AN-012: Extending the Positioning Range of the TMC429 (or TMC428) to 32-bit or more

Understanding the Algorithm

The TMC429 and the TMC428 have 24-bit registers for position information. These are the registers X_TARGET and X_ACTUAL available for each of the three motor drivers. When doing a positioning movement, the motion direction is determined by the sign of the 24-bit difference between both values. This way, the maximum displacement per move is (223)−1, i.e. 8388607.

The algorithm for position extension requires two additional registers within the microcontroller: A 32-bit target position (XTARGET32) and a 32-bit actual position (XACTUAL32).

Starting a positioning sequence:

  1. XTARGET32 becomes updated with the with the new target position.
  2. The regular service procedure becomes executed.

Regular service procedure:

This procedure becomes executed on a regular base, taking into account maximum velocity, it needs to be executed at least each 2 seconds assuming maximum TMC429 clock frequency (32MHz).

  1. Each control cycle starts with reading X_ACTUAL from the TMC429.
  2. X_ACTUAL becomes checked for an underflow or overflow, by comparing to the lower 24 bits of the old value stored in XACTUAL32. In case of an underflow, the upper 8 bits of XACTUAL32 become decremented by 1. In case of an overflow, they become incremented by one. The lower 24 bits of XACTUAL32 become updated with the value read from X_ACTUAL.
  3. This step is required in positioning mode, only: XTARGET32 becomes compared to XACTUAL32. In case, the difference is larger than 222, X_TARGET becomes set to XACTUAL32 + 222, using the lower 24 bits of the result. In case the difference is smaller than −222, X_TARGET becomes set to XACTUAL32 − 222 the same way. In case the difference is in between, X_TARGET becomes set to XTARGET32 using the lower 24 bits.

As an additional extension, the position compare mechanism must be switched off in case the upper eight bits of XACTUAL32 do not match the upper 8 bits of XTARGET32. When a position snapshot is triggered by a reference switch, the upper 8 bits of XACTUAL32 need to be copied to the 32-bit copy of the snapshot register, too.

An implementation in C

A sample implementation is shown here. It uses the TMC429 library provided with the TMC429+TMC26x-EVAL and extends this so that the 32-bit positioning extension can be used in a transparent way. For this purpose, the routines Write428Int and Read428Int are extended in a way that accessing the XTARGET or XACTUAL register will access the 32-bit variables holding the 32-bit position values instead. The routine Check32BitExtension() is provided to be called on a regular basis (at least every two seconds), and the routine TrackPosition() is just helper function for the Check32BitExtension() routine.

