SPRUJ53B April 2024 – September 2024 TMS320F28P550SJ , TMS320F28P559SJ-Q1
The receiver receives data on every clock edge. While there are specific patterns that determine the a start of a frame, and denote the end of a frame, these patterns are able to occur at any point during normal operation inside of the frame. If there ever is a point at which the receiver fails to detect a successful frame, the module must be reset to make sure that subsequent frames are received properly.
When any of the following errors occur in a received frame, the receiver can be required to be reset and resynchronized with the transmitter:
The receiver core status (RX_VIS_1.RX_CORE_STS) can be monitored to determine if the receiver core has entered into an error state requiring a soft reset to resume communication. Incorrect frame type and end of frame errors always cause this bit to become set. A soft reset is required in these cases. A frame watchdog timeout always requires a reset due to the fact that the receiver state machine is still expecting more information when the watchdog timed out. RX_CORE_STS can be used to determine if a noise event was the cause of the failed frame. The ping frame watchdog also does not cause RX_CORE_STS to be set. Similar to the frame watchdog, a corrupt receiver may not be the reason for the ping frame to have timed out. The transmitter could have gone offline and never sent a ping frame. Alternately, during idle time, a noise event could have occurred, thereby putting the receiver into a corrupt state. As the receiver is able to detect this during the ping frame watchdog timeout interrupt handler, this type of event is not lost and the application can act appropriately.
As the receiver is clocked by RXCLK, not SYSCLK, a noisy clock or data line can cause some internal design constraints to be violated, putting the receiver core logic into undefined states. Make sure that the clock and data lines satisfy the Electrical Characteristics and timing requirements of the FSI module found in the device data sheet. Failure to do so can cause the receiver state machine to go into an unrecoverable error state. The receiver can only be recovered by undergoing a soft reset. To determine the state of the receiver core after an unexpected frame error, the application must check the receiver core status bit.
In addition to the above errors, buffer overrun or underrun can warrant a soft reset to resynchronize with the local application software. Refer to Section 25.3.3.8 for more information on the receive buffers. The requirement of resetting the receiver due to overrun or underrun is up to the application.
After the receiver has been placed into soft reset, the application must notify the other device's transmitter to begin a new synchronization phase. The simplest way to achieve this is through a ping or error frame sent with a designated tag. If the application is not using the FSITX on the device with the detected error, some other method must be established. The other device must stop transmitting and begin a new synchronization phase.