NXC: I²C drivers for Hitechnic Tetrix motor muxer ?

[gloomyandy]

Looks to me like you are getting an error from the i2cwrite function. Probably worth adding an extra print if you get an error return. At a guess I would say that you can't do two writes to the device very close together and adding the sound function delays things enough so that the hardware has sorted itself out and is ready to accept the next command. Strange stuff hardware (especially "toy" hardware that is probably only tested with the standard firmware and programming language)...

Andy

[HaWe]

you're probably right -
but strangely I get the error at the first i2cwrite (modeMsg), not the direct following one (powerMsg)

Code: Select all

#define printf2( _x, _y, _format1, _format2, _value1, _value2) { \
  string sval1 = FormatNum(_format1, _value1); \
  string sval2 = FormatNum(_format2, _value2); \
  string s =sval1+sval2; \
  TextOut(_x, _y, s); \
}

void TXMotorOn(byte NXTport, byte TXmotor, char percentage)
{
  const byte retLen = 0;
  int i, j;
  char modeMsg[];
  char powerMsg[];
  byte devAddr, modeReg, powerReg;
  char result;

  // constant values for testing
  devAddr = 0x02;
  modeReg = 0x44 ;
  powerReg= 0x45 ;
  
    // Create the message arrays
  ArrayBuild(modeMsg, devAddr, modeReg, 0);
  ArrayBuild(powerMsg, devAddr, powerReg, percentage);

  // Send the first message as soon as the bus is ready
  while ((I2CCheckStatus(NXTport) == STAT_COMM_PENDING) && (I2CCheckStatus(NXTport) != NO_ERR)) Yield();
  if(I2CWrite(NXTport, retLen, modeMsg) != NO_ERR)  
    { printf2(0,8,"rErr%2d"," vErr=%4d",modeMsg[1],modeMsg[2]); return;  } // Error msg displayed!!

  printf2(0,56,"mReg%2d"," mVal=%4d",modeMsg[1],modeMsg[2]);
  
  // Send the second message when the first one is done
  while ((I2CCheckStatus(NXTport) == STAT_COMM_PENDING) && (I2CCheckStatus(NXTport) != NO_ERR)) Yield();
  if(I2CWrite(NXTport, retLen, powerMsg) != NO_ERR)   
    { printf2(0,0,"pErr%2d"," vErr=%4d",powerMsg[1],j=powerMsg[2]); return;  } //*no* Error displayed!!
  
  printf2(0,48,"pReg%2d"," pVal=%4d",powerMsg[1],j=powerMsg[2]);
}

task main(){

  SetSensorType(S1, SENSOR_TYPE_LOWSPEED);


  TXMotorOn(0, 1, 50);
  Wait(2000);

  TXMotorOn(0, 1,  0);  // stop
  Wait(1000);

  TXMotorOn(0, 1,-50);
  Wait(2000);

  TXMotorOn(0, 1, -128); // coast
  Wait(2000);

  while (true);
}

that means (CMIIW) that already the first of the two i2cwrites fails, then the execution of the subroutine breaks and returns to main - so the 2nd i2cwrite has no chance at all to be executed any more...

but why-oh-why does it then execute the 3rd cmd (go reverse -50) ???

that's really weird...

[HaWe]

BTW:
the Tetrix muxer works with NXT-G and RobotC.
IIRC the Mindsensors guys once made some changes to the I2c protocol.....? (can't find the thread right now....)

[gloomyandy]

It may be that the device need a little time to settle after it has been powered on/initialized... I notice that your working code has a sound being played between initializing the port and sending the first command, perhaps that gives time for things to settle?

[mightor]

IIRC the Mindsensors guys once made some changes to the I2c protocol.....? (can't find the thread right now....)

The only change they made was to allow it to work with much longer cables. It was a small timing fix. The protocol has not changed.

If you have trouble with I2C commands, the first thing to try is put more time between each one.

