I'm using TIM6-DAC-DMAMUX-DMA to create a periodic signal as shown in the following picture. The yellow curve is DAC output, and green curve is a toggled GPIO for debug. I catched an unexpected interrption time point which is described as followed:

[Jerry1]

Registers Setting:

DMA_S0CR: MSIZE=PSIZE=16bit, DBM=1, MINC=1, CIRC=1, DIR=M2P, TCIE=1, EN=1

DMA_S0NDTR = 8

DMA_S0M0AR points to a buffer: 0x8000, 0x6000, 0x4000, 0x2000, 0x0000, 0x2000, 0x4000, 0x6000

DMA_S0M1AR points to a buffer: 0x7000, 0x5000, 0x3000, 0x1000, 0x0000, 0x1000, 0x3000, 0x5000

When TCIF0 interruption occurs, I toggle the green GPIO to indicate transfer completed.

Trigger: TIM6 generates TRGO when EN and UPDATE.

I got an interrupt at time point 1(red arrow), Is this the time when DMA finished reading from memory? why doesn't it occur at the next trig ? (There are only 7 trigs from begining, not 8).

But, the important qestion is, how can I got the interrupt at time point 2 ?, because I think this time point is a right time to stop the DAC output when a full periodic signal has finished.

[waclawek.jan]

Read the DAC chapter - DMA writes to DHR register when it's empty, and it gets transferred to DOR register by the trigger (coming from timer).

At 1, what happened was, that a trigger arrived from timer, 0x4000 got transferred from DHR to DOR thus the output pin's level changed appropriately, DOR got empty, subsequenly DMA tranferred 0x6000 into DAC_DHR (where it remained until the next trigger), and as that was the 8th transfer, it threw the TC interrupt.

> But, the important qestion is, how can I got the interrupt at time point 2 ?

One way to achieve that would be to count to 8+1 (and subsequently to 8) in an interrupt triggered by the same timer which triggers the DAC.

JW