Dynamic Braking of 3-Phase DC Motor

[Phrak]

To brake a 3ph DC, the best I've come up with, apparently, is to place a pair of power FETS, back-to-back across each pair of leads.

I would like to manage a hard short without additional load beyond the motor winding resistance, then duty cycle to get controlled braking.

As far as having the (positive) load current in a NPN MOSFET directed from source to drain, I have only heard of this done once, and that was at low currents.

Is this at all feasible to do this up near the rated current, or am I completely wrong about this?

 

[rbj]

uh, what's a 3-phase DC?

what does phase have to do with DC?

 

[NoTime]

3-phase dc uses an electronic commutator.
http://www.pmdcorp.com/advanced-motion-control/brushless-dc-motor-controller.cfm

I think you would melt the windings if you tried this at the motors max rated power.

 

[Phrak]

NoTime, I need to dump about 800 joules in 3 seconds. At the maximum rated torque the motor disipiates about 120 watts. That's 6 or 7 seconds worth of power. It could have a heating problem at high braking duty agrivated by high torque duty.

How about those FETs?

 

[rbj]

to repeat, what is "3-phase DC"?

i don't work in electromechanical energy conversion (a.k.a. "power") but i used to teach EE at the university level. i used to know what 3-phase power is, what 3-phase motors were about, what DC motors were about, what variations like stepper motors or other electronically driven motors were, what "regenerative braking" is (at least in the context of an electric car), but i never even heard of the term "3-phase DC". it sounds like an oxymoron to me.

 

[zeitghost]

Is it not an electronically commutated dc motor?

 

[stewartcs]

to repeat, what is "3-phase DC"?

i don't work in electromechanical energy conversion (a.k.a. "power") but i used to teach EE at the university level. i used to know what 3-phase power is, what 3-phase motors were about, what DC motors were about, what variations like stepper motors or other electronically driven motors were, what "regenerative braking" is (at least in the context of an electric car), but i never even heard of the term "3-phase DC". it sounds like an oxymoron to me.

It's not "3-phase DC", it's a "3-phase DC Motor". It just means that the DC motor has three phase windings, not that it is powered by 3-phase DC, which as you already noted is nonsensical.

The phase windings are switched in a specific order to make the motor rotate.

CS

 

[stewartcs]

I would like to manage a hard short without additional load beyond the motor winding resistance...

Why?

As far as having the (positive) load current in a NPN MOSFET directed from source to drain, I have only heard of this done once, and that was at low currents.

Define low currents.


CS

 

[NoTime]

NoTime, I need to dump about 800 joules in 3 seconds. At the maximum rated torque the motor disipiates about 120 watts. That's 6 or 7 seconds worth of power. It could have a heating problem at high braking duty agrivated by high torque duty.

How about those FETs?

Think of what happens if you short the output of a generator.
Say a motor is 90% efficient. So for your 120 watt motor, 100 watts goes out the shaft and 12 watts gets dissipated as heat in the motor.

800 joules in 3 seconds is nearly 300 watts for a motor designed to get rid of 12w.

 

[stewartcs]

Think of what happens if you short the output of a generator.
Say a motor is 90% efficient. So for your 120 watt motor, 100 watts goes out the shaft and 12 watts gets dissipated as heat in the motor.

800 joules in 3 seconds is nearly 300 watts for a motor designed to get rid of 12w.

Sounds like you need more time to dump the load, or you need some mechanical braking to assist.

CS

 

[RonL]

NoTime, I need to dump about 800 joules in 3 seconds. At the maximum rated torque the motor dissipates about 120 watts. That's 6 or 7 seconds worth of power. It could have a heating problem at high braking duty aggravated by high torque duty.

How about those FETs?

The FETs are neat, and I'm guessing your using something like hard drive motors. Can you say anything about how big the system can be ?
As stewartcs suggested something mechanical might be needed to absorb that much wattage. Are you looking for ideas ?

 

[Phrak]

Think of what happens if you short the output of a generator.
Say a motor is 90% efficient. So for your 120 watt motor, 100 watts goes out the shaft and 12 watts gets dissipated as heat in the motor.

800 joules in 3 seconds is nearly 300 watts for a motor designed to get rid of 12w.

Sorry, NoTime. I wasn't clear about that. The power dissipation of 120 watts is the motor's power loss at maximum continuous rating (55 amps through 30 milliohm). The maximum shaft output is 800 watts or so.

My application is for a vehicle that takes about 3 seconds to stop. If this occurs once every 7 seconds, the motor itself will be dissipating 113 watts, average in heat.

This assumes that the heat capacity of the wire is sufficient to absorb excess heating at the low pulse rate of 1/7th Hz. Also it involves the thermal resistance of the windings to ambient, and all that, but I'm not sophisticated enough to understand how all that is calculated.

 

[Phrak]

Why?

Hey, stewartcs.

Why do I want a dead short?

I'll be using the motor to both power and brake a vehicle. I can't obtain an acceptable deacceleration rate otherwise. .. Maybe I could supplement with dissipating up to 20% with an external load, but not much