Change PWM Duty Cycle

D

drkidd22

Member

• Jan 21, 2012

• #1

Hello,

I have a 60hz PWM signal which can change from two stages when a switch is fliped.
It changes from 5% duty cycle to 12.5% duty.

What I want to do is create a circuit that take the PWM as an input and that at 5% duty cycle my voltage is 0V and when at 12.5% duty cycle I want the full 5Vdc out. So basically make the 5% duty cycle a 0% duty cycle and the 12.5% duty cycle a 100% duty cycle.

I was thinking of using a differential op-amp maybe, but want to know if anyone here has any other suggestions.

 

jpanhalt

Well-Known Member

Most Helpful Member

• Jan 21, 2012

• #2

How about a low- pass filter to integrate the PWM, then a comparator to switch to 100% when the critical voltage is passed.

John

 

Last edited: Jan 21, 2012

D

drkidd22

Member

• Jan 22, 2012

• #3

Ok so I did a quick simulation in multisim and created the low pass filter with cut-off at 62Hz and using the lm311 comparator.
The only problem is that I'm not yet able to switch the output as I described before.

When V3Max is the + input to LM311, the voltage is 355mV and when V4Low is the + input to LM311 the voltage is 200mV.
I'm close to getting this working, but the output from LM311 doesn't witch from 0-5V output when I toggle the switch.

 

Attachments

• freq.png

  16.3 KB · Views: 212

Last edited: Jan 22, 2012

alec_t

Well-Known Member

Most Helpful Member

• Jan 22, 2012

• #4

You could just use a Schmitt input gate as the threshold detector:-

Edit: Add another inverter to get the output polarity you specified.

 

Last edited: Jan 22, 2012

ronv

Well-Known Member

Most Helpful Member

• Jan 22, 2012

• #5

Try a cutoff frequency of about 3 to 6 Hz. Here is a nifty calculator so you can see the response.

**broken link removed**

 

D

drkidd22

Member

• Jan 22, 2012

• #6

alec_t said:

You could just use a Schmitt input gate as the threshold detector:-

Edit: Add another inverter to get the output polarity you specified.

Click to expand...

Yeah that's what I want to do, specially with the Schmitt gate, but the requirements I have are a little different. I updated my circuit and just have to figure out the Schmitt thing out so that it trigger only as in your image but with my settings.

 

Attachments

• freq.png

  14.2 KB · Views: 184

ronv

Well-Known Member

Most Helpful Member

• Jan 22, 2012

• #7

If I read your schematic correctly you only have 3.15 volts for the input signal. If that is true you may need to stay with the comparator. What you have now should work.

 

D

drkidd22

Member

• Jan 22, 2012

• #8

Yeah, What I have is only 3.15V. The circuit works ok in the sim, but there seems to be some jitter at the output for a few milliseconds before the output CMND complety changes stare. How can I get a smoother transition? I'm trying to get a smooth signal like alec_t posted on his wave form.

 

Last edited: Jan 22, 2012

alec_t

Well-Known Member

Most Helpful Member

• Jan 23, 2012

• #9

How can I get a smoother transition?

Click to expand...

Try adding some positive feedback (3.3 meg?) from pin 7 to pin 2 of the comparator.

 

alec_t

Well-Known Member

Most Helpful Member

• Jan 23, 2012

• #10

I simulated your circuit with added +ve feedback, but reliable jitter-free 2-way switching wasn't achievable. The jitter is due to the ripple on the waveform at the comparator input. The problem is that with simply smoothing the PWM signal (your circuit) instead of using a leaky rectifier (my circuit) the difference between the average voltage for the 5% and 12.5% PWM ratios is so small; less than the hysteresis provided by the feedback.

 

ronv

Well-Known Member

Most Helpful Member

• Jan 23, 2012

• #11

PWM Filter

The hysterisis needs to be larger than the ripple. So you need a lower frequency filter and more hysterisis. Try these values.

 

D

drkidd22

Member

• Jan 23, 2012

• #12

Interesting, I will give that a try tonight when I get home.

 

D

drkidd22

Member

• Jan 23, 2012

• #13

That helped a little, will do some more testeing at work tomorrow with the scioe and stuff.

 

Last edited: Jan 23, 2012

D

drkidd22

Member

• Jan 24, 2012

• #14

ronv said:

The hysterisis needs to be larger than the ripple. So you need a lower frequency filter and more hysterisis. Try these values.

Click to expand...

Thanks ronv, that did the trick. I'm not familiar with comparators, but this really helped be out. I found an LM393 and built the circuit. I will play around with inverting and non-inverting. You might have seen my other post need-help-switching-signals
it's related to this one.

I've attached an image of the results from my actual circuit.
I also added an image with feedback resistor removed.

 

Attachments

• 20120125_775207.png

  28.3 KB · Views: 160
• 20120125_780618.png

  33.4 KB · Views: 168

Last edited: Jan 24, 2012

ronv

Well-Known Member

Most Helpful Member

• Jan 24, 2012

• #15

Cool. Is the long time constant ok for your purpose? It takes almost 1/2 second to make the step.
Is this for a plane or something else?

 

D

drkidd22

Member

• Jan 24, 2012

• #16

It's actually around where I need it, might need to play around with it when I'm a little more advanced on the project and during testing.
It for some crazy toy I'm working on. I got this RC car and an rc Airplane that I want to combine into one ;-). So it's help switching signals for air mode or plane mode using one of the toggle switchies of the transmitter.