Announcement

Collapse
No announcement yet.

Almost off topic electronics question!

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Almost off topic electronics question!

    Hi, I have built this motor speed controller to control a PMDC motor on the power feed I made for my mill, it works very well driving a geared motor from an old rechargeable drill..





    Unfortunately I have already built my power feed system using a PMDC motor from an old main frame computer tape drive system. The motor is much higher voltage and runs without apparent stress from a 72volt supply. I would appreciate comments on this modification to the drive system to handle the higher voltage.



    The modifications are marked in the red squares.

    Thanks.

  • #2
    The only difference to the control circuit is the feedback resistance. I would have suggested making it three times larger for the motor voltage which is roughly three times higher. In that case it would be a 56k resistor.

    There may be some interference to circuit operation from switching the higher voltage, so you might want to add a small value of cap across the 4.7k resistor just to the left of this feedback resistor. I'm thinking something like .01, or even .0047 uf. It might affect circuit operation as well, so it would be something to experiment with.

    The other concern I have is with the mos output transistors. I haven't looked up the specs on those, but I notice they are unchanged in the higher motor voltage circuit. I would think they should have a voltage rating of at least 200 in order to not be stressed. I've seen circuits where a 50 or 60 volt rated part is used in a 12v motor supply, but where the voltage is higher, a different part number is used.

    This 24 volt cordless drill I just got has a current pass device rated at 60 volts, but I think that's just on the edge of being too low a voltage rating. You do want the circuit to be reliable.

    Other than that, I think that's a good circuit to use. It has the benefit of the charge/discharge transistors driving the mosfet gates, which should make them run cooler.
    I seldom do anything within the scope of logical reason and calculated cost/benefit, etc- I'm following my passion-

    Comment


    • #3
      Agree with the previous. Those are 60V mosfets, you need a much higher voltage.

      I also highly applaud the use of an SMPS chip as a pwm driver. I suggest more folks may want to do that when rolling their own.

      The SMPS chip has an on-board reasonably linear ramp oscillator, and many of the simpler types are capable of running at fairly low frequencies where many of the more obnoxious problems don't occur.

      They are made to respond to a variable voltage input, so are quite capable of taking input from a pot, or even a voltage from a micro controller, etc.

      They usually have an on-board reference voltage available for purposes such as powering the pot.

      It's possible to apply stabilizing feedback, as the chip has an on-board op-amp.

      Pulse-by-pulse current limiting, which is usually missing from simple PWM setups, is often provided on-board the chip (not with the 494, apparently, I have never used the TL494, although I knew "of" it, and didn't know much about them). However, it could be added externally.

      The SMPS chip often has a pretty capable gate driver output, which may not require an external booster, depending on outboard power part and the frequency.

      Lots of advantages, relatively few problems, since this is essentially what they are made to do. Some cannot be made to run at low frequencies without troublesome oscillator external part values, that is one of the few issues with the SMPS chip as a simple motor drive PWM source.

      You may be able to avoid such small resistances on the gates...... At what appears to be only 3 or 4 kHz, and a gate charge of maybe 80nC per each, you should be able to go higher, to 22 or more, which reduces electronic noise, at the cost of a little extra dissipation.

      Higher voltage parts will likely have similarly low gate charge requirements, so that should be OK, given the low frequency.

      I am not sure that C3 is needed, it will draw a goodly amount of current at turn-on, and may just amount to an un-needed loss.
      Last edited by J Tiers; 05-03-2012, 11:21 PM.
      1601

      Keep eye on ball.
      Hashim Khan

      Comment


      • #4
        Thanks for the comments and advice. I have an IRFP450 500V 14A MOSFET on hand so I will be swapping that into the circuit!

        Comment


        • #5
          I think C3 is there to limit the amount of noise from the brushes. Maybe you can lower it to 10n to reduce current draw but even at this 220n there shouldn't be that much. (60mA at 72V/4kHz now)
          When shopping for new mosfets, try to get some that have protection features inside. (like a zener to protect the gate from spikes at the source pin)

          Igor

          Comment


          • #6
            Thanks Igor.

            Comment


            • #7
              One other thing - I'm assuming the 72V supply is unregulated...?

              As the revised circuit stands it regulates the voltage at the junction between drivers and motor relative to ground* - i.e. if the supply droops by say 5V under load then the voltage across the motor also drops by 5V.

              In the original circuit it appears that droop is compensated for by using the supply voltage to provide the error amp reference (at A1+) via the 20k and 2k2 divider.

              I suspect that for the modified circuit this divider should be connected to the 72V supply, with values adjusted to suit.

              Cheers

              .
              * edit: Not quite true - actually relative to the low voltage supply...
              Last edited by Barrington; 05-04-2012, 08:25 AM.

              Comment


              • #8
                Another point: D2 is not named.
                It should be a high-speed type, preferably ultrafast, and it should be rated for the full motor current.

                Benta.

