to be chosen in a way that it can take over the full bridge current for a short moment of time only. During this time, the forward voltage needs to be lower than the MOSFETs forward voltage. A small 5A diode like the SK56 can take over a current of 20A at a forward voltage of roughly 0.8V. Even in this constellation, an optional snubber element at the output can reduce overshoot and ringing (see schematic). Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 17 +VM HS1 Z 12V BM1 Motor 220R LS1 10nF GNDP 1R optional snubber (example values) figure 11: parallel Schottky diode avoids current spikes due to bulk diode recovery, optional snubber reduces overshoot and ringing 5.2.9 Short to GND detection An overload condition of the high side MOSFET (“short to GND”) is detected by the TMC603, by monitoring the BM voltage during high side on time. Under normal conditions, the high side power MOSFET reaches the bridge supply voltage minus a small voltage drop during on time. If the bridge is overloaded, the voltage cannot rise to the detection level within a limited time, defined by an external resistor. Upon detection of an error, the error output is activated. By directly tying it to the enable input, the chip becomes disabled upon detection of a short condition and the error flip flop becomes set. A variation of the short to GND detection delay allows adaptation to the slope control, as well as modification of the sensitivity of the short to GND detection. BHx 0V VVM VVM- BMx Short detection 0V /ERROUT, ENABLE Driver off via ENABLE pin 0V tS2G Short to GND monitor phase Short to GND detected Valid area VBMS2G inactive tS2G delay BMx voltage monitored inactive delay Short detected figure 12: timing of the short to GND detector Pin Comments RS2G The resistor connected to this pin controls the delay between switching on the high side MOSFET and the short to GND check. A 20µA current out of this pin (resistor value of 220 k to GND) results in a 500ns delay, a lower current gives a longer delay. Disconnecting the pin disables the function. Keep interconnection between IC and resistor short, to avoid stray capacitance to adjacent signal traces of modulating the set current. Resistor range: 47 k to 1 M 5.2.10 Error logic The TMC603 has three different sources for signaling an error: Undervoltage of the low side supply Undervoltage of the charge pump Short to GND detector Upon any of these events the error output is pulled low. After a short to GND detector event, the error output remains active, until it becomes cleared by the CLR_ERR. By tying the error output to the Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 18 enable input, the TMC603 automatically switches off the bridges upon an error. The enable input then should be driven via an open collector input plus pull-up resistor, or via a resistor. Pull-up resistor can be internal to microcontroller +VCC TMC603 error logic Drive with open drain output, if feedback is provided ENABLE CLR_ERR undervoltage VLS undervoltage VCP 100k S: priority D short to GND 1 short to GND 2 short to GND 3 D /ERR_OUT S Q R Q D GND Feedback connection for automatic self-protection figure 13: error logic Pin Comments /ERR_OUT Error output (open drain). Signals undervoltage or overcurrent. Tie to ENABLE for direct self protection of the driver. The internal error condition generator has a higher priority than the CLR_ERR input, i.e. the error condition cannot be cleared, as long as it is persistent. CLR_ERR Reset of error flip-flop (active high). Clears error condition. The error condition should at least be cleared once after IC power on. ENABLE Enables the power drivers (low: all MOSFETs become actively switched off) 5.2.11 Paralleling gate drivers for higher gate current In order to double gate driver current in a BLDC application, two TMC603 can be switched in parallel to have the double gate driver current while taking advantage of all features. Therefore it is important to parallel the gate driver inputs and outputs of the second IC to the first IC, and to also parallel the ERR_OUT and ENABLE input. The driver strength of both ICs adds taking into account their respective slope control resistor. The switching regulator and charge pump of one device can supply both ICs! Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 5.3 19 Current measurement amplifiers The TMC603 amplifies the voltage drop in the three lower MOSFET transistors in order to allow current measurement without the requirement for current sense (shunt) resistors. This saves cost and board space, as well as the additional power dissipation in the shunt resistors. Optional shunt resistors can be used, e.g. source resistors for each lower MOSFET or a common shunt resistor in the bridge foot point if a more precise measurement without the need for calibration and temperature compensation is desired. For the TMC603A, the FILTx pins in this mode are switched as inputs for the sensing of the shunt resistors. The internal amplifier conditions the signal for a standard microcontroller. The TMC603 CURx outputs deliver a signal centered to 1/3 of the 5V VCC supply. This allows measurement of both, negative and positive signals, while staying compatible to a 3.3V microcontroller. The current amplifier is an inverting type. +VCC R SENSE_HI ENRS_TEST amplify 5x or 20x D R 1/3 VCC D SWC BMx FILTx_RSx R A A track & hold stage A autozero BLx SAMPLEx automatic sample point delay D CURx add 1/3 VCC offset D figure 14: schematic of current measurement amplifiers Pin Comments CURx Output of current measurement amplifier. The output signal is centered to 1/3 VCC. SENSE_HI Switches current amplifiers to high sensitivity (high level). Control by processor to get best sensitivity and resolution for measurement. SAMPLEx Optional external control for current measurement sample/hold stage. Set to positive level, if unused FILTx_RSx Input for optional external sense resistor. To enable, tie pin ENRS_TEST to VCC. This feature has been added in TMC603A. The voltage drop over the MOSFET (or shunt resistor) is calculated as follows: whereas x is the ADC output value, x0 is the ADC output value at zero current (e.g. 85 for an 8 bit ADC with 5V reference voltage), ADCMAX is the range of the ADC (e.g. 256 for an 8 bit ADC), VADCREF is the reference voltage of the ADC and ACUR is the currently selected amplification (absolute value) of the TMC603. With this, the motor current can be calculated using the on resistance RDSON (at 10V) of the MOSFET: Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 20 For a shunt resistor based measurement, the same formula is true: For the shunt resistor measurement, care has to be taken not to exceed the voltage range which can be accepted by the measurement input, i.e. the shunt resistor should be selected in a way that the voltage drop is at maximum 0.3V at full motor current. This is the maximum voltage which can be measured. A lower sense resistor gives less power dissipation, but lower currents show with less resolution. 