DC motor control using PWM on rectified unfiltered current

Thread Starter

onlyonce18

Joined May 8, 2016
48
Hello everyone.

I´m trying to control the speed a DC Motor using a rectified unfiltered current and a PLC. The PLC has PWM capabilities, up to a 200kHz frequency. The DC motor was used in a winding machine, with speed controled by thyristor card with a potentiometer, until that card burned up, so I thought this is an opportunity to try and learn something new.

I have searched the internet but couldn't find how to do it or if it is the best approach, although found some references about using PWM on a sinusoidal current as being complex and difficult.

I read that thyristor drives are common DC motor controllers and that they work by changing the firing time of the thyristor allowing the sinusoid rectified current to pass through and reach the DC motor, attaining a very acceptable speed control - looks like PWM with a sinusoid signal instead a quadratic one.

For this project I will use an H-bridge to rectify 220 AC, the rectified current will have a frequency of 120 Hz (double the AC source current). Using a PLC to generate a pulse at 6 kHz (using this freq each half sine wave is divided in 50 parts, each of 167 microseconds) to activate a MOSFET gate to allow the rectified unfiltered DC to go through the MOSFET and reach the DC motor. Speed would be controlled by changing the pulse width.

I'm not an electronic engineer and have a very basic knowledge of electronics, so I'm trying to check the feasibility of this project.

So.... I need your comments and advice:

  • First of all, would this work? If not what can be done to make it work reasonably?
  • I though about using a filter after the H-Bridge, but since it is a regular DC motor I believe it is not necessary, and if I do I would need a big capacitor and a bigger heat sink for all components. Do I need to think more about this?
  • Is 6kHz enough? PLC can manage up to 200kHz and up/down and rise/fall times for MOSFETS are on the nanoseconds, so there is plenty of room.

DC Motor Characteristics (I couldn't find more details) 120V 1HP 1600 RPM. Max speed in real world use is 180 RPM
Energy source: 220V AC 60 Hz

I also have made a circuit, but I have some doubts:

  • Do i need to put something more to protect the Low side (specifically to protect the PLC) or a fuse is enough. Is there another way to arrange the components or another circuit to protect the PLC?
  • Do I need to put a Varistor or the flyback diode is enough?
  • Heat, at the flyback diode, rectifier and Power MOSFET, do I need to take special measures or a good heatsink is ok? Are there better options?
  • I though about using a potentiometers to find the right pull down resistors for both MOSFETS, can them be calculated so there is no need to make the tests?
  • Motor max voltage is 120 V and the PWM could reach 311V over some milliseconds, will it damage the motor?

I'm attaching a PDF file for the circuit and the datasheets of some components.

Thanks to everyone!
 

Attachments

MaxHeadRoom

Joined Jul 18, 2013
28,617
SCR bridge type controllers, KB etc, do not use or require a smoothed P.S. hence they run direct off of the 120v/240v AC supply, with analogue control
However it is usually necessary to use a bridge with smoothing capacitors when implementing PWM.
Max.
 

Bernard

Joined Aug 7, 2008
5,784
I do not think this circuit will pass AAC line operated equipment standards, & if 24 V supply is not derived from
the rectified 220 , then there will be a ground mismatch.
 

crutschow

Joined Mar 14, 2008
34,280
I see no fundamental problem with using PWM on a full-wave rectified unfiltered sinewave to control the motor, no filter needed.

6KHz is in the audible range so that is a consideration if such a noise from the motor would be annoying.

Isolation is likely required so you could use opto isolators to drive the MOSFET gates.
You would need a simple DC supply at the output of the optos to drive the MOSFET gates.

Heat depends upon the motor current, the ON resistance of the MOSFETs, and the voltage drop of the flyback diodes (Schottkys have a lower drop).

The MOSFET gate drive required is determined by the MOSFET gate capacitance and the PWM frequency.
You should be able to calculate that and it shouldn't require a pot.

Why are you running a 120V motor from a 220Vac supply?
Are you in a location where there is no 120V available?
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
I see no fundamental problem with using PWM on a full-wave rectified unfiltered sinewave to control the motor, no filter needed.

6KHz is in the audible range so that is a consideration if such a noise from the motor would be annoying.

Isolation is likely required so you could use opto isolators to drive the MOSFET gates.
You would need a simple DC supply at the output of the optos to drive the MOSFET gates.

Heat depends upon the motor current, the ON resistance of the MOSFETs, and the voltage drop of the flyback diodes (Schottkys have a lower drop).

The MOSFET gate drive required is determined by the MOSFET gate capacitance and the PWM frequency.
You should be able to calculate that and it shouldn't require a pot.

Why are you running a 120V motor from a 220Vac supply?
Are you in a location where there is no 120V available?
Yes, to get 120V I need to use a transformer.
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
I do not think this circuit will pass AAC line operated equipment standards, & if 24 V supply is not derived from
the rectified 220 , then there will be a ground mismatch.
Thanks, the PLC power source is connected to the same 220V line. Will that make the circuit compliant?
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
I see no fundamental problem with using PWM on a full-wave rectified unfiltered sinewave to control the motor, no filter needed.

6KHz is in the audible range so that is a consideration if such a noise from the motor would be annoying.

Isolation is likely required so you could use opto isolators to drive the MOSFET gates.
You would need a simple DC supply at the output of the optos to drive the MOSFET gates.

I'll see how to include an opto isolator, I might be able to replace the smaller MOSFET and connect the optoisolator output to the power MOSFET.
Thanks for the suggestion!

