tips for smoothing motor PWM

[Hank Fletcher]

I have a small motor that I want to run at a low voltage of about 1V. What I plan to do is a 20% PWM signal from a 5V source. The motor is rated for max 3V.

What's the best way for smoothing the PWM signal to the motor, so that it receives a relatively constant 1V? Where motors are concerned, after the PWM frequency is over, say, 1kHz, does it really even make a difference?

 

[Nigel Goodwin]

You don't smooth PWM to feed a motor, it would be pointless - and lower performance of the motor.

 

[Hank Fletcher]

Pointless how? I don't understand how it would effect motor performance.

 

[Hank Fletcher]

I just don't understand the reasoning behind using capacitors for limiting back EMF here:
https://www.electro-tech-online.com/threads/value-of-the-capacitor-parallel-with-the-dc-motor.21305/

and why the same reasoning wouldn't apply to a PWM situation, especially if the PWM signal is above the motor's rated voltage. Aren't you in effect, by using the high-voltage PWM signal, sending multiple "spikes" of a sort to the motor?

Or did your answer presume the caps mentioned for the the purpose in the other post, i.e. since there are (presumably) caps for back-EMF purposes, additional caps for PWM are redundant?

I'm just trying to understand the reasoning behind it, is all.

 

[Nigel Goodwin]

Pointless how? I don't understand how it would effect motor performance.

The high voltage pulses give more torque to the motor, slowing a motor down by reducing it's voltage at a DC level reduces torque considerably.

 

[ericgibbs]

I just don't understand the reasoning behind using capacitors for limiting back EMF here:
https://www.electro-tech-online.com/threads/value-of-the-capacitor-parallel-with-the-dc-motor.21305/

and why the same reasoning wouldn't apply to a PWM situation, especially if the PWM signal is above the motor's rated voltage. Aren't you in effect, by using the high-voltage PWM signal, sending multiple "spikes" of a sort to the motor?

Or did your answer presume the caps mentioned for the the purpose in the other post, i.e. since there are (presumably) caps for back-EMF purposes, additional caps for PWM are redundant?

I'm just trying to understand the reasoning behind it, is all.

hi Hank,
Ref the back emf caps in the link.
You know that when current flowing in an inductor [motor winding] is rapidly switched OFF a back emf is produced.
The voltage level is Vbemf= - Lind * di/dt.

This voltage is a 'spike' which can be many times greater that the motor supply voltage.
As its a spike, its made of many frequencies or harmonics.

You recall that the reactance[resistance] of a capacitor = 1/[2 * pi * freq * capF]
That means the higher frequencies in the spike are are effectively shorted by the cap, so limit the voltage level
of the spike to a 'safe' limit for the drive transistors.

Hope this explains a little better.

 

[Ubergeek63]

The high voltage pulses give more torque to the motor, slowing a motor down by reducing it's voltage at a DC level reduces torque considerably.

Torque has nothing to do with voltage, other than the fact that a higher voltage can drive more current.

The inductive spike is potentially damaging to both the drive transistor AND the motor itself, though typically the transistor blows long before the motor insulation breaks down.

That said the best approach for a unidirectional drive is a simple diode across the motor to provide a path for the armature inductance's current while the transistor is off.

 

[crutschow]

You definitely don't want to put a capacitor at the output of a PMW circuit (except perhaps a very small one) since large current spikes will occur in the PWM circuit when it turns on. This will cause the PWM circuit to heat up and reduce its efficiency (approaching the poor efficiency of a simple series regulator).

Instead a diode is added across the motor windings (cathode to the + terminal) to limit any inductive spike and protect the transistors. This provides a path for the motor current when the PWM signal is off. The motor inductance thus acts to smooth the pulses from the PWM circuit, rather similar to a switching regulator.

 

[Speakerguy]

An RC snubber on a PWM output is another method that works better than a diode to prevent voltage spikes, especially if you optimize it properly. You need a scope to do that. Without a scope, a 100R and a 100nF capacitor works pretty good for most purposes.

Attached is an LTSpice (SwitcherCAD) simulation that will let you compare the two. Feel free to change parameters, I don't know what the inductance of your motor is or what diode you might use. An RC snubber is so easy I wonder it is not more commonly recommended than a diode?

 

[crutschow]

An RC snubber is so easy I wonder it is not more commonly recommended than a diode?

The RC snubber takes care of the high frequency voltage spikes due to parasitic inductance from wires and circuit traces.

But you also need a diode (or a synchronous switch in some cases) for both dc motors and switching regulator inductors in PWM circuits. This provides the path for the primary inductor current to flow between pulses when the PWM switch is off, to give an average dc current. This current is actually fundamental to the way a PMW circuit operates. If that current is snubbed with an RC circuit you're wasting all the energy stored in the inductance.

 

[Speakerguy]

Gotcha. Made sense to me when you used the switching regulator example. You're right, the RC would absolutely kill efficiency.

 

[Boncuk]

You definitely don't want to put a capacitor at the output of a PMW circuit (except perhaps a very small one) since large current spikes will occur in the PWM circuit when it turns on. This will cause the PWM circuit to heat up and reduce its efficiency (approaching the poor efficiency of a simple series regulator).

Instead a diode is added across the motor windings (cathode to the + terminal) to limit any inductive spike and protect the transistors. This provides a path for the motor current when the PWM signal is off. The motor inductance thus acts to smooth the pulses from the PWM circuit, rather similar to a switching regulator.

Hi Carl,

I agree with you. Using a cap across the motor more or less nulls PWM. However a small cap helps considerably reducing the noise resulting from PWM frequency. I tried that with a CPU cooling unit for a power transistor. The noise frequency was obviously in resonance with the heat sink and it was pretty loud and annoying. Using a 2µ2 electrolytic cap almost reduced the noise to zero with minimum effect on motor starting - it even started smoothly at very low rpm.

BTW, outdated Pentium IV coolers make good active heat sinks for power applications.

Hans

 

[Nigel Goodwin]

Torque has nothing to do with voltage, other than the fact that a higher voltage can drive more current.

So it's EVERYTHING to do with voltage then - you rather contradicted yourself there!