5.3.1 Current measurement timing Current measurement is self-timed, in order to only provide valid output voltages. Sampling is active during the low side ON time. The sampling is delayed by an internal time delay, in order to avoid sampling of instable values during settling time of the bridge current and amplifiers. Thus, a minimum ON time is required in order to get a current measurement. The output CURx reflects the current during the measurement. The last value is held in a track and hold circuit as soon as the low side transistor switches off. External control Current sense out Bridge voltage drop Control signals Internal sample control SAMPLEx 0V BLx 0V BMx VVM 0.25V 0V -0.25V CURx VVCC/3 0V tBLHICURX Phase tBLHICURX Hold (undef.) CURx tracking -BMx Hold CURx tracking -BMx Hold figure 15: timing of the current measurement The SAMPLEx pins can be used to refresh the measurement during long on time periods, e.g. when the motor is in standstill, with the low side being continuously on, e.g. in a hall sensor based block commutation scheme with the chopper on the high side. In this application, all SAMPLEx pins can be tied together to one microprocessor output. For advanced applications, where a precise setting of the current sampling point is desired, e.g. centered to the on-time, SAMPLEx pins can be deactivated at the desired point of time, enabling the hold stage. 5.3.2 Auto zero cycle The current measurement amplifiers do an automatic zero cycle during the OFF time of the low side MOSFETs. The zero offset is stored in internal capacitors. This requires switching off the low side at least once, before the first measurement is possible, and on a cyclic basis, to avoid drifting away of the zero reference. This normally is satisfied by the chopper cycle. If commutation becomes stopped, e.g. due to motor stand still, the respective phase current measurement could drift away. After the first switching off and on of the low side, the measurement becomes valid again. Therefore, you should integrate a timer in your commutation, which checks for the low side on time exceeding for example 10ms. If the on time of the respective low side reaches this time limit, you can either use the sample input SAMPLEx to refresh the current measurement, by switching it high for at least 1µs, or you switch off the low side for a short time of a few microseconds. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 21 5.3.3 Measurement depending on chopper cycle If the low side on-time on one phase tBLHICURX is too short, a current measurement is not possible. The TMC603 automatically does not sample the current if the minimum low side-on time requirement is not met. This condition can arise in normal operation, e.g. due to the commutation angle defined by a sine commutation chopper scheme. The respective CURx output then does not reflect the phase current. Thus, the CURx output of a phase should be ignored, if the on-time falls below the minimum low side on-time for current measurement (please refer to maximum limit). The correct current value can easily be calculated using the difference of the remaining two current measurements. This results from the fact that the sum of all three currents equals zero (I U+IV+IW = 0). This way, all motor currents are always known from the measurement of two phase currents. It is important to know all three phase currents for a sine commutated motor. For block commutation, there is always one low side active and the full current can be seen at this low side. 5.3.4 Compensating for offset voltages In order to measure low current values precisely, the “zero” value (x0) of 1/3 VCC should be measured via the ADC, rather than being hard coded into the measurement software. This is possible by doing a first current measurement during motor stand-still, with no current flowing in the motor coils, e.g. during a test phase of the unit. The resulting value can be stored and used as zero reference. However, the influence of offset voltages can be minimized, by using the high sensitivity setting of the amplifiers for low currents, and switching to low sensitivity for higher currents. 5.3.5 Getting a precise current value using MOSFET on-resistance The on-resistance of a MOSFET has a temperature co-efficient, which should not be ignored. Thus, the temperature of the MOSFETs must be measured, e.g. using an NTC resistor, in order to compensate for the variation. Also, the initial R DSON depends upon fabrication tolerance of the MOSFETs. If exact measurement is desired, an adjustment should be done during initial testing of each product. For applications, where an adjustment is not possible, external sense resistors can be used instead. A single resistor in the GND line often is sufficient for block commutation. For sine commutation, three sense resistors should be used. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 5.4 22 hallFX™ sensorless commutation hallFX™ provides emulated hall sensor signals. The emulated hall sensor signals are available without a phase shift and there is no error-prone PLL necessary, like with many other systems, nor is the knowledge of special motor parameters required. Since it is based on the motors’ back-EMF, a minimum motor velocity is required to get a valid signal. Therefore, the motor needs to be started without feedback, until the velocity is high enough to generate a reliable hallFX™ signal. Position signal generation (PSG) Switch Cap filters Induction pulse supressor (IPS) U Low pass LPU ULP BM1 E1 D H1 W 30k Low pass LPV VLP BM2 E1 E2 PSG IPS D H2 E2 30k E3 Low pass LPW WLP BM3 V H1 E3 D H3 H2 H3 30k ENRS_TEST D A SP_SUP A FILT3 FILT2 FILT1 SC_CLK A A CSUP figure 16: hallFX™ block diagram and timing A switched capacitor filter for each coil supplies the measured effective coil voltages. Its filter frequency can be adapted to the chopper frequency and the desired maximum motor velocity. An induction pulse suppressor unit gates the commutation spikes which result from the inductive behavior of the motor coils after switching off the current. The gating time can be adapted by an external capacitor to fit the motor inductivity and its (maximum) velocity. Pin Comments SP_SUP A capacitor attached to this pin sets the spike suppression time. This pin charges the capacitor via an internal current source. If more exact timing is required, an external 47k pull-up resistor to VCC can be used in parallel to the internal current source. The capacitor becomes discharged upon each valid commutation. The capacitor can optionally be left away, and the suppression can be done in software. FILTx_RSx These pins provide the filtered coil voltages when ENRS_TEST is tied to GND. For most applications this will be of no use, except when an external back-EMF commutation is realized, e.g. using a microcontroller with ADC inputs. Because of the high output resistance and low current capability of these pins, it is advised to add an external capacitor of a few hundred picofarad up to a few nanofarad to GND, if the signals are to be used. This prevents noise caused by capacitance to adjacent signal traces to disturb the signal. Hx Emulated hall sensor output signal of hallFX™ block. SCCLK An external clock controls the corner frequency of the switched capacitor filter. A 1.25 MHz clock gives a filter bandwidth of 3kHz. A lower clock frequency may be better for lower motor velocities. 5.4.1 Adjusting the hallFX™ spike suppression time hallFX™ needs two minimum motor- and application-specific adjustments: The switched capacitor clock frequency and the spike suppression time should be adapted. Both can easily be deducted from basic application parameters and are not very critical. The SCCLK frequency should be matched to the chopper frequency of the system and the maximum motor velocity. The spike suppression time needs to be adapted to the desired maximum motor velocity. