Go Down

Topic: Understanding DC motor freewheeling (Read 18542 times) previous topic - next topic

crxguy52

I've used a few DC motors before with a flyback diode, and haven't thought much of it. I've recently been thinking about building a larger DC motor driver - 36V, 100A for a golf cart. In doing so. I've been reading about motor controller design. I'm a bit confused about how back EMF is treated, this is how I understand it:

When an inductive load, like a motor, is rapidly disconnected from it's voltage source, the inductor will try and maintain constant current. If it is left with nowhere to go, voltage will spike over across the motor's leads. This can burn out components, mostly solid state, or cause arcing in switches

To prevent this, there are three options:

1. Disconnect the motor's leads completely, and let it freewheel

2. Connect a flyback diode to regulate the voltage across the leads, but also dissipate current in the winding \ diodes. This has a motor braking affect. See the flyback diodes below:


3. Using an additional MOSFET, PWM the motor between shorting (to gain current momentum) and connected to the battery regularly (using the current momentum to charge the battery). This is similar to a boost converter. Switching must be done very quickly to avoid braking the motor with great force (shorted) or driving the motor (connected regularly). Times are determined by the resistance and inductance of the motor.  See below for sample circuit:



A) This makes sense, but how to electronic PWM motor controllers let motors freewheel? Without disconnecting the controller, some sort of current is going to have to be disappated (either through a flyback diode or the battery), resulting in motor braking.

B) Is there a way to maintain the option to use regenerative braking and still freewheel?

My primary sources are:
http://electronics.stackexchange.com/questions/56186/how-can-i-implement-regenerative-braking-of-a-dc-motor
http://zeva.com.au/Projects/Speedy/

MarkT


I've used a few DC motors before with a flyback diode, and haven't thought much of it. I've recently been thinking about building a larger DC motor driver - 36V, 100A for a golf cart. In doing so. I've been reading about motor controller design. I'm a bit confused about how back EMF is treated, this is how I understand it:

Firstly I'll stop you there, back EMF means two dinstinct things in (DC) motors - the back EMF generated by the spinning windings
in the motor's magnetic field - this is proportional to the rotational speed.

Secondly the back-EMF of a winding when the current is cut off due to its self-inductance - you seem to be talking about this primarily.

Both forms of back-EMF are voltages induced due to a change in the magnetic flux linking a winding, but the mechanisms are
conceptually different.
Quote


When an inductive load, like a motor, is rapidly disconnected from it's voltage source, the inductor will try and maintain constant current. If it is left with nowhere to go, voltage will spike over across the motor's leads. This can burn out components, mostly solid state, or cause arcing in switches

To prevent this, there are three options:

1. Disconnect the motor's leads completely, and let it freewheel

Surely there is a smiley missing - disconnecting the leads is what generates the back EMF spike!
Quote


2. Connect a flyback diode to regulate the voltage across the leads, but also dissipate current in the winding \ diodes. This has a motor braking affect. See the flyback diodes below:


Only works for unidirectional motor - if you are driving bi-driectionally from an H-bridge each arm of the bridge needs diodes - 4 in total.
There is minimal (no?) braking effect from this as the current reduces rapidly and then there is no current (unless PWM is involved).
Quote


3. Using an additional MOSFET, PWM the motor between shorting (to gain current momentum) and connected to the battery regularly (using the current momentum to charge the battery). This is similar to a boost converter. Switching must be done very quickly to avoid braking the motor with great force (shorted) or driving the motor (connected regularly). Times are determined by the resistance and inductance of the motor.  See below for sample circuit:

You mean buck converter, not boost converter - the point is this is PWM via half an H-bridge and has diodes built-in (all power MOSFETs have
diodes built-in due to the vertical current flow).  The PWM rate has to match the inductance of the windings (be high enough to reduce
current oscillations to a low enough level).

The important thing with any circuit connected to an inductor is that there is an alternate path for current to continue flowing when
any switch (semiconductor or otherwise) opens.  Without an alternative the rate of change of current will be extremely high, thus the
rate of flux-linkage changes very rapidly, hence the induced EMF is very high (and something is damaged).
Quote



A) This makes sense, but how to electronic PWM motor controllers let motors freewheel? Without disconnecting the controller, some sort of current is going to have to be disappated (either through a flyback diode or the battery), resulting in motor braking.

