Part Number: MSPM33C321A
Other Parts Discussed in Thread: SYSCONFIG
I am using the LP-MSPM33C321A and I have used the example project hsadc_max_freq_dma as a starting point for my development and have configured my the clocks exactly the same as the example project does which uses errata workaround #1 for the SYSPLL that may lock on to the wrong frequency. I am using the DMA0 and timerA with the exact same configuration that CCS uses in the example except that I am using transfer mode: DL_DMA_FULL_CH_REPEAT_SINGLE_TRANSFER_MODE
To give more context, I have set MCLK to 160MHz and use PA27 to take in 1us pulses from a signal generator at 10kHz. However, I am only able to sample at ~2.5MSPS instead of the advertised 9.43MSPS and I know this by toggling pin PA22 with an ISR. I am using the ping pong buffer feature, so I am using early interrupts when one half of a buffer is full and an interrupt when the buffer finishes. My ADC sample count is 2048 samples/half and my frequency from PA22 is 1.22 kHz(or 1,22y halves/second) which leaves me at ~2.51MSPS. If MCLK is set 160MHz and I use TIMERA0_0 to trigger SOC with a TIMER_LOAD_VALUE of 16, should I not be seeing 9.54MSPS from my calculations? It is as if my results are being divided by 4.