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 23 Calculating the commutation frequency fCOM of the motor: SRPM is the rotation velocity in RPM nPOLE is the pole count of the actual motor, or the double of the number of pole pairs The spike suppression time can be chosen as high, as the commutation frequency required for maximum motor velocity allows. As a thumb rule, we take half of this time to have enough spare. Example: Given a 4 pole motor operating at 4000 RPM: CSUP = 6.25nF. The nearest value is 6.8nF. 5.4.2 Adjusting the hallFX™ filter frequency The filter block needs to separate the motors’ back EMF from the chopper pulses. Thus, the target is, to filter away as much commutation noise as possible, while maintaining as much of the back EMF signal as possible. Therefore, we need to find a cut-off frequency in between the chopper frequency and the electrical frequency of the motor. Since we do not want to change the frequency within the application, we use the nominal or maximum motor velocity to calculate its electrical frequency. The chopper frequency is given by the system, typically about 20 kHz. The electrical frequency of the motor is: Since the filter has a logarithmic behavior, as a thumb rule we can make a logarithmic mean-value as follows: With the cut-off frequency being about 1/390 of the switched capacitor clock frequency f SCCLK the following results as a thumb rule: The result shall be checked against minimum limit of 250 kHz and maximum limit of 4 MHz, however, the actual frequency is quite uncritical and can be varied in a wide range. Example: Given a 4 pole motor operating at 4000 RPM with a 20 kHz chopper frequency: fEL = 133 Hz fCUTOFF = 1.6 kHz fSCCLK = 0.64 MHz The result is well within the limits, however, the frequency in a practical application can be chosen between 300 kHz and 1.5 MHz. 5.4.3 Block commutation chopper scheme for hallFX™ hallFX™ works perfectly with nearly every motor. You can use a standard block commutation scheme, but the chopper must fulfill the following: The coils must be open for some percentage of the chopper period, in order to allow the back-EMF of the motor to influence the coil voltages. The motor direction Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 24 is determined by the start-up scheme, since the hallFX™ signals depend on the direction. Thus, the same commutation scheme is used for turn right and turn left! Only a single commutation table is required. You find the required commutation table in chapter 8.3. Bridge control signals (high active) HallFX signals Motor turning forward H1 0V H2 0V H3 0V BH1 0V BL1 0V BH2 0V BL2 0V BH3 0V BL3 0V Hall vector 3 Chopper on high side 1 5 4 Motor turning reverse 6 2 (chopper events not shown) Chopper on low side 5 1 3 2 6 4 Example: 50% chopper on high and low side showing 3 chopper events figure 17: hallFX™ based commutation A chopper scheme fulfilling the desired coil open time per chopper period is shown here: Both, the high side driver and the low side driver are chopped with the same signal. The coil open time automatically is inverted to the duty cycle. In a practical application, the motor can run with a duty cycle of 15% to 25% (minimum motor velocity at low load) up to 90% to 95% (maximum motor velocity). The exact values depend on the actual motor. With a lower duty cycle the motor would not start, or back EMF would be too small to yield a valid hallFX™ signal. With a higher duty cycle, the back EMF would not be visible at the coil voltages, because the coils would be connected to GND or VM nearly the whole time. The minimum resulting coil open time thus is 5% to 10%. This simple chopper scheme automatically gives a longer measuring time at low velocities, when back EMF is lower. The actual borders for the commutation should be checked in the actual application. Provide enough headroom to compensate for variations due to motor load, mechanics and production stray. 5.4.4 Start-up sequence for the motor with forced commutation In order to start the motor running with hallFX™, it must reach a minimum velocity. The microcontroller needs to take care of this by starting the motor in a forward control mode, without feedback – just like a stepper motor. In order to allow a smooth transition to feedback mode, the same chopper scheme should be used as described above. Alternatively, the chopper scheme can be changed a few electrical periods before you switch to hallFX™. This allows for example to start-up the motor using a sine commutation, to get a smooth movement also at low motor velocities. In a practical application, only a few percent up to 10% of the maximum motor velocity are sufficient for hallFX™ operation. Turn the motor up to a minimum velocity, where you safely get correct hallFX™ signals. Since rotation of the motor cannot be measured during this phase, the motor needs to be current controlled, with a current which in every case is high enough to turn the mechanical load. Current control can be done by feedback control, or by adapting the duty cycle to the motor characteristics. Further, the minimum starting speed and acceleration needs to be set matching the application. For sample code, please see www.trinamic.com. Upon reaching the threshold for hallFX™ operation, a valid hall signal Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 25 becomes available and allows checking success of the starting phase. The turning direction of the start-up sequence automatically determines the direction of motor operation with hallFX™. You can check velocity and direction, as soon as valid hallFX™ signals are available. When you experience commutation sequence errors during motor operation, probably motor velocity has dropped below the lower threshold. In this case, the motor could be restarted in forward control mode, or you could switch to forward control mode on the fly. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 5.5 26 Power supply The TMC603 integrates a +12V switching regulator for the gate driver supply and a +5V linear regulator for supply of the low voltage circuitry. The switching regulator is designed in a way, that it provides the charge pump voltage by using a Villard voltage doubler circuit. It is able to provide enough current to supply a number of digital circuits by adding an additional 3.3V or 5V low voltage linear or switching regulator. If a +5V microcontroller with low current requirement is used, the +5V regulator is sufficient, to also supply the microcontroller. +VM VM+10V charge pump BAV99 (70V) BAS40-04W (40V) 220n 220n 100n (2x) 4µ7 Tantal 25V 16V LSW SS16 TP0610K or BSS84 (opt. BC857) 12V supply 100µ (150mA with sel. transistor) 100n (2x) optional supply filter components when supply ripple is high due to low filter capacity for transistor bridges SMD induct. 