PWM + dither 1

[Deaks]

The principles of PWM are clear to me, but does anyone have any pictorial representation of what dither looks like when superimposed on the PWM signal. We use these principles in the drivers for solenoids on hydraulic proportional valves.
Please direct me to a more appropriate forum if needed.

 

[itsmoked]

This is a fine place for your question.
Unfortunately your question makes little sense.

What dither? And why do you expect any dither?

What kind of system are you speculating about? If there is some dither in the output what it would look like would depend on where the dither originates.

Perhaps if you would describe more of what you are doing we can give you a valid answer.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

There are two uses (to my knowledge) of dither and PWM.
The more common is to broaden the spectrum so that peaks in the EMI spectrum get lower and broader. That sometimes saves the bacon for those who didn't filter their outputs well enough.
The other use is to "vibrate" the mean value of a PWM DC signal in proportional valves and such devices so that stiction is overcome. In the latter case, the PWM duty cycle is varied with a low frequency.

Picture, see

http://www.gke.org/pub/files/PWM w dither 1.bmp

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[roydm]

Dither used to be quite common in instrumentation, chart recorders and such like as Skogs says to overcome hysteresis with a small step change at low frequency.
Roy

 

[zappedagain]

Scroll up and down and you'll see what it looks like:

      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|
      _____
_____|
     ______
____|
      _____
_____|
       ____
______|

Normally you wouldn't want to dither at a constant frequency like this, but it shows the effect.

John D

 

[itsmoked]

Keith Cress
kcress -

http://www.flaminsystems.com

 

[zeusfaber]

In most applications I've seen, you shouldn't need to bother too much about what the dither looks like on the PWM.

Your proportional valve demand is likely to exist in some other form - either digital or analog - upstream of the pulse width modulator, and to my mind, it's easier to add a simple dither signal to that rather than trying to do something clever to the pulse train (especially if you're using a pair of solenoids to drive a directional valve).

A.

 

[MacGyverS2000]

Johm,

Would you mind reposting that with a frequency of 30Hz? I'm working on a school project that requires it...













Dan - Owner

http://www.Hi-TecDesigns.com

 

[zappedagain]

scroll faster!

 

[MacGyverS2000]

scroll faster!

<chuckle> I tried, but my finger on the scroll wheel was adding in too much dither...

Dan - Owner

http://www.Hi-TecDesigns.com