Showing results for 
Search instead for 
Did you mean: 

STM32F4 Timer DMA requests

Associate II

Hello Team,

I am using the STM32F469NIH6 to implement a timer triggered DMA and I need some clarity on a few points here-

I see that that there are three possible requests possible for a timer triggered DMA. I understand that the TIM8_UP is to trigger a DMA request with an update event (which occurs when the counter overflows/underflows)


What is the TIM8_CH4/TRIG/COM all about? How is it different from the TIM8_CH1 request?  Any leads on this would be of help !



Thank you,

Janani Sunil


> TIM8_CH1 request

That's triggered by a Capture or Compare event on TIM8_CH1, if TIM8_DIER.CC1DE is set.


That's triggered by one of the Capture/Compare on TIM8_CH4, or the Trigger event (i.e. edge of TRGI signal as selected by TIM8_SMCR.TS), or the Commutation event (that is a rarely used event I am not going to discuss here); again, if respective DMA-enable bit is set in TIM8_DIER.


Hi @waclawek.jan ,

Thank you for your response. 

I am trying to operate timer 8 in ETR Mode (in slave trigger mode) and I have enabled TIM8_CH1 DMA request for the same (Master TRGO OC mode). I am supplying an external PWM signal to the respective ETR pin allocated for timer 8.

Upon configuring this, when I probe the ETR pin, I see that my master signal that I apply is getting corrupted. 

Am I missing out on any configuration here?


Thank you,


> when I probe the ETR pin

Which pin is that?

> I see that my master signal that I apply is getting corrupted.

What hardware is this, your own, or some "known good" board like Nucleo or Disco?

Read out and check/post content of timer and relevant GPIO registers.



PA0 is the ETR pin that I am using for TIM8_CH1

I am using the SDP-K1, which has an STM32F469NIH6 mounted on it. 

The timers and the GPIOs seem to be configured right. 


Thank you again for your inputs.




> The timers and the GPIOs seem to be configured right.

Yet you complain about a timer-related pin working strangely.

How else do you propose to debug it? The mcu works out of its registers, not out of source code or CubeMX clicks.