typedef struct
{
 int XTarget; //Target position (32 Bit)
 int XActual; //Actual position (32 Bit)
 int XOld; //last 24 bit position (for position counting)
 UCHAR ExtPosFlag; //TRUE when positioning more than 8388607 steps
} TExtPos;
TExtPos ExtendedPositions[3]; //32 bit position registers for each axis
void Write428Int(UCHAR Address, int Value)
{
 UCHAR Write[4], Read[4];
 UCHAR Motor;
 int Value2;
 Motor=(Address & 0x60)>>5;
 if(Address<0x60)
 {
 switch(Address & 0x1e)
 {
 case IDX_XTARGET:
 //When changing XTARGET the value will be corrected in a way that 
 //at first we don’t move more than 8388607 microsteps. 
 //During moving the value of XTARGET will be adapted successively 
 //in Check32BitExtension() until we are less than 8388607 
 //microsteps away from our 32 bit target position.
 ExtendedPositions[Motor].XTarget=Value;
 if(abs(ExtendedPositions[Motor].XTarget- 
 ExtendedPositions[Motor].XActual)>8388607)
 {
 if(ExtendedPositions[Motor].XTarget>ExtendedPositions[Motor].XActual)
 Value=ExtendedPositions[Motor].XActual+8388607;
 else
 Value=ExtendedPositions[Motor].XActual-8388607;
 ExtendedPositions[Motor].ExtPosFlag=TRUE;
 }
 else ExtendedPositions[Motor].ExtPosFlag=FALSE;
 break;
 case IDX_XACTUAL:
 //When changing XACTUAL the value of XTARGET will be corrected in a way that 
 //at first we don’t move more than 8388607 microsteps. 
 //During moving the value of XTARGET will be adapted successively 
 //in Check32BitExtension() until we are less than 8388607 
 //microsteps away from our 32 bit target position.
 ExtendedPositions[Motor].XActual=Value;
 ExtendedPositions[Motor].XOld=Value & 0x00ffffff;
 if(abs(ExtendedPositions[Motor].XTarget-ExtendedPositions[Motor].XActual)>8388607)
 {
 if(ExtendedPositions[Motor].XTarget>ExtendedPositions[Motor].XActual)
 Value2=ExtendedPositions[Motor].XActual+8388607;
 else
 Value2=ExtendedPositions[Motor].XActual-8388607;
 Write[0]=IDX_XTARGET|MOTOR_NUMBER(Motor)<<5;
 Write[1]=Value2 >> 16;
 Write[2]=Value2 >> 8;
 Write[3]=Value2 & 0xff;
 ReadWrite428(Read, Write);
 ExtendedPositions[Motor].ExtPosFlag=TRUE;
 }
 else ExtendedPositions[Motor].ExtPosFlag=FALSE;
 break;
 }
 }
 Write[0]=Address;
 Write[1]=Value >> 16;
 Write[2]=Value >> 8;
 Write[3]=Value & 0xff;
 ReadWrite428(Read, Write);
}
void TrackPosition(UCHAR Axis, int New24BitPosition)
{
 int Diff24_1;
 int Diff24_2;
 if(New24BitPosition>ExtendedPositions[Axis].XOld)
 {
 Diff24_1=New24BitPosition-ExtendedPositions[Axis].XOld;
 Diff24_2=0xffffff - Diff24_1 + 1;
 if(Diff24_1<=Diff24_2)
 {
 ExtendedPositions[Axis].XActual+=Diff24_1;
 }
 else
 {
 ExtendedPositions[Axis].XActual-=Diff24_2;
 }
 }
 else if(New24BitPosition<extendedpositions[axis].xold) {="" diff24_1="ExtendedPositions[Axis].XOld-New24BitPosition;" diff24_2="0xffffff-Diff24_1+1;" if(diff24_1<="Diff24_2)" extendedpositions[axis].xactual-="Diff24_1;" }="" else="" extendedpositions[axis].xactual+="Diff24_2;" extendedpositions[axis].xold="New24BitPosition;" void="" check32bitextension(uchar="" axis)="" uchar="" read[4],="" write[4];="" int="" actual24bitposition;="" value;="" rampmode;="" check="" the="" actual="" ramping="" mode="" write[0]="Axis<<5|IDX_REFCONF_RM|TMC428_READ;" readwrite428(read,="" write);="" rampmode="Read428[3];" 24="" bit="" position="" and="" update="" "virtual"="" 32="" register.="" in="" this="" routine,="" write428int()="" or="" read428int()="" must="" not="" be="" used!="" actual24bitposition="(Read[1]<<16)|(Read[2]<<8)|(Read[3]);" trackposition(axis,="" actual24bitposition);="" target="" register="" will="" adapted="" successively="" to="" when="" there="" is="" a="" move="" of="" more="" than="" 8388607="" microsteps="" progress.="" if(extendedpositions[axis].extposflag="" &&="" if(abs(extendedpositions[axis].xtarget-extendedpositions[axis].xactual)="">8388607)
 {
 if(ExtendedPositions[Axis].XTarget>ExtendedPositions[Axis].XActual)
 Value=ExtendedPositions[Axis].XActual+8388607;
 else
 Value=ExtendedPositions[Axis].XActual-8388607;
 }
 else
  Value=ExtendedPositions[Axis].XActual-8388607;
 }
 else
 {
 ExtendedPositions[Axis].ExtPosFlag=FALSE;
 Value=ExtendedPositions[Axis].XTarget;
 }
 //Don’t use Write428Int() here!
 Write428Datagram(Axis<<5|IDX_XTARGET, Value >> 16, Value >> 8, Value & 0xff);
 }
}
</extendedpositions[axis].xold)>


作者

Olav_Kahlbaum_blue

Olav Kahlbaum

Olav Kahlbaum works as a Staff Engineer for Firmware at Trinamic part of Analog Devices based in Germany. After finishind his studys of Computer Science in 1998 he joined the Trinamic team in 2000. He mainly develops firmware for stepper motor control modules. During his career at Trinamic he has become one of the inventors of TMCL and has specialized very much in communication protocols like CANopen, EtherCAT and Modbus.