                Comment


                • #9
                  Originally posted by ikdor
                  I think C3 is there to limit the amount of noise from the brushes. Maybe you can lower it to 10n to reduce current draw but even at this 220n there shouldn't be that much. (60mA at 72V/4kHz now)


                  Igor
                  That isn't quite right................

                  What will happen is that as the mosfets turn on, which they do rather fast with the low gate resistance, the output node switches rapidly from the power supply voltage +diode drop (73V) to near ground.

                  The capacitor must be charged (or discharged for other polarity) in order to have its voltage changed. This is in addition to any motor current (from continuous conduction), and must be supplied by the transistors.

                  If this switching transition takes a full microsecond, there will be a transfer of charge amounting to (0.22e^-6*73)/1e^-6, meaning a pulse of about 16A in the capacitor. If switching is faster, the pulse current is higher.

                  My suspicion is that the noise from the pulse currents is competitive with any noise from the motor, and that heavy filtering for the motor noise is best located in the DC supply. There may also be resonance of capacitor with motor or lead inductance.

                  A considerably smaller capacitor, preferably with a damping resistor, or even no capacitor at all, is suggested.

                  The comment on reference divider seems to be correct. I hadn't looked at that.
                  1601

                  Keep eye on ball.
                  Hashim Khan

                  Comment


                  • #10
                    Just a quick question, but how does the TL494 chip or those small transistor handle the base current needed, as there is no current limiting resistors between the voltage source and the chip OR between the output and the transistor bases?
                    Amount of experience is in direct proportion to the value of broken equipment.

                    Comment


                    • #11
                      The IC outputs are set up as emitter follower configuration with 2.2k res, is that what you you meant?
                      Max.

                      Comment


                      • #12
                        Having now had a look at the datasheet I would suggest the following:-

                        In the original circuit the feedback divider from 'MOTOR -' is (4.7/(18+4.7)) = ~20%. This is then divided by 2 by the two 47k resistors to ~10% at the A1- input.

                        The resistive divider on the supply (i.e. 'MOTOR +') to A1+ is also ~10%. (This is not a coincidence...)

                        As a first approximation the voltage across the motor will be 5x the voltage from the control pot (0-5v), i.e. 25v max.

                        So - a nominal 0-75v range requires a gain of 15x.

                        To achieve this the feedback divider from the 'MOTOR -' should be v/15 and the resistive divider on the supply ('MOTOR +') v/30.

                        Suitable resistor values might be 51k/3k6 and 47k/1k6 respectively.

                        Cheers

                        .

                        Comment


                        • #13
                          If you do as suggested here, be absolutely sure that both power supplies have the SAME ground. With a heavy gauge ground conductor between them.


                          Originally posted by Barrington
                          One other thing - I'm assuming the 72V supply is unregulated...?

                          As the revised circuit stands it regulates the voltage at the junction between drivers and motor relative to ground* - i.e. if the supply droops by say 5V under load then the voltage across the motor also drops by 5V.

                          In the original circuit it appears that droop is compensated for by using the supply voltage to provide the error amp reference (at A1+) via the 20k and 2k2 divider.

                          I suspect that for the modified circuit this divider should be connected to the 72V supply, with values adjusted to suit.

                          Cheers

                          .
                          * edit: Not quite true - actually relative to the low voltage supply...
                          Paul A.

                          Make it fit.
                          You can't win and there is a penalty for trying!

                          Comment


                          • #14
                            Originally posted by MaxHeadRoom
                            The IC outputs are set up as emitter follower configuration with 2.2k res, is that what you you meant?
                            Max.
                            Almost - I'm wondering how that Q1 base is handling the use, as there is no current limiting resistor on it. When TL494 output is high, the supply voltage goes to the base of Q1, makes it conduct and the resulting collector current plus the base current charges the FETs gate through that 4.7 ohm resistor. But, there is nothing limiting the above base current, so how will the transistor Q1 or the TL494 output handle this?
                            Amount of experience is in direct proportion to the value of broken equipment.

                            Comment


                            • #15
                              Originally posted by Barrington
                              One other thing - I'm assuming the 72V supply is unregulated...?
                              Unregulated.

                              As the revised circuit stands it regulates the voltage at the junction between drivers and motor relative to ground* - i.e. if the supply droops by say 5V under load then the voltage across the motor also drops by 5V.

                              In the original circuit it appears that droop is compensated for by using the supply voltage to provide the error amp reference (at A1+) via the 20k and 2k2 divider.



                              I suspect that for the modified circuit this divider should be connected to the 72V supply, with values adjusted to suit.

                              Cheers

                              .
                              * edit: Not quite true - actually relative to the low voltage supply...
                              Thanks, I cant claim to understand the operation of this circuit but that certainly makes part of it clearer, obviously I need to move both reference voltage divider chains to the 72V supply.
                              Last edited by The Artful Bodger; 05-04-2012, 05:34 PM.

                              Comment

                              Working...
                              X