2023-07-10 10:06 AM
I am using the Nucleo-F091RC board to communicate with a Motor that allows me to change it's speed with CAN. I have the nucleo board attached to an Adafruit CAN Pal board which has a TJA1051T/3 from NXP as the CAN transceiver but I have yet to actually connect the motor to the CAN_H and CAN_L outputs from the transceiver. I am powering the board with 3.3V generated by the nucleo board and the CAN signals attached to PA12 and PA11 and Silent pin connected to PB12 generating a low signal all the time to make sure the transceiver is in "normal" mode. I have been able to get the nucleo board to output a CAN message every 65ms and I can see it on the CAN_TX pin being output from the nucleo board to the adafruit CAN Pal. When I attach my scope to the output of the transceiver (CAN_H) the signal is always High...no data.
Here is my CAN setup code:
static void MX_CAN_Init(void)
{
/* USER CODE BEGIN CAN_Init 0 */
/* USER CODE END CAN_Init 0 */
/* USER CODE BEGIN CAN_Init 1 */
/* USER CODE END CAN_Init 1 */
hcan.Instance = CAN;
hcan.Init.Prescaler = 16;
hcan.Init.Mode = CAN_MODE_NORMAL;
hcan.Init.SyncJumpWidth = CAN_SJW_2TQ;
hcan.Init.TimeSeg1 = CAN_BS1_2TQ;
hcan.Init.TimeSeg2 = CAN_BS2_3TQ;
hcan.Init.TimeTriggeredMode = DISABLE;
hcan.Init.AutoBusOff = DISABLE;
hcan.Init.AutoWakeUp = ENABLE;
hcan.Init.AutoRetransmission = DISABLE;
hcan.Init.ReceiveFifoLocked = DISABLE;
hcan.Init.TransmitFifoPriority = ENABLE;
if (HAL_CAN_Init(&hcan) != HAL_OK)
{
Error_Handler();
}
/* USER CODE BEGIN CAN_Init 2 */
HAL_GPIO_WritePin(CAN_Silent_GPIO_Port, CAN_Silent_Pin, CAN_SILENT_DEFAULT);
/* USER CODE END CAN_Init 2 */
}
Since the motor can't talk back to me I am not setting up anything to do with receiving data from the CAN bus. Can anyone help me with why I would be seeing data coming from the MCU but then the transceiver not then transmitting that data onto the bus? Does the transceiver need to be connected to a CAN Bus in order for data to appear on it? I've tested similar CAN devices that transmit data and they do not seem to have to be connected to the bus for data to appear on CAN_H.
Solved! Go to Solution.
2023-07-12 11:17 AM
My apologies...I used the oscilloscope probes instead of the logic probes and now I see both CAN_H and CAN_L signals as I would expect. I am still new to CAN and I wasn't aware that the "zero" for CAN is actually 2.5V...with that in mind here is the CAN_H and CAN_L with the decoded message coming from the Nucleo board:
As for why the motor is not responding I believe I may have to take that up with the motor manufacturer. Thank you for your assistance!
2023-07-10 01:49 PM
Ok, So I've been able to hook it up to the motor now and I am getting a ton of traffic on the CAN Bus (I'm only connected to the motor, there is nothing else on the bus). As soon as I do this the CAN_TX line out of the Nucelo board doesn't do anything..it's just goes low and then high really quick and then the code hangs and nothing works. On the CAN_H pin of the CAN bus I see a ton of data coming through. Am I required to set up the RX side of things in my code and then clear some register out so it doesn't cause the code the hang? If so, what do I need to set up for the STM32F091RC MCU?
2023-07-10 02:23 PM
Yes, the TJA1051 CAN bus needs to be connected to the motor CAN bus, along with the terminating resistors. Otherwise it'll detect a bus fault and go into a bus-off state.
