Hello,

I'm trying to analyze an incoming PWM signal on DQa.3 of a S7-1214C DC/DC/DC. The signal comes from a Teknik clearpath motor and represents the actual speed, or actual torque of the motor. 

I configured HSC_1 to Period, single phase. And use the first example of this Siemens example.

It works well, but it seems that the HSC calculates the time from rising to rising edge. The teknik motor feedbacks on a steady 45Hz. And that's exactly what I see. But what I would like to know is the time the HSC calculates from rising to falling edge, in that way I could determine the speed/torque of the motor.

It would be greatly appreciated if someone could give me some pointers :-)

Thanks,

Rob. 

A frequency of 45Hz sounds slow. What is the smallest time increment that you wish to "see"?

Three possible ideas come to mind:

1.
Put the PWM signal through a low pass filter circuit and read the resulting analog signal.

2.
Use hardware interrupts to capture the +edge and -edge of the PWM. Read a timer or real time clock or PLC clock pulse to determine the duration of the on/off pulse.

3.
Use a time interrupt. Each interrupt the input level is checked and a counter is increased. One counter for +level and another for -level.

Hi dP,

Thanks for your response. 

45Hz is rather slow. it gives a pwm period of 22.22ms. I would like to 'see' 1/100th of that signal, roughly 0.2ms. I think that's too fast to be able to analyze with OB1. Introducing hardware interrupts on the rising and falling edge(your option 2) is actually a very good idea and absolutely worth trying. 

Option 1 and 3 are also good ideas. 

I'll look into it tomorrow(time to go home over here ;-)

Rob. 

 Hi Hubby, 

hubby_d

Along with what #dP already suggested, if I understand correctly, you have a single pulse every 22.2222msec with the pulse width varying as motor speed / torque vary.

You want to determine how wide each pulse is within that 22.222msec pulse train.

This is correct, and exactly what I would like to do.

unknown user

Do you know the min and max pulse widths?  Obviously they _can_ be 0 and 100%, but that is difficult to tell isn't an error or system failure.  But it would make sense they limit the range to say 10 and 90% where 10% is 0% speed and 90% pulse width = 100% speed.

 Yes. The minimum pulsewidth is 5%(but closer to 0% in reality) and the maximum pulsewidth is 95%.

Today is the first day I could work on the project again.  I tried the method to put two seperate hardware interrupts on the rising and falling edge of the signal. I tried to operate a timer, but it seems impossible on a s7-1200 to get a resolution higher than 1 milliseconds. Also the cyclic interrupt can be no faster than 1 millisecond

HSC can do it, but how can I use the HSC to measure the pulsewidth? In earlier attempts I could only get the time in between two rising edges on the signal. Any suggestions?

(In addition, how can I use the HSC without a hardware input?)