Pulse counter with MK60DN512 and MQX 3.7

[feraro]

Hello.
I am working with micro K60DN512, CodeWarrior 10.3 and MQX 3.7. Need to measure a time interval of 2 msec around 3500 pulses. I've tried using dual-edge FlexTimer, but does not give me the accuracy I need. Could anyone advise an alternative?.
Thank you very much.

[paulmartin]

Hi,

Could you specify a bit more what you actually need to achieve.

Are the pulses uniform/ constant frequency or random? What accuracy do you need? Do you need a continuous measurement, like a pulse rate  measurement every 2 msec or is it a single measurement?

one alternative for example is the clock input of the FTM counter. It counts up with every pulse of the input up to I think a quarter of the bus clock.

second alternative: the quadrature decoder mode of the FTM can count up and down.

Martin

[feraro]

Hi Martin.

Thank you very much for your answer.
Let me explain better. Every 10 msec, for 2 msec, pulses reach me, at a constant frequency which can be between 1.4 and 1.7 MHz. My purpose is to measure the frequency. I tried dual-edge mode but I have the problem that a system clock frequency of 50 MHz, the values ​​that are read between 28 and 35 respectively. And this for my application is very low resolution. Would need to get the frequency in MHz with two decimals.
I need something to count the pulses in an interval of time without working the mic, so that I can check then.

Thanks again.

Fernando.

[paulmartin]

Counting pulses is not the problem. The problem will be fixing the interval and synchronizing with the 10msec cycle.

If you know that the pulses are send for exactly 2 msec this is a possible solution:

Use the clock input FTM_CLKIN0/1 ( if you dont use EXTAL/XTAL pins ) or the quadrature decoder mode ( multiple options here: any pin with FTMX_QD_PHA ).

Both allow you to count pulses.

Then your measurement cycle could be like this.

-reset the counter to 0

-wait 10msec ( time for other computations )

-read the counter and compute frequency

(accuracy should be fine since you have about 3000 pulses)

If the pulses are sent for a variable amount of time, you would need more intervention from the processor.

- trigger an interrupt on the first pulse and then reset the counter to 0

- wait for as long as you are guaranteed to have pulses

- read the counter and compute frequency

or

- set the modulo of the counter to lets say 2500 and trigger another interrupt on this overflow event. But then you need a second FTM module to know how much time has passed for the 2500 pulses.

( this also suffers from the possibility that another interrupt may cause a variable delay between the overflow and your code to read the time)

Does your application have some kind of synchronization mechanism, like a pulse on a separate line 1 msec before the next set of pulses arrive. Or are the pulses somehow triggered by the MCU?