I think you are confusing the two types of back-EMF - the switching transients are typically in the realm of microseconds or at
worst a few milliseconds.
When you switch off all the MOSFETs in the Hbridge the last transient spike will dissipate very quickly and the motor is no longer carrying
any current.  Voltage, yes, the rotational back-EMF will be present.

Also when the switch opens the current during the transient is still flowing in the same direction, so it is still trying to push the rotor
round normally, so rather than braking the transient is actually powering the motor (albeit very briefly).
Quote


B) Is there a way to maintain the option to use regenerative braking and still freewheel?


No, that would imply generating energy from nowhere.
Quote

My primary sources are:
http://electronics.stackexchange.com/questions/56186/how-can-i-implement-regenerative-braking-of-a-dc-motor
http://zeva.com.au/Projects/Speedy/
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

crxguy52

Thanks for the explanation! Makes much more sense now. Just to clarify:

There are two causes of back EMF:

1. Rotor spinning and generating a voltage - This is low, and not as much of a concern for blowing things up. It linearly increases with speed, matching the control voltage at full speed
2. EMF from the inductance of the motor - This can potentially be very high, and can destroy components. It occurs when voltage is suddenly removed from the motor, and is caused by rapidly stopping the flow of current.

As I understand it, there are two methods of controlling back EMF from point 2 above:

3. Flyback diode will break down at a certain voltage, and short the motor to prevent a large voltage spike. Once the spike dissipates, it goes back to essentially being open, leaving only the EMF from point 1.

4. Synchronous Rectification - can be used in two ways:
     a. To dissipate the initial voltage spike by PWM-ing the flyback MOSFET, then leaving the flyback MOSFET off and letting the motor freewheel. The only voltage across the half H bridge would be that of the EMF from point 2.
     b. The initial voltage spike is dissipated as in point a, but the flyback MOSFET is left on until the current switches directions and reaches a safe value. Once this occurs, the flyback MOSFET is turned off and the drive MOSFET is turned on long enough to recharge the battery. The cycle then repeats.

Does that sound right? If so, using synchronous rectification as in A would have the advantage that instead of dissipating heat in the flyback diodes (which are connected to the same heatsink as the drive MOSFETS, thus heating them up), you essentially burn off the current from the motor's inductance in the motor's windings. Using brake regen would have obvious advantages.

Thanks for the help

MarkT



3. Flyback diode will break down at a certain voltage, and short the motor to prevent a large voltage spike. Once the spike dissipates, it goes back to essentially being open, leaving only the EMF from point 1.



No, the flyback diode conducts in the normal forward direction - there's no high voltages at all.  You ideally use a
fast rectifier, not a common-or-garden 1N4001, since a slow response allows some spiking.

I hesitated to use the phrase "synchronous rectification" since that applies to AC->DC conversion, strictly speaking.

Once you have a full H-bridge with 4 diodes there is no way to get spikes - there are different ways to switch it
though - google "slow decay mode" and "fast decay mode".  Here synchronous rectification == fast decay mode,
and generates less heat in the MOSFETs (diodes drop 1V or more, an on-MOSFET is typically a lot lower).
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

crxguy52

To preface this, I'm looking at building a half H bridge, so this is the lens through which I'm viewing this problem.

I think I understand now why using a MOSFET is better than a flyback diode. A flyback diode has a voltage drop across it (even if it's just 0.3V), which will dissipate the current continuing to run through the inductor (after the source is turned off), into heat. A MOSFET will let the current keep circulating, essentially continuing to power the motor until it dissipates. Care must be taken not to leave the MOSFET on too long, or the EMF from the motor spinning will generate a current the other way, braking the motor. The diode wouldn't act as a motor brake because it's oriented such that the current can flow through it if back EMF is generated, but not if the motor tries to act as a generator (the current would be reversed, so the diode would block it).

I looked up fast and slow decay modes, and it explained a lot. A slow decay mode would be using the flyback MOSFET to short the motor until the inductance EMF is dissipated, and with a half bridge, there isn't a way to fast decay. This is the circuit I'm envisioning building:



Am I correct in the operation of the flyback MOSFET? If so, I'll stop bugging you :)

1. To dissipate back EMF from motor inductance, they flyback MOSFET would be turned on until the current has dissipated. It must be turned off at this point, or else the current will reverse and motor braking will occur.

2. Regenerative braking can be achieved by activating the flyback MOSFET as in 1, but allow the current to reverse to a safe level. The flyback MOSFET is then turned off, and the drive turned back on. This will cause the current to flow back into the battery. When the current is at or near zero again, the cycle is repeated.


Thanks for the help

Go Up