24V PWM Control of N Mosfet from 24V PLC Output

Discussion in 'The Projects Forum' started by piashaw, May 26, 2014.

  1. piashaw

    Thread Starter New Member

    Apr 20, 2013
    12
    0
    Hi

    Trying to find out the "correct" way of running an N channel Mosfet from a 24V PLC Output (as most Mosfets have a max Vgs of 20V)

    As the PLC only has 1 24V 100mA output and I wish to PWM a motor at 33KHz with a 10 bit Duty Cycle, I was intending to use a TC4432 or equiv Mosfet driver as the 1.5A Peak Output Current will enable the Mosfet to switch fast enough.

    Problem is that Output Signal Voltage relates to Vdd which will be 24V (which also enables 24V input signal).

    Was thinking of using a Zener but it needs to cope with 1.5A, and I'm not certain which ones would be fast enough. Secondly it is often pointed out that with a Zener, the circuit might oscillate.

    So what is the PROPER way of dealing with this. 24V PLC outputs are common place and the PWM outputs on my PLC are internal outputs as opposed to on an external card.

    Thanks
     
    Last edited: May 26, 2014
  2. crutschow

    Expert

    Mar 14, 2008
    13,009
    3,233
    You just need to power the driver with a voltage less then the maximum allowed by the Vgs rating. You can use a linear regulator to drop the 24V to the desired voltage (say 15V).
     
  3. wayneh

    Expert

    Sep 9, 2010
    12,120
    3,046
  4. piashaw

    Thread Starter New Member

    Apr 20, 2013
    12
    0
    @wayneh Was trying to find a 30V Vgs but haven't succeeded so far. Only needs to output an eddy current brake with 24V Max 8Amp. Thanks for the suggestion. Do you think the 25V rating of the mosfet is "safe" enough for the 24V usage if under constant usage?

    @crutschow Had considered that but the datasheet states Input Voltage VDD + 0.3V to GND, so would you suggest a simple voltage divider for the 24V output from the PLC? Not certain if this would have any effect on the timing of the driver. (Have no experience of PWM at this speed with this range of duty cycle) and don't want to delay rise and fall times at low duty cycle.
     
  5. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,542
    2,369
    What PLC is it? as most PLC's are too slow due to the scan time to drive a PWM signal of any significant frequency, usually a dedicated special module is used?
    But in any event, why can't you use a restive voltage divider?
    Max.
     
    Last edited: May 26, 2014
  6. piashaw

    Thread Starter New Member

    Apr 20, 2013
    12
    0
    It's a Saia Burgess Unit. Their internal PWM outputs are up to to 32.xx KHz and I can select duty cycle of 0-1000. But only output 24V 100mA
    With the voltage divider, I read in a few places that it could have an effect on the on/ off timings of the fet due to the internal capacitance charge. Unfortunately the mosfet drivers I have seen have an output voltage relating to the Vdd as opposed to the Input voltage. I cannot see a problem using the linear regulator as proposed by crutschow and a voltage divider on the driver input, but want to make sure.
     
  7. piashaw

    Thread Starter New Member

    Apr 20, 2013
    12
    0
    But it still makes me wonder what the "industry" standard way of doing it is ,as 24V outputs are exceedingly common, and I assume they PWM drive Mosfets reasonably regularly.
     
  8. #12

    Expert

    Nov 30, 2010
    16,284
    6,797
    I think this is what crutschow said:
     
  9. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,542
    2,369
    Usually these are simple digital I/O.
    On all the PLC's I have worked with, you would use a dedicated module specific to the application.
    Is this a dedicated PWM module or a regular I/O?
    If dedicated module I would have expected it to be able to be conditioned for a particular output?
    Max.
     
  10. inwo

    Well-Known Member

    Nov 7, 2013
    2,433
    315
  11. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
    Last edited: May 26, 2014
  12. piashaw

    Thread Starter New Member

    Apr 20, 2013
    12
    0
    @#12 Yes that's what I understood, but by lowering the Vdd to 15V, I also need to lower the input signal to 15V as max input signal and output signal are both related to Vdd. That's why I am trying to work out if a divider would cause issues on the input side of a mosfet driver chip, not enabling it to switch off/ on fast enough.
    @inwo Basically yes. Here is the manual for the model I am using. http://www.sbc-support.com/uploads/tx_srcproducts/26-856_EN11_Manual_PCD2M5_CC-2014-01-07_dpi300.pdf
    @MaxHeadRoom pg 3-1 (pg14) states On the motherboard there are also 6 digital inputs (4 interrupt inputs or one encoder connection) and 2 outputs. The option to configure the inputs as interrupt or encoder inputs and the outputs as pulse-width-modulated (PWM) means that the PCD2.M5_can be used as a "low-cost solution" for machinery and systems.

    For this particular PLC, I have up to 32.226KHz Frequency with a duty Cycle of 0-1000


    @ronv Thanks, maybe thi is the wrong place to ask, but I am now confused as to what acceptable rise and fall times are. How do I calculate this eg on a 33Khz frequency with a .001 duty cycle?

    Thanks everybody for your input.
     
  13. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
    Switching loses can get a little complicated, but in general you can look at your FET data sheet to see what the NC value is. This will tell you the switching time of the FET with 1 amp constant current to charge and discharge the gate. Since you have a constant voltage not constant current the time is ~ twice as long. So if your FET has a gate charge of 100NC and you drive it with 24 volts thru 100 ohms to limit the current to 100ma peak the time would be about 2 usec for it to switch. This will make the power go up quite a bit as well as messing up your duty cycle. The power lost here is added to the power lost in the steady state load. In your case a driver is probably needed depending on the load and the FET used.
    You can use your driver at a lower voltage as it has a clamp diode on the input to keep the voltage from going above Vcc, but I think I would add another schottky to Vcc and about 1k in series with the input signal.
     
Loading...