Heat depends upon the motor current, the ON resistance of the MOSFETs, and the voltage drop of the flyback diodes (Schottkys have a lower drop).

The MOSFET gate drive required is determined by the MOSFET gate capacitance and the PWM frequency.
You should be able to calculate that and it shouldn't require a pot.

Why are you running a 120V motor from a 220Vac supply?
Are you in a location where there is no 120V available?
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
The voltage would only reach 311vdc if you used filter cap on rectified 240v, otherwise the mean level will be less than 240v.
For only 180rpm with that motor, you would only require 13.5v DC with no load.
If you chose to go with no filter the results may be OK, maybe not, I have never come across a PWM drive that had no capacitors, this includes VFD's that are feed by a 3ph bridge where the ripple is a few % and the freq. is 360°, not 100% 120hz.
I use 5Khz to 6Khz and never noticed any noise, but my hearing is not what it was!:)
When using low rpm application with a high rpm motor it is usually more efficient and cost effective to go with a much small motor and gearing.
Max.
 
Last edited:

crutschow

Joined Mar 14, 2008
34,280
................
If you chose to go with no filter the results may be OK, maybe not, I have never come across a PWM drive that had no capacitors, this includes VFD's that are feed by a 3ph bridge where the ripple is a few % and the freq. is 360°, not 100% 120hz.
...........
Max, I see no significant differences between using PWM on an unfiltered rectified sinewave and an unfiltered SCR phase-control drive as far as the motor is concerned.
Actually the PWM circuit should give a smoother current flow.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
I just question why VFD's and large BLDC drives with 3 ph rectifiers always use a Very large bank of capacitors if it was not that necessary?
I guess if the OP tries it, he can let us know how it went.
Max.
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
I just question why VFD's and large BLDC drives with 3 ph rectifiers always use a Very large bank of capacitors if it was not that necessary?
I guess if the OP tries it, he can let us know how it went.
Max.
I will post the results. I will change the circuit to include an optoisolator and change the flyback diode for a schottky.

Thank you all for your advice and comments
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
Sorry for taking this long, finally I was able to test the circuit on a DC motor.

Had to increase the number of MOSFETs to manage the heat, so I put 2 more in parallel with the first one on a big heatsink.

Since the circuit was drawing a lot of current I used cables, which I had to twist to stop some noise that was causing problems at some frequencies.

Now it is possible to control the motor torque by variating the PWM, however the MOSFETs get very hot even at small tests. I have changed the gate resistors from 470K to 100K to ensure the MOSFETs were at the saturation region, but the heating problem remains, even at a very low torque (almost no load and the motor running at less than 60 rpm).

What could be the heating cause?

I´m attaching the current circuit diagram and the MOSFET datasheet.

Thanks for your comments!
 

Attachments

jpanhalt

Joined Jan 18, 2008
11,087
The immediate answer is to add a relatively large (e.g., ≥470 uF) capacitor across the supply as in this schematic form Microchip AN898:

upload_2016-7-8_13-28-13.png

I sense you may want to know why. Years ago, I found the educational articles from 4QD-Tec drives immensely helpful. Here is one on point and a more general link. I don't recall how big your motor is. The size of your motor will determine the size of capacitor to use.

http://www.4qdtec.com/pwm-01.html
http://www.4qdtec.com/

John

PS: you may want to edit your schematic and just show DC voltage, not the mains source.
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
The immediate answer is to add a relatively large (e.g., ≥470 uF) capacitor across the supply as in this schematic form Microchip AN898:

View attachment 108870

I sense you may want to know why. Years ago, I found the educational articles from 4QD-Tec drives immensely helpful. Here is one on point and a more general link. I don't recall how big your motor is. The size of your motor will determine the size of capacitor to use.

http://www.4qdtec.com/pwm-01.html
http://www.4qdtec.com/

John

PS: you may want to edit your schematic and just show DC voltage, not the mains source.

Thank you for the URLs, indeed I would like to understand what is happening and how the solution works.

About the schematic, I left the main source to make more evident that the current, even though is not alternating phases, is a semi sinusoid curve (after the rectifier bridge there is no filtering to smooth the result). I have been looking for some information about applying a PWM on a sinusoid current but couldn´t find much about it...

Thanks again!
 

Thread Starter

onlyonce18

Joined May 8, 2016
48
IMO I would say your gate resistors are way to high, 500ohm maximum as a general rule.
Max.
Will try that too. With the 100K resistor the Vgs was 12V, measured when using a 24V source instead the 220V, to make some tests.

The optocoupler output has 50mA according to its specifications, and the test results for the mosfet indicate 250 μA at a threshold voltage of 2V, so current shouldn´t be a problem...?

Thanks again!
 

jpanhalt

Joined Jan 18, 2008
11,087
I didn't mention it, but I agree completely on the gate resistors. 100K is way too high for the turn-on resistors. I use a value of <22 Ω if that stops ringing. As for your turn off resistor, that depends on how the gates are being driven. It may not be necessary at all; although, I have used a 100K there to ensure the gates are off when power is first applied to the device. Once my gate drive is running, its ability to sink current turns off the gate just fine. Some people even add a diode (cathode toward the driver) to hasten turn off. I would not recommend that, if your driver can't sink that current.

John
 

Alec_t

Joined Sep 17, 2013
14,280
When you reduce the value of the resistors in series with the FET gates you should also reduce the value of the gate-to-source resistors so that the maximum Vgs of 20V is not exceeded.
 
Top