2023-07-10 02:32 PM
Thank you, that answers one question and explains why when the motor is connected I get traffic on the CAN bus. That said, for the first 5 seconds after power up I get TX data out of the MCU and nothing on the CAN bus and then after about 5 seconds the MCU hangs and stops working all together and then the CAN Bus comes alive with a ton of traffic (presumably from the motor and not the MCU).
As I mentioned in my last post, I have not implemented any CAN receive code since I am not trying to talk back and forth with the motor...just give it commands. Do I need to implement some sort of receive routines in order to not crash my code when there is data on the bus coming back at me?
I've attached screen captures of the CAN_TX line from the MCU as well as the CAN_H signal on the CAN Bus to show you what I'm seeing.
2023-07-11 03:14 AM
Hello,
Transmit and Receive are independent. So if your application needs only transmitting messages it's ok and no need for Rx implementation.
Apart of this:
hcan.Init.Prescaler = 16; --> try to decrease this value
hcan.Init.TimeSeg1 = CAN_BS1_2TQ; --> try to increase this value
hcan.Init.TimeSeg2 = CAN_BS2_3TQ; --> try to increase this value
And select BS1 > BS2. So sample point will be around 80%.
2023-07-11 04:00 AM
If the transmit messages are not properly ACKnowledged, the TX FIFO might get full, then it stops sending data, so check all TX buffers / FIFOs / memories.
I recommend to receive CAN data anyway, for the sake of learning and controlling the peripheral - and out of curiosity what's happening on the bus...
2023-07-11 04:59 AM
Thank you! I have changed to:
hcan.Init.Prescaler = 8;
hcan.Init.TimeSeg1 = CAN_BS1_5TQ;
hcan.Init.TimeSeg2 = CAN_BS2_6TQ;
These settings allow me to maintain 500k baud rate and it gives me a MUCH more stable CAN Bus and the whole transmission on the CAN_TX line can now be properly decoded by my oscilliscope. Unfortunently after about 5 seconds the transmission cuts out still. After doing some searching I also came across people recommending enabling the Automatic Bus-Off Management. When I do this, now after the 5 seconds you can see it cut out but then come right back...and it does this over and over.
When I probe the CAN_H and CAN_L signals of the actual CAN Bus I can see data transmitting on the CAN_L, but the CAN_H just goes from high to low very briefly occasionally...it doesn't seem to have a steady stream of data on it. I am suspecting a couple things:
1) The motor is broken. I have tried to control it using a known good controller and it doesn't work either
2) Because the motor is broken (if that is the case) the CAN bus is turning it self off when it cuts out due to not getting a proper ACK back from the motor when it sends it a message.
Does this sound reasonable? Is there anything else I should be doing? I've attached my main.c and main.h files for reference...these are the only places I'm doing anything with the CAN Bus.
2023-07-11 06:10 AM
Hello,
You have an issue with your system clock config:
--> RCC_OscInitStruct.OscillatorType = RCC_OSCILLATORTYPE_HSI|RCC_OSCILLATORTYPE_HSI48; ??? HSI or HSI48?
Also for CAN communication it's NOT RECOMMENDED to use any of the internal source of Clock HSI, HSI48 etc.. This introduce some instability in the communication. So you need to use HSE and an external crystal by soldering X3 or in your case you can use MCO source generated by ST-Link module (which in turn is generated by X1 a 8MHz crystal), for this you need to solder SB16 and SB50 (please refer to the board schematic) and you need to configure HSE in bypass mode:
RCC_OscInitStruct.HSEState = RCC_HSE_BYPASS;
2023-07-11 06:22 AM - edited 2023-07-11 06:42 AM
I suggest this config for the system clock and CAN at 500kb/s
Please check the availability of the 8Mhz on SB50 with an oscilloscope.
2023-07-11 07:38 AM
You might want to start with the loopback mode to check if your transmissions are okay.
Even better, in case you have 2 boards you should try to set up communication between these, without the motor.
And have you checked all hardware connections, including the correct motor pins and the wiring? Maybe it's kust a broken cable.