PWM clock count

Georg Tornqvist said:

...generate a PWM frequency of 10.0 KHz, 50% duty ... unable to count the transitions of the PWM squarewave.

Would it not prove quick/simple to route that PWM output signal into one of the (many) ARM counter inputs - and have that counter do the "dirty work?"   Such is very much "in the background" - should be almost, "automatic."

You "cloud" the issue w/4 channels - (violation of KISS) - why not get a single channel to perform to your satisfaction - and once that's achieved (as/if needed) we can extend the capability...   Doing "too much" ... "too soon"  most always proves NOT in posters' best interest...

Hi cb1,

now i'm working only with one channel, i think i have not explained well, i wrote of four channel as a future improvement. Now i'm working to your idea, to let a counter do the job, and stop the pwm when the time has gone.

Thanks for your answer, regards

G.

It would appear that "one here" has qualified for a, "non self-awarded" Verified Answer. (as featured in Robert's sticky post entry, yesterday...)

(strange forum action here - this post was aimed at Robert - not George - yet forum improperly cites "response target"...)

@Georg  (that was my grandfather's name)

Georg Tornqvist said:

how to configure the timer in capture mode to trigger an interrupt

May I note that your request - quoted above - rises above the goal of "KISS?"   Kindly look here:

The highlighted "Edge Count" seems best for your objective - don't you agree?   You ask about adding the complexity of interrupts - yet might it make more sense to feed your (selected) Edge Count configured, Timer input - with a low frequency test signal, initially?   I'd suspect that the timer would perform as a counter - and not require the complexity & demands enforced by interrupts - at least in your early, investigative stage.

You may further simplify your study by using a free MCU GPIO to provide a "known" number of pulses into your timer - and then simply reading the appropriate timer register - to confirm that the count is proceeding as expected.

Once you've mastered this (simpler) method - we can continue into interrupt - but too often the (early) introduction of interrupts confounds - we find it far more productive to follow KISS - one small battle - yielding one small solution - at a time...   Are we in agreement?

Georg Tornqvist said:

...output pin with a jumper with the input pin of the counter, or there is a better/simple way to do it?  

Mon Ami - could there (possibly) be a "simpler/quicker" means?   Firm/my limited ability/skill lures us (always) to KISS.

Do read again - the "Generator as input signal" (to the Counter/Timer) is too complex!   (Do NOT add complexity - certainly NOT NOW!)   Instead - as past written - employ one of the MCU's (ideally nearby) free GPIO pins - and program it to produce 5-10 low frequency, output bit toggles.   Dare I state that, "nothing" could be as simple - or helpful - in leading you to the end solution?  (the "generator" adds complexity - "eats our time/effort" - your attempt to "move so forcefully/directly" will delay & confound - not speed/ease!   (Not to ask, "How I know?")

One wonders, "How/why" the jumpered interconnect between GPIO output pin - and Timer input pin - can (ever) be judged (anything) other than best/simplest?

Isn't the bigger question - why would one (want) to expend that additional time/effort?

We've achieved your stated, "PWM Clock Count" (Forum title) objective - have we not?   (indeed we have - yet post is unmarked)

"Now" you (seem) to seek, "PWM Generator's (proper) awareness" of its output count - do you not?   Have you exhaustively read/reviewed all available MCU documentation & API software examples - before adding this task?   (such would appear a good "first step")

Possibly I can devise (some) means to "tease such awareness" into the "timer" (not the PWM Generator) when that timer is employed w/in Timer-PWM mode.   Yet - such seems "outside of normal/expected usage" - and may have limited utility.   (i.e. you may have to SELL your justification for this extra analysis...)

Georg Tornqvist said:

...not able to...reset the counter before inserting a new (match) value.

Pity that - yet shouldn't your time/effort be directed to solving that - rather than "forcing" the PWM Generator into an (unnatural) count recognition?

I haven't used this family MCU in ages - yet I recall some restrictions in both, "when & how" the "TimerMatchSet()" may be properly invoked.   (purely a "swag" - perhaps the Timer must be "disabled" prior to such setting?)   I do believe your issue IS covered - more likely w/in the MCU manual than API.  (although I'd exhaustively search both)

While inelegant - you could always "duplicate" the "code block" used to implement your initial match - which works.

Single stepping should pay huge dividends here - first in examining the Timer values pre & post "Successful" match.   And then after you've changed that match value.   In both cases - reverting to the GPIO pulse rather than the faster PWM Generator (as Timer/Counter input signal) - should, "speed, ease, enhance" your investigation.   And that is "pure" (uber effective) KISS...