While I have SPI working correctly on the MSP430FR5969, there are a few intervals that seem much larger than expected. The following logic analyzer capture is for 3 wire SPI with a software controlled CS pin. It is transmitting 0x4F, 0x00, and 0x25 (correctly). For this test, MISO is left floating. SMCLK = 1 MHz.
I believe the issue is two-fold. First, after the CS pin is toggled and UCA1TXBUF is written to, there is a very long delay before data transmission starts. I know that I am not using DMA, but should the delay really be this large? Second, the delay between bytes also seems unusually large.
// Init CS pin (active low) P4DIR |= BIT3; P4OUT |= BIT3; // Configure SPI pins P2SEL1 |= BIT4; // USCI_A1 operation P2SEL1 |= BIT5 + BIT6; // USCI_A1 operation // Configure USCI_A1 for SPI operation UCA1CTLW0 = UCSWRST; // **Put state machine in reset** UCA1CTLW0 |= UCMST + UCSYNC + UCMSB; // 3-pin, 8-bit SPI master, MSB UCA1CTLW0 |= UCSSEL__SMCLK; // SMCLK UCA1BR0 = 0x02; // /2 UCA1BR1 = 0; // UCA1MCTLW = 0; // No modulation UCA1CTLW0 &= ~UCSWRST; // **Initialize USCI state machine** UCA1IE |= UCRXIE; // Enable USCI_A1 RX interrupt
The above code initializes the eUSCI module.
*SPI_CS_PORT &= ~SPI_CS_PIN; while (!(UCA1IFG & UCTXIFG)); // USCI_A1 TX buffer ready? UCA1TXBUF = buffer_read(&SPI_tx_buffer);
The above code is used to start SPI transmission.
#pragma vector=USCI_A1_VECTOR
__interrupt void USCI_A1_ISR(void)
{
switch(__even_in_range(UCA1IV, USCI_SPI_UCTXIFG))
{
case USCI_NONE: break;
case USCI_SPI_UCRXIFG:
if (buffer_count(&SPI_tx_buffer)) // Check TX byte counter
{
UCA1TXBUF = buffer_read(&SPI_tx_buffer);
}
else
{
*SPI_CS_PORT |= SPI_CS_PIN;
SPI_busy = 0;
__bic_SR_register_on_exit(LPM0_bits); // Wake-up CPU
}
SPI_rx_byte = UCA1RXBUF;
UCA1IFG &= ~UCRXIFG;
break;
case USCI_SPI_UCTXIFG:
break;
default: break;
}
}
The above code is the interrupt routine.
The delay seems much longer than what it would take to execute the code in the service routines and transfer the data to and from the eUSCI buffer registers. Does anyone have any suggestions? Have I set up the eUSCI incorrectly? Even if it did take that long between bytes, why does it take so long for the first byte to be sent after writing to UCA1TXBUF?
Thank you for your help!
