HI TI Team!
I will try to describe my problem on prototype that we are preparing for mass production.
Let me first quickly describe what we want to achieve. Our CAN protocol is designed like this that in our grid there is master which is always asking and node always replying. Master is usually asking for parameter every 8ms. Parameters are incremented by master if it received answer successfully. Thant means that whole protocol is circling ( N*8ms) and then start owner after last parameter. If MASTER does not get right answer on certain parameter it will ask again after 64ms for 20 times then it will throw error. Answer is from node is created as soon it receive new data.
Software that we are using is your standard code for pooling data:
if (TCAN_Int_Cnt != 0 )
{
TCAN_Int_Cnt--;
TCAN4x5x_Device_ReadInterrupts(&dev_ir); // Read the device interrupt register
TCAN4x5x_MCAN_ReadInterrupts(&mcan_ir); // Read the interrupt register
if (dev_ir.SPIERR) // If the SPIERR flag is set
TCAN4x5x_Device_ClearSPIERR(); // Clear the SPIERR flag
if (mcan_ir.RF0N) // If a new message in RX FIFO 0
{
//TCAN4x5x_MCAN_RX_Header MsgHeader = {0}; // Initialize to 0 or you'll get garbage
uint8_t numBytes = 0; // Used since the ReadNextFIFO function will return how many bytes of data were read
//rx_data = {0}; // Used to store the received data
TCAN4x5x_MCAN_ClearInterrupts(&mcan_ir); // Clear any of the interrupt bits that are set.
numBytes = TCAN4x5x_MCAN_ReadNextFIFO(RXFIFO0, &RXMsgHeader, rx_data); // This will read the next element in the RX FIFO 0
// numBytes will have the number of bytes it transfered in it. Or you can decode the DLC value in MsgHeader.DLC
// The data is now in rx_data[], and message specific information is in the MsgHeader struct.
tcan_msg_ID();
}
}
iside function "tcan_msg_ID();" is this part of code
else if(0x6A9)
{
can_msg_buff = 0;
can_msg_buff = rx_data[0];
switch(can_msg_buff)
{
case 0x10:
tcan_msg_ID_ack(STAT_OK,0x255,0x10,0x00,0x00,0x00,0x00);
break;
case 0x22:
tcan_msg_ID_ack(STAT_OK,0x255,0x22,0x00,0x00,0x00,0x00);
break;
default:
tcan_msg_ID_ack(STAT_OK,0x255,can_msg_buff,0x00,0x00,0x00,0x00);
break;
......
......
}
void tcan_msg_ID_ack(uint8_t can_Status, uint16_t msg_ID, uint16_t par_ID, uint8_t data_0,uint8_t data_1,uint8_t data_2,uint8_t data_3)
{
TCAN4x5x_MCAN_TX_Header TXMsgHeader = {0};
TXMsgHeader.ID = msg_ID;
tx_data[0] = can_Status;
tx_data[1] = par_ID & 0x00FF;
tx_data[2] = par_ID>>8;
tx_data[3] = data_0;
tx_data[4] = data_1;
tx_data[5] = data_2;
tx_data[6] = data_3;
tx_data[7] = 0x00;
TXMsgHeader.RTR = 0; //Remote Transmission Request flag
TXMsgHeader.XTD = 0; // Extended Identifier flag
TXMsgHeader.ESI = 0; // Error state indicator flag
TXMsgHeader.DLC = MCAN_DLC_8B;
TXMsgHeader.BRS = 1; //Bit rate switch used flag
TXMsgHeader.FDF = 1; //CAN FD Format flag
TXMsgHeader.reserved = 0; //Reserved //1
TXMsgHeader.EFC = 0; //Event FIFO Control flag, to store tx events or not
TXMsgHeader.MM = 0; //Message Marker, used if @c EFC is set to 1 //1
TCAN4x5x_MCAN_WriteTXBuffer(0, &TXMsgHeader, tx_data);
TCAN4x5x_MCAN_TransmitBufferContents(0);
}
System example from.
0x6A9 is a master asking for parameter DB0 = 0x10
0X255 is node answering on parameter DB0 = 0x01( status OK), DB1 = 0x10
Now lets go to problem.
It was really hard to reproduce problem but i manage to do it with external PEAK CAN dongle.
These are the settings on PEAK dongle:
Dongle is acting as MASTER in my system. I am sending message 0x6A9 cyclic ever 10ms and message 0x583 every 1ms before and after 06A9 just to create traffic. I am expecting answer on ID 0x255 with parameter 0x40.
Most of the time answer is right but now and then there is wrong answer. Of course i have set filter in TCA4550 for message 0x583 but i newer replay on it, my software is just listening this message.
Can you explain this behaver?
Regards.
Sebastian