- Xander

[HaWe]

ok, I will try, but as I already said:
I (actually it was YOU who wrote this code) always send a couple of i2c cmds - 1st the mode cmd, then the power cmd.
Now it's not the time BETWEEN the i2c cmds that causes the failure, it's already the 1st one of the couple (mode) that doesn't execute (instead the funtion executes the line behind

Code: Select all

if(I2CWrite(NXTport, retLen, modeMsg) != NO_ERR) {...return}

and breaks the subroutine.

If it was a matter of time delay between both of them, the 1st one should be executed anyway and the 2nd one should cause the error (show error msg and then return to main) - but it doesn't.

The "sound-check" that I inserted is started before the subroutine is called and so actually must not have any effect of the call inside the subroutine - but nevertheless it obviously has.

And I really don't get that.

[mightor]

I wish I had one of each of these controllers so I could test this code. It would make debugging this a lot easier.

Try to capture the actual error code that call returns, we know it's not NO_ERR, but what is it?

- Xander

[HaWe]

hi,
I added some debugging lines, and it says
error=-32

Code: Select all

#define printf2( _x, _y, _format1, _format2, _value1, _value2) { \
  string sval1 = FormatNum(_format1, _value1); \
  string sval2 = FormatNum(_format2, _value2); \
  string s =sval1+sval2; \
  TextOut(_x, _y, s); \
}

void TXMotorOn(byte NXTport, byte TXmotor, char percentage)
{
  const byte retLen = 0;
  int i, j;
  char modeMsg[];
  char powerMsg[];
  byte devAddr, modeReg, powerReg;
  char IOresult;

  // constant values for testing
  devAddr = 0x02;
  modeReg = 0x44 ;
  powerReg= 0x45 ;
  
    // Create the message arrays
  ArrayBuild(modeMsg, devAddr, modeReg, 0);
  ArrayBuild(powerMsg, devAddr, powerReg, percentage);

  // Send the first message as soon as the bus is ready
  while ((I2CCheckStatus(NXTport) == STAT_COMM_PENDING) && (I2CCheckStatus(NXTport) != NO_ERR)) Yield();
  IOresult=I2CWrite(NXTport, retLen, modeMsg);
  if (IOresult != NO_ERR)
    { printf2(0,8,"mErr%2d"," Err=%4d",modeMsg[1],IOresult); return;  }  // Error debug: error =-32

  printf2(0,56,"mReg%2d"," mVal=%4d",modeMsg[1],modeMsg[2]);
  
  // Send the second message when the first one is done
  while ((I2CCheckStatus(NXTport) == STAT_COMM_PENDING) && (I2CCheckStatus(NXTport) != NO_ERR)) Yield();
  IOresult=I2CWrite(NXTport, retLen, powerMsg);
  if (IOresult != NO_ERR)
    { printf2(0,0,"pErr%2d"," Err=%4d",powerMsg[1],IOresult); return;  }  // Error debug: no error
  
  printf2(0,48,"pReg%2d"," pVal=%4d",powerMsg[1],j=powerMsg[2]);
}

task main(){

  SetSensorType(S1, SENSOR_TYPE_LOWSPEED);

  TXMotorOn(0, 1, 50);
  Wait(2000);

  TXMotorOn(0, 1,  0);  // brake
  Wait(1000);

  TXMotorOn(0, 1,-50);
  Wait(2000);

  TXMotorOn(0, 1, -128); // coast
  Wait(2000);
  while (true);
}

[HaWe]

ps:
Xander, I can lend you an extra Tetrix muxer (=controller), and 2 extra Tetrix motors, and 1 extra encoder.

The stuff actually is not mine but from a friend (Frank from the German Mindstorms Forum who is also interested in a NXC driver). I also could send you my set, but it's better if you his stuff , so that I can test simultaneously the things you do.

Additionally I can lend you my own Servo muxer (plugged to the motor muxer) with 2 servos , in the moment in don't use them.

Send me a mail with your address, I will arrange that if you wish!

[mightor]

Helmut,

Let me think about it I have two spare motors and two spare servos at the moment. I was recently given a VEX Cortex Super Bundle to play with, so I have plenty of hardware.

I have a fair bit of stuff I am testing and playing with at the moment, I'll have to get back to you on your offer

- Xander