1µH or 4R7 1µ LSW: 220µH for 100kHz VM TMC603 voltage regulators SWOUT startup current VLS VCP 5V linear regulator VM-12V / 2mA driver triangle OSC COSC COSC: 470p ->100kHz 14k +VCC 5VOUT VCC 5V supply 100nF +VCC Q S Q R R R1 5/12 VLS R 150mV triangle 1/5R 4/5R R R2 10R dutycycle limit GND figure 18: power supply block with example values Pin Comments COSC Oscillator capacitor for step down regulator. A 470pF capacity gives 100kHz operation. Do not leave this pin unconnected. Tie to GND, if oscillator is not used. SWOUT Switch regulator transistor output. The output allows driving of a small signal Pchannel MOSFETs as well as PNP small signal transistors 5VOUT Output of internal 5V linear regulator. Provided for VCC supply 5.5.1 Switching regulator The switching regulator has been designed for high stability. It provides an upper duty cycle limit, in order to ensure switching operation even at low supply voltage. This allows the combination with a Villard voltage doubler. The application schematic shows a number of standard values, however, the coil and oscillator frequency can be altered: The choice of the external switching regulator transistor depends on the desired load current and the supply voltage. Especially for high switching frequencies, a low gate charge MOSFET is required. The following table shows an overview of available transistors and indicative operation limits. For a higher output current, two transistors can be used in parallel. In this case the switching frequency should be halved, because of the higher gate charge leading to slower switching slopes. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 27 transistor type manufacturer gate charge (typ.) max. frequency max. voltage max. load current BC857 div. - (bipolar) 100 kHz 40V 80 mA BSS84 Fairchild, NXP 0.9 nC 300 kHz 50V 120 mA TP0610K Vishay 1.3 nC 230 kHz 60V 150 mA NDS0605 Fairchild 1.8 nC 175 kHz 60V 150 mA TP0202K Vishay 1 nC 300 kHz 30V 350 mA For the catching diode, use a Schottky type with sufficient voltage and current rating. The choice of a high switching frequency allows the use of a smaller and less expensive inductor as well as a lower capacitance for the Villard circuit and the switching regulator output capacitor. However, the combination of inductor, transistor and switching frequency should be carefully selected and should be adapted to the load current, especially if a high load current is desired. Choice of capacitor for the switching frequency (examples): COSC frequency fOSC inductivity example 470 pF 100 kHz 220 µH 220 pF 175 kHz 150 µH 100 pF 300 kHz 100 µH Remark Not recommended for VVM < 14V The switcher inductivity shall be chosen in a way, that it can sustain part of the load current between each two switching events. If the inductivity is too low, the current will drop to zero and higher frequency oscillations for the last part of each cycle will result (discontinuous mode). The required transistor peak current will rise and thus efficiency falls. For a low load current, operation in discontinuous mode is possible. If a high output current is required, a good design value for continuous mode is to target a current ripple in the coil of +/-40%. To give a coarse hint on the required inductor you can use the following formula for calculating the minimum inductivity required for continuous operation, based on a ripple current which is 100% of the load current: VVM is the supply voltage. For low voltage operation (15V or less), the output voltage sinks from 12V to 0.85*VVM. The formula can be adapted accordingly. IOUT is the current draw at 12V. For 40% current ripple, you can use roughly the double inductivity. If ripple is not critical, you can use a much smaller inductivity, e.g. only 5% to 50% of the calculated value. But at the same time switching losses will rise and efficiency and current capability sink due to higher losses in the switching transistor. If the TMC603 does not supply additional external circuitry, current draw is very low, about 20mA in normal operation. This would lead to large inductivity values. In this case we recommend going for the values given in the table above in order to optimize coil cost. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 28 Example: fOSC = 175 kHz, IOUT = 0.2 A, VVM = 48 V: For continuous operation, a 330µH or 470µH coil would be required. The minimum inductivity would be around 100µH. Note: Use an inductor, which has a sufficient nominal current rating. Keep switching regulator wiring away from sensitive signals. When using an open core inductor, please pay special care to not disturbing sensitive signals. 5.5.2 Charge pump The Villard voltage doubler circuit relies on the switching regulator generating a square wave at the switching transistor output with a height corresponding to the supply voltage. In order to work properly the load drawn at +12V needs to be higher than the load drawn at the charge pump voltage. This normally is satisfied when the IC is supplied by the step down regulator. For low voltage operation, the charge pump voltage needs to be as high as possible to guarantee a high gate drive voltage, thus, a dual Schottky diode should be used for the charge pump in low voltage applications. 5.5.3 Filter capacitors for switching regulator and charge pump The filter capacitors in the switching regulator and the charge pump are required to provide current for the high current spikes which are caused by switching up to three MOSFETs at the same time. The required amount of charge can be estimated when looking at the MOSFETs gate charge. The gate voltage should not drop significantly due to the switching event, e.g. only 100mV. Additionally, the 12V filter capacitor provides charge for load spikes on the 12V net and filter switching ripple. In applications, where board space is critical, lower capacitance values can be used. Choice of filter capacitors in the switching regulator for different current requirements (example): 12V load current power MOSFET 12V filter capacitor gate charge (electrolytic/ceramic) charge pump filter capacitor (tantalum / ceramic) <20mA <20nC 22µF (or 4.7µF ceramic) 1µF (e.g. ceramic) <20mA <50nC 22µF (or 10µF ceramic) 2.2µF (e.g. ceramic) <50mA >50nC 47µF (or 10µF ceramic) 4.7µF 100mA >50nC 100µF (or 10µF ceramic) 4.7µF 5.5.4 Supply voltage filtering and layout considerations As with most integrated circuits, ripple on the supply voltage should be minimized in order to guarantee a stable operation and to avoid feedback oscillations via the supply voltages. Therefore, use a ceramic capacitor of 100nF per supply voltage pin (VM to GND, VLS to GND and VCC to GND and VCP to VM). Please pay attention to also keep voltage ripple on VCC pin low, especially when the 5V output is used to supply additional external circuitry. It also is important to make sure, that the resistance of the power supply is low when compared to the load circuit. Especially high frequency voltage ripple >1MHz should be suppressed using filter capacitors near the power bridge or near the board power supply. The VM terminal is used, to detect short to GND situations, thus, it has to correspond to the bridge power supply. In high noise applications, it may make sense to filter VCP supply separately against ripple to GND. A large low ESR electrolytic capacitor across the bridge supply (VM to GND) should also be used, because it effectively suppresses high frequency ripple. This cannot be accomplished with ceramic capacitors. GND and GNDP pins should be tied to a common, massive GND plane. Pay attention to power routing: Use short and wide, straight traces. The PCB power supply should be placed near the driver bridge, where most current is consumed, to avoid current drop in the plane between critical components like TMC603 and microcontroller. This is especially is important to get a precise current measurement. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 29 5.5.5 Reverse polarity protection Some applications need to be protected against a reversed biased power supply, i.e. for automotive applications. A highly efficient reverse polarity protection based on an N channel MOSFET can simply be added due to the availability of a charge pump voltage. This type of reverse polarity protection allows feeding back energy into the supply, and thus is preferable to a pure diode reverse polarity protection. +Terminal -Terminal BC846 10k Reverse polarity power MOS (i.e. same type as bridge transistors) 10k +VM protected (to bridge) VM VCP figure 19: adding a reverse polarity protection 5.5.6 Standby with zero power consumption In battery powered applications, a standby function often is desired. It allows switching the unit on or off without the need for a mechanical high power switch. In principle, the bridge driver MOSFETs can switch off the motor completely, but the TMC603 and its switching regulator still need to be switched off in order to reduce current consumption to zero. Only a low energy standby power supply will remain on, in order to wake up the system controller. This standby power supply can be generated by a low current zener diode plus a resistor to the battery voltage, buffered by a capacitor. The example in the schematic uses a P channel MOSFET to switch off power for the TMC603 and any additional ICs which are directly supplied by the battery. Before entering standby mode, the motor shall be stopped and the TMC603 should be disabled. +Vbattery +VM to bridge, only 100K 220n FDC5614P POWER SWITCH electronic ON switch +VM switched, 3A max. 27k VM 10µ TMC603 enable ENABLE HSx (only shown for one high side MOSFET) figure 20: low power standby 5.5.7 Low voltage operation down to 9V In low voltage operation, it is important to keep the gate driving voltages as high as possible. The switching regulator for VLS thus is not needed and can be left out. Tie the pin COSC to GND. VLS becomes directly tied to +VM, which is possible as long as the supply voltage does not exceed 14V (16V peak). However, now a source for the Villard voltage doubler is missing. A simple solution is to use a CMOS 555 timer circuit (e.g. TLC555) oscillating at 250 kHz (square wave) to drive the voltage doubler. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) +9V...14V 30 VM+10V charge pump 100n BAS40-04W 470n 16V 12V supply VCC DISCH 100n (2x) 100n (2x) RESET OUT TLC555 THRES CONT (150mA with sel. transistor) optional supply filter components when supply ripple is high due to low filter capacity for transistor bridges SMD induct. 1µH or 4R7 22k 1µ TRIG GND 150p VM SWOUT VLS VCP 5VOUT COSC 5V supply VCC 100nF TMC603 figure 21: low voltage operation 5.6 Test output The test output is reserved for manufacturing test. It is used as an input for a normal application. Tie to GND or VCC in application. Pin Comments ENRS_TEST Enable sense resistor input and output for test voltages. Output resistance 25kOhm +-30%. Reset: ENABLE(=low); Clock: SCCLK (rising edge). Test voltage sequence: 0: 0V 1..3 / 4..6 / 7..9: Gate_HS_Off, Gate_LS_On, Gate_LS_Off (driver 1/2/3) 10..14: currently unused 15: 0V (no further counts: Reset for restart) 5.7 ESD sensitive device The TMC603 is an ESD sensitive CMOS device and also MOSFET transistors used in the application schematic are very sensitive to electrostatic discharge. Take special care to use adequate grounding of personnel and machines in manual handling. After soldering the devices to the board, ESD requirements are more relaxed. Failure to do so can result in defect or decreased reliability. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 31 6 Absolute Maximum Ratings The maximum ratings may not be exceeded under any circumstances. Operating the circuit at or near more than one maximum rating at a time for extended periods shall be avoided by application design. Parameter Supply voltage Supply and bridge voltage max. 20000s Symbol VVM Min Max Unit -0.5 50 V 55 V Low side driver supply voltage VVLS -0.5 14 V Low side driver supply voltage max. 20000s VVLS -0.5 16 V -0.5 60 V 65 V VM-10 VM+16 V VVCC -0.5 6.0 V Logic input voltage VI -0.5 VCC+0.5 V Analog input voltage VIA -0.5 VCC+0.5 V Voltages on driver pins (HSx, LSx, BMx) VDRVIO -0.7 0.7 V Relative high side driver voltage (VHSX – VBMX) VHSBM -20 20 V Charge pump voltage (related to GND) Charge pump voltage max. 20000s VVCP Charge pump voltage during power up / down Logic supply voltage Maximum current to / from digital pins and analog low voltage I/Os IIO +/-10 mA 5V regulator continuous output current I5VOUT 40 mA 5V regulator short time output current I5VOUT 150 mA Junction temperature TJ -50 150 °C Storage temperature TSTG -55 150 °C ESD-Protection (Human body model, HBM), in application VESDAP 1 kV ESD-Protection (Human body model, HBM), device handling VESDDH 100 V 7 Electrical Characteristics 7.1 Operational Range Parameter Symbol Min Max Unit Ambient temperature TA -40 125 °C Junction temperature TJ -40 140 °C 10 50 V 9 14 V Supply voltage (standard circuit) Supply voltage (low voltage application: VVLS=VVM) VVM Low side driver supply voltage VVLS 9 13 V Differential charge pump voltage measured to VM (VVCP – VVM) VCPD 8 12 V Logic supply voltage VVCC 4.75 5.25 V Slope control resistor with VVM <30V RSLP 60 500 k Slope control resistor with VVM >30V RSLP 100 500 k Short to GND control resistor RS2G 47 1000 k Output slope tSLP 100 1000 ns Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 7.2 32 DC Characteristics and Timing Characteristics DC characteristics contain the spread of values guaranteed within the specified supply voltage range unless otherwise specified. Typical values represent the average value of all parts measured at +25°C. Temperature variation also causes stray to some values. A device with typical values will not leave Min/Max range within the full temperature range. NMOS low side driver note 1) Parameter DC-Characteristics VVCC = 5.0 V, VVLS = 12V Symbol Conditions Min Typ Max Unit Gate drive current LSx low side switch ON ILSON VLSX = 5V RSLP = 68k 150 mA Gate drive current LSx low side switch OFF ILSOFF VLSX = 5V RSLP = 68k -150 mA Gate drive current LSx low side switch ON ILSON VLSX = 5V RSLP = 100k 75 100 125 mA Gate drive current LSx low side switch OFF ILSOFF VLSX = 5V RSLP = 100k -75 -100 -125 mA Gate drive current LSx low side switch ON ILSON VLSX = 5V RSLP = 220k 50 mA Gate drive current LSx low side switch OFF ILSOFF VLSX = 5V RSLP = 220k -50 mA Gate Off detector threshold VGOD VLSX falling 1 V QGD protection resistance after detection of gate off RLSOFFQGD VLSX = 2V 15 Delay LS driver switch on BLx to LSx at 50% tLSON RSLP = 100k CLSX = 100pF 35 70 140 ns Delay LS driver switch off BLx to LSx at 50% tLSOFF RSLP = 100k CLSX = 100pF 80 160 320 ns Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) NMOS high side driver note 1) Parameter 33 DC-Characteristics VVCC = 5.0 V, VVLS = 12V, VCPD = 10.5V Symbol Conditions Min Typ Max Unit Gate drive current HSx high side switch ON IHSON VHSX = 5V RSLP = 68k 150 mA Gate drive current HSx high side switch OFF IHSOFF VHSX = VM+5V RSLP = 68k -150 mA Gate drive current HSx high side switch ON IHSON VHSX = 5V RSLP = 100k 75 100 150 mA Gate drive current HSx high side switch OFF IHSOFF VHSX = VM+5V RSLP = 100k -75 -100 -125 mA Gate drive current HSx high side switch ON IHSON VHSX = 5V RSLP = 220k 50 mA Gate drive current HSx high side switch OFF IHSOFF VHSX = VM+5V RSLP = 220k -50 mA Gate Off detector threshold high side VHSX-VBMX, BM level high VGOD VHSX falling VBMX > VGOBM 0 V Gate Off detector threshold high side VBMX, BM level low VGOBM VBMX falling 3.5 V VBMX = 24V VHSX = VBMX+2V 300 mA QGD protection current after detection of gate off IHSOFFQGD Delay HS driver switch on BHx to HSx at 50% tHSON RSLP = 100k VM = 24V CHSX = 100pF 75 150 300 ns Delay HS driver switch off BHx to HSx at 50% tHSOFF RSLP = 100k VM = 24V CHSX = 100pF 60 120 240 ns Min Typ Max Unit Break-before-make block note 1) Parameter Timing-Characteristics VVM = 48 V, RSLP = 100K Symbol Conditions Break-before-make delay LSx off to HSx on tBBMLH Measured at 1V gate-source voltage 160 ns Break-before-make delay HSx off to LSx on tBBMHL Measured at 1V gate-source voltage 290 ns 1) See timing diagram in figure 6: bridge driver timing Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) RSLP input and RS2G input Parameter Typical voltage at RSLP and RS2G input, depending on the external resistor 34 DC-Characteristics VVCC = 5.0 V Symbol Conditions VRSLP VRS2G RSLP = 100 k Min Typ Max 3.8 Unit V RS2G = 100 k 4.0 RSLP = 470 k RS2G = 470 k Short to GND detector Parameter DC-Characteristics, Timing-Characteristics VVM = 24 V Symbol Short to GND detection level (VVM – VBM) VBMS2G Short to GND detector delay (HSx going active to short detector active / ERR_OUT falling) tS2G Supply current Parameter Conditions Min Typ Max Unit 1 1.5 2.3 V RS2G = 68k 200 320 450 ns RS2G = 150k 500 750 1000 ns RS2G = 220k 700 1000 1300 ns RS2G = 470k 1400 2000 2600 ns DC-Characteristics VVCC = 5.0 V, VVLS = 12V, VCPD = 10.5V, RSLP = 100k, VVM = 48V Symbol Conditions Min Typ Max Unit 0.45 0.68 mA 4.6 6.9 mA VM supply current IVM VLS supply current IVLS VCP supply current IVCP 1.6 2.4 mA VCC supply current IVCC 2.9 4.4 mA Min Typ Max Unit 7 7.85 8.5 V 5.8 6.6 Undervoltage detectors Parameter DC-Characteristics VVCC = 5.0 V Symbol VLS undervoltage level VVLSUV VCP undervoltage level (VVCPVM) VCPDUV VCP voltage OK level (VVCP-VM) not including I5VOUT Conditions VVCP falling VVCP rising VCP undervoltage detector Hysteresis Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG 7.1 0.5 V 7.8 V V TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Switching regulator / Charge pump Parameter 35 DC-Characteristics VVCC = V5VOUT Symbol Conditions Min Typ Max Unit -1.5 -2.2 -3.0 mA Switch output drive current (on) ISWOUT VSWOUT = VVM Switch output drive current (off) ISWOUT VSWOUT = VVM - 5V Switch start-up drive current during VCC undervoltage ISWOUT VSWOUT = VVM VVM = 24V VVLS < 2V Switch output drive voltage (on) VVM - VSWOUT VSWOUT Switch regulator output voltage V12VOUT 10 mA -0.4 -0.8 mA ISWOUT = 0 8 12 15 V VVM > 16V 11 12 13.1 V VVLSUV < VVM < 16V 0.85 VVM Oscillator output resistance RCOSC Lower oscillator threshold voltage VCOSC 1/3 VVCC V Upper oscillator threshold voltage VCOSC 2/3 VVCC V Oscillator threshold voltage for maximum duty cycle limit VCOSC 6/15 VVCC V Maximum duty cycle switch regulator TJ = 25°C V 14.1 k DCSWOUT VVLS = 10V fCHOP = 100kHz 63 70 77 % Switch frequency nominal fSW COSC = 470pF 70 100 130 kHz Switch frequency range (design reference value only) fSW 300 kHz Charge pump voltage (design reference value only) Linear regulator Parameter VCPD 0 (off) VVLS = 12V IVCP = 1.6mA Symbol Conditions Min Typ Max Unit 4.75 5.0 5.25 V 60 mV V5VOUT I5VOUT = 10mA TJ = 25°C Output resistance R5VOUT Static load Output current capability V DC-Characteristics Output voltage Deviation of output voltage over the full temperature range 10.6 2 V5VOUT(DEV) I5VOUT = 10mA TJ = full range I5VOUT 30 VVLS = 12V 100 mA VVLS = 8V 60 mA VVLS = 6.5V 20 mA Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Digital logic level Parameter 36 DC-Characteristics VVCC = 5.0 V +/-10% Symbol Conditions Min Typ Max Unit Input voltage low level VINLO -0.3 0.8 V Input voltage high level VINHI 2.0 VVCC+0.3 V 0.4 V Output voltage low (H1, H2, H3, ERR_OUT) VOUTLO IOUTLO = 1mA Output voltage high (H1, H2, H3) VOUTHI IOUTHI = -1mA Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG 0.8VVCC V TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Current measurement block 37 DC-Characteristics, Timing-Characteristics VVM = 24 V, VVCC = 5.0 V Parameter Symbol Conditions Min Typ Max Unit Amplification of voltage VFILTXRSX (or VBMX) to VCURX ACURLO+ SENSE_HI = GND -4.72 -4.82 -4.92 V/V ACURHI+ SENSE_HI = VCC -20.4 -20.8 -21.2 V/V Zero current level at CURX V0CURX VVCC/3 -50mV VVCC/3 -11mV VVCC/3 +30mV V Measurement voltage range at VBMX VBMX SENSE_HI = GND -300 300 mV SENSE_HI = VCC -70 70 mV VCURX output voltage swing low VCURX 0.1 V VCURX output voltage swing high VCURX Ripple voltage / hold step noise at output note 2) VCURX Minimum low side on time for current measurement (Delay from BLx going active to CURx tracking amplified signal) tBLHICURX SAMPLEx = VCC Delay from SAMPLEx going active to CURx tracking amplified signal tSMPHICURX SAMPLEx = VCC Delay from BLx or SAMPLEx going inactive to CURx hold 0.02 VVCC-1.2 VVCC-0.6 V VBMX = 0V SENSE_HI = GND 17 26 mV VBMX = 0V SENSE_HI = VCC 50 75 mV 5.3 7.2 µs 3.5 tBLHICURX /2 µs tBLXLO 0 µs Sample and hold drop during hold period dVCURX 0.001 1.6 V/s Auto zero drop of current amplifier during sampling period (low side on) dVCURX 0.003 3 V/s Minimum initial auto zero period (low side off or SAMPLEx low) after power on tBLXLO0 tSMPXLO0 5 µs Minimum refreshing time for auto zero during continuous measurement, e.g. each 10ms tSMPXLO 1 µs Minimum sample period after tBLHICURX for a 100% current step tBLXHIADD 1 µs 0.45 mA Output current limit at CURx ICURX Current sourcing 2) Note on first ICs TMC603 rather than TMC603A: CURx outputs are sensible to ripple voltage on VCC pin and frequency below 5MHz. Ripple voltage is amplified by 1/3 * Set amplification, i.e. factor 1.5 with SENSE_HI low and factor 6 with SENSE_HI high. Thus, it is suggested to use 5VOUT only for VCC supply, if possible, if exact measurements are required. This is corrected for TMC603A, ripple does not become amplified. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) Switched capacitor filter 2 order nd Parameter DC-Characteristics, AC-Characteristics VVM = 24 V, VVCC = 5.0 V Symbol Conditions Attenuation of voltage VBMX to VFILTX AFILTLO VBMX > 0.9V Output resistance of FILTX RFILTX Output current limit at FILTX IFILTX Noise voltage on FILTX 3dB bandwidth 38 Min fCUTOFF 3.0 fSCCLK = 2.5MHz Switched capacitor filter clock frequency for normal operation hallFX™ unit Parameter fSCCLK 30 Unit V 40 20 VFILTXNOISE VBMX = 12V fSCCLK = 1.25MHz fSCCLK = 1.25MHz Max VBMX/15 – 0.9V 21 Current sourcing Typ k µA 20 mV 1/390 fSCCLK Hz 3.2 3.4 6.4 0.25 kHz kHz 4 MHz Typ Max Unit 50 150 mV -400 0 400 mV DC-Characteristics, AC-Characteristics VVM = 24 V, VVCC = 5.0 V Symbol Conditions Min Noise voltage of comparators VCOMPNOISE VBMX = 12V including switched capacitor filter Offset voltage of comparators VCOMPOFFS VBMX = 12V including switched capacitor filter and input attenuation Spike suppression comparator threshold VSP_SUP VSP_SUP rising 2.0 VVCC/2 2.8 V Spike suppression capacitor charging current ISP_SUP VSP_SUP = 1V 15 25 35 µA Spike suppression capacitor discharging current ISP_SUP VSP_SUP = 1V -0.5 -1 -1.5 mA Dead time for spike suppression tSP_SUP CSP_SUP = 1nF 60 100 180 µs Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 39 8 Designing the application 8.1 Choosing the best fitting power MOSFET There is a huge choice of power MOSFETs available. MOSFET technology has been improved dramatically in the last 20 years, and gate drive requirements have shifted from generation to generation. The first generations of MOSFETs have a comparatively high gate capacity at a moderate RDSON. Their gate-source capacity is two to five times as high as the capacity of the gate-drain junction. These MOSFETs have a high gate charge and thus require high current gate drive, but they are easy to use, because internal feedback is low. In the early 2000s new MOSFETs have emerged, where RDSON is much lower, and gate-source capacity has been improved by minimizing structural overlap. Thus, the capacitance ratio has shifted, and feedback has become quite high. These MOSFETs thus are much more critical, and power drives have to actively force the gate off to prevent the bridges from cross-conduction due to feedback from the drain to gate. Latest generation MOSFETs, like the Vishay W-Fet technology, further reduce RDSON, while reducing the capacity between the channel and the drain. Thus, these MOSFETs have lowest gate charge, and again, are easier to control than the previous generation of MOSFETs. Further enhancements of MOSFETs have been done, to reduce the reverse recovery charge of the bulk diode. The bulk diode reverse recovery charge otherwise is a source for high current spikes an oscillations in push-pull output stages driving inductive loads like motor coils. When choosing the MOSFET, the following points shall be considered: Maximum voltage VDSS: Choose at least a few volts above your maximum supply voltage, taking into account that the motor can feed back energy when slowing down, and thus the supply voltage can rise. On the other hand, a transistor rated for a higher voltage is more expensive and has a higher gate charge (see next chapter). On-resistance RDSON: A low RDSON gives low static dissipation, but gate charge and cost increases. Take into account that a good part of the power dissipation results from the switching events in a chopped drive system. Further, to allow a current measurement, the RDSON should be in a range, that the voltage drop can be used for measurement. A voltage drop of 50mV or higher at nominal motor current is a good target. Gate charge QG and switching speed: The switching speed of the TMC603 application depends on the gate charge and the gate drive current setting. The switching speed should be compared to the required chopper frequency. Choose the chopper frequency low to reduce dynamic losses. When the application does not require slow, EMV optimized switching slopes, choose a low gate charge transistor to reduce dynamic losses. Gate threshold voltage VGS(TH): Most MOSFETs have a specified on-resistance at a gate drive voltage of 10V. Some MOSFETs are optimized for direct control from logic ICs with 5 or even 3.3V. They provide a low gate threshold voltage of 1V to 2V. MOSFETs with higher gate threshold voltage should be preferred, because they are less sensible to effects of the drain gate capacity and cross conduction. Reverse recovery charge QRR of bulk diode: A lower reverse recovery charge QRR and lower reverse recovery time tRR reduce peak currents in the bridge and allow for faster switching. Snubber elements at the output are required for high reverse recovery charge transistors. Otherwise, Schottky diodes should be used to bridge the bulk diode. Package, size and cooling requirements Cost and availability Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 40 8.1.1 Calculating the MOSFET power dissipation The power dissipation in the MOSFETs has three major components: Static losses (PSTAT) due to voltage drop, switching losses (PDYN) due to signal rise and fall times, losses due to diode conduction (PDIODE). The diode power dissipation depends on many factors (back EMF of the motor, inductivity and motor velocity), and thus is hard to calculate from motor data. Normally, it contributes for a few percent to some ten percent of overall power dissipation. Other sources for power dissipation are the reverse recovery time of the transistors and the gate drive energy. Reverse recovery also causes current spikes on the bridges. If desired, you can add Schottky diodes over the (chopper) transistors to reduce the diode losses and to eliminate current spikes caused by reverse recovery. The following sample calculation assumes a three phase BLDC motor operated in block commutation mode with dual sided chopper. At each time, two coils conduct the full motor current (chopped). where IMOTOR is the motor current, e.g. 10A RDSON is the on-resistance of the MOSFETs at a gate voltage of about 10V, e.g. 20mΩ tDUTY is the actual duty cycle of the chopper, e.g. 80% = 0.8 VVM is the motor supply voltage, e.g. 24V or 48V fCHOP is the chopper frequency, e.g. 20kHz tSLOPE is the slope (transition) time, e.g. 300ns Example: With the example data for a 10A motor at 24V, we get the following power dissipation: PSTAT = 3.2W PDYN24 = 2.88W For comparison: The motor output power is 10A*24V*0.8=192W The dynamic and static dissipation here are in a good ratio, thus the choice of a 20mΩ MOSFET is good. At 48V, the dynamic power dissipation doubles: PDYN48=5.76W Here, the dynamic losses are higher than the static losses. Thus, we should reduce the slope time. Given that the gate capacity would not allow for faster slopes than 300ns, we could go for a 30mΩ MOSFET, which has a lower gate charge and thus allows faster slopes, e.g. 200ns. With these modifications we get a static loss of 4.8W and a dynamic loss of 3.84W. This in sum is 8.64W, which is slightly less than the 8.96W before. At the same time, system cost has decreased due to lower cost MOSFETs. The loss is still low when compared to a motor power of 384W. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 8.2 41 MOSFET examples There is a huge number of MOSFETs on the market, which can be used in combination with the TMC603. The user choice will depend on the electrical data (voltage, current, RDSon) and on the package and configuration (single / dual). The following table gives a few examples of SMD MOSFETs for different motor currents. The MOSFETs explicitly are modern types with a low total gate charge. With dual configurations, only three MOSFET packages are required to control a BLDC motor, but the current which can be reached is significantly lower due to thermal restrictions of the packages. For the actual application, we suggest to calculate static and dynamic power dissipation for a given MOSFET, as described in the previous chapter. Especially for sine commutation and chopper frequencies above 20kHz, transistors with a gate charge below 100nC should be preferred. Transistor type manufacturer unit RDSon voltage mΩ V package & max. motor configuration current (*) total gate charge @10V A nC IBP019N06L3 Infineon 1.9 60 D2PACK 30 124 IPP032N06N3 Infineon 2.9 60 TO220 30 125 IRFB3306 International Rectifier 4.2 60 TO220 / D2PACK 30 85 SiE876DF Vishay 6.1 60 PolarPAK 20 51 SI7164DP Vishay 6.25 60 PowerPAK SO-8 15 50 SUM75N0609L Vishay 9.3 60 D2PAK (TO263) 25 47 FDD10AN06A0 Fairchild 10.5 60 DPAK (TO252A) 20 28 FDD5353 Fairchild 12.3 60 DPAK 15 46 SI7964DP Vishay 23 60 PowerPAK SO-8 (dual) 9.6 43 SI4946 Vishay 55 60 SO-8 (dual) 4.5 19 SiE868DF Vishay 2.3 40 PolarPAK 30 95 SI7994DP Vishay 5.6 30 PowerPAK SO-8 (dual) 10 52 (*) Remark: The maximum motor current applicable in a given design depends upon PCB size and layout, since all of these transistors are mainly cooled via the PCB. The data given implies adequate cooling measures taken by the user, especially for higher current designs. Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 8.3 42 Programming a block commutation for hallFX™ In order to operate a motor using a hall sensor or hallFX™, the user processor needs to provide a commutation decoder. Also, commutation checking makes sense, to determine the direction of operation. The commutation logic decodes the hall sensor signal to provide standard block commutation patterns. There are six different valid hall sensor codes. Each of these represents a different position of the rotor. In order to turn the rotor, a magnetic field has to be provided by the motor’s stator coils, which is shifted by an commutation angle of +90° or by -90° for CW respectively CCW rotation. Since the hall sensor provides a 60° resolution, the commutation logic can keep the phase angle always between +60° to +120° respectively -60° to -120°. The mean value is the desired +/-90°. In block commutation, one motor phase terminal is open (Z) at each phase pattern, while the current flows through the other two phases. One of these two phases is switched to the motor supply voltage (1), the other one to GND (0). For hallFX™, both of these are chopped between (Z) and (1), respectively (Z) and (0) in order to modulate the motor power. The commutation table shows the block commutation decoder logic. Hall pattern H1 H2 H3 U1 V1 W1 U0 V0 W0 1 1 1 0 Z 0 1 Z 1 0 2 0 1 0 1 0 Z 0 1 Z 3 0 1 1 1 Z 0 0 Z 1 4 0 0 1 Z 1 0 Z 0 1 5 1 0 1 0 1 Z 1 0 Z 6 1 0 0 0 Z 1 1 Z 0 Z: Coil output open 0: Coil output pulled low or negative PWM 1: Coil output pulled high or positive PWM U0, V0, W0: Pattern with positive direction (dir = 0). This is the pattern for hallFX™ in both directions. U1, V1, W1: Pattern with negative direction (dir = 1) U U DIR=1 V V W 6 5 4 W 2 3 1 U U DIR=0 V W 1 2 3 V W 4 5 6 figure 22: commutation sequence 8.4 Driving a DC motor with the TMC603 The TMC603 can also be used for DC motor applications, using a full bridge or a half bridge for motor PWM operation with or without reverse direction operation. For single half bridge applications, all TMC603 gate drivers can be paralleled, taking advantage of the three time increase in gate drive capability up to 450mA. This way a motor current of up to 100A can be driven. The drive system can use the shunt less current sensing capability for best efficiency. A Schottky diode across the nonchopped transistor optimizes slopes and electromagnetic emission characteristics (see chapter 5.2.8). Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 43 9 Table of figures FIGURE 1: APPLICATION BLOCK DIAGRAM ..................................................................................................... 5 FIGURE 2: PINNING / QFN52 PACKAGE (TOP VIEW) ....................................................................................... 6 FIGURE 3: APPLICATION DIAGRAM ................................................................................................................ 8 FIGURE 4: THREE PHASE BLDC DRIVER..................................................................................................... 10 FIGURE 5: PRINCIPLE OF HIGH-SIDE DRIVER (PAT. FIL.) ............................................................................... 10 FIGURE 6: BRIDGE DRIVER TIMING ............................................................................................................. 12 FIGURE 7: EXAMPLES FOR MICROCONTROLLER PWM CONTROL .................................................................. 13 FIGURE 8: MOSFET GATE CHARGE AS AVAILABLE IN DEVICE DATA SHEET VS. SWITCHING EVENT ................. 13 FIGURE 9: QGD PROTECTED DRIVER STAGE .............................................................................................. 14 FIGURE 10: EFFECT OF BULK DIODE RECOVERY .......................................................................................... 16 FIGURE 11: PARALLEL SCHOTTKY DIODE AVOIDS CURRENT SPIKES DUE TO BULK DIODE RECOVERY, OPTIONAL SNUBBER REDUCES OVERSHOOT AND RINGING ................................................................................... 17 FIGURE 12: TIMING OF THE SHORT TO GND DETECTOR .............................................................................. 17 FIGURE 13: ERROR LOGIC ......................................................................................................................... 18 FIGURE 14: SCHEMATIC OF CURRENT MEASUREMENT AMPLIFIERS ............................................................... 19 FIGURE 15: TIMING OF THE CURRENT MEASUREMENT ................................................................................. 20 FIGURE 16: HALLFX™ BLOCK DIAGRAM AND TIMING ................................................................................... 22 FIGURE 17: HALLFX™ BASED COMMUTATION ............................................................................................ 24 FIGURE 18: POWER SUPPLY BLOCK WITH EXAMPLE VALUES ........................................................................ 26 FIGURE 19: ADDING A REVERSE POLARITY PROTECTION.............................................................................. 29 FIGURE 20: LOW POWER STANDBY ............................................................................................................ 29 FIGURE 21: LOW VOLTAGE OPERATION ...................................................................................................... 30 FIGURE 22: COMMUTATION SEQUENCE ...................................................................................................... 42 Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG TMC603A DATA SHEET (V. 1.16 / 2010-May-14) 44 10 Revision History 10.1 Documentation Revision Version Author Description (BD=Bernhard Dwersteg) 0.94 BD TMC603 initial release with preliminary electrical data 0.96 BD Added package dimensions 0.98 BD Added microcontroller PWM control examples 0.99 BD Added reverse polarity protection and MOSFET examples 1.00 BD Added low power standby and low voltage operation 1.01 BD Removed “preliminary” indication, modifications in electrical characteristic tables 1.02 BD Slightly corrected a few values 1.03 BD Added transistor examples and temperature information to tables 1.04 BD Slight beautifications / rewording 1.05 BD Added mathematical background for QGD protection, discussion on MOSFET bulk diode and DC motor application 1.06 BD Added minimum output voltage swing of current amplifiers 1.10 BD TMC603A preliminary specs, changed date format YYYY-MON-DD 1.11 BD Added 5Vout temperature deviation and detailed current measurement refreshing using sample input 1.12 BD Added block commutation example and notes on capacitor selection, ESD 1.14 BD TMC603A electrical data update 1.15 BD Some cosmetic changes 1.16 BD Some cosmetic changes Table 1: Documentation Revisions Copyright © 2009 TRINAMIC Motion Control GmbH & Co. KG