Hi Kerry!
Thanks for the reply!
Coming to you queries :
Do you test the FlexCAN based on the RT1170 SDK, use the interrupt_transfer project or your own write code?
No, we are using our own custom board and not the SDK. Yes, we have used the interrupt transfer project code itself that has been provided and modified it to receive 5 frames in 5 different message buffers using unique identifier and filtering for each of the frames.
Do you test this code?
Yes, all my statements/observations are based on testing the code.
If yes, do you modify it or not? Whether you try one RT1175 board as on CAN node, and other CAN node use the PC side + CAN testing tool, then send 5 frame to the RT board continuously?
Yes, we modified the interrupt_transfer code as explained earlier. Both the nodes are custom nodes wherein the master node is RT1175 and slave node is a TI DSP controller. I have used Texas instruments' RM57 (uses same message buffers concept as provided by RT1175), TMS570 and NXP's 9s12XDP512 controllers and they all seem to receive the frames properly without any delay.
In the case of RT1175, whenever the Slave node is adding some delay (say 1ms) in between the 5 frames, I am able to receive all the frames whereas when there's no delay, I only receive the 1st frame continuously! This does not happen for other TI or power PC controllers.
So is the delay mandatory in the case of RT1175 in between the frames in order to receive all frames correctly? I don't think delay should be an issue because other controllers working at very low clock speeds didn't require so.
What do you suggest next?
Regards,
Chetan.