Distortion of pwm signals

Discussion in 'General Electronics Chat' started by yassser, Dec 9, 2012.

  1. yassser

    Thread Starter Member

    Jul 25, 2011
    91
    0
    Hello ,

    I'm making a dc motor control system , I have a control ct. and a power ct (H-bridge configuration) , the PWM control signals need to be transferred a long way to the power ct. in a noisy environment , when I used the oscilloscope to view the signals , I found the signals to be very distorted as shown in the figure when it reached the power ct. which caused fatal results .
    [​IMG]

    I understand this is because the inductance of the cable , but I can't understand it mathematically and physically , can someone please illustrate it for me , and is there some way to prevent this kind of distortion ?

    note: i use a shielded cable to transfer the signal
     
  2. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    I need to see a drawing of the driver circuitry to figure out why you aren't getting a fast enough move in the positive direction.

    Schematic, please.
     
    Last edited: Dec 9, 2012
  3. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
  4. yassser

    Thread Starter Member

    Jul 25, 2011
    91
    0
    [​IMG]this is the drive ct.
     
  5. yassser

    Thread Starter Member

    Jul 25, 2011
    91
    0
    the frequency is 10khz , the distance is about 10 meters
     
  6. nigelwright7557

    Senior Member

    May 10, 2008
    487
    71
    You will need to reduce the 10k resistor or find a push/pull driver circuit.
     
  7. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    That's an open collector driver...it has no positive drive. You must provide more pull up current with a resistor.
     
    Last edited: Dec 9, 2012
  8. yassser

    Thread Starter Member

    Jul 25, 2011
    91
    0
    i do have a 10K pull-up resistor , may be I need to reduce it .

    I checked the max232 line driver , the problem is that its delay is 500ns , these pwm signals have to be synchronized and these high delay times isn't acceptable , I could use a transistor to level shift the signals to high voltage to reduce noise effect and use a smaller pull up resistance , would that work?
     
  9. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    I see the 10k pull up resistor. It is too much resistance.
    This driver can sink as much as 50 ma. The smallest possible resistance is 100 ohms. Try something around 1k ohms.
     
  10. yassser

    Thread Starter Member

    Jul 25, 2011
    91
    0
    okay , i will .

    but can you please explain to me why does the signal look like if a cap is being charged , why does it have this shape ? and would reducing the resistance reduce the time constant ? i mean what is the mathematical reason ?
     
  11. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    The shielded cable is a capacitor. It has picofarads per meter. You can treat it as if the driver only sees a capacitor.
     
  12. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    The 10k resistance and the high capacitance of the long cable slow down the risetime.
     
  13. crutschow

    Expert

    Mar 14, 2008
    12,977
    3,220
    The signal looks like a cap being charged because it is. Look up the capacitance of the shielded cable. It's likely somewhere around 50pF/ft. For 10 meters that would be 1.6nF. (You can also measure it with a capacitance meter). So you need to reduce the driver impedance to charge that capacitance and provide the RC risetime you want. Remember that the signal rises to 63% of its final value in one RC time-constant and will go to 98% of its final value in 4 time-constants.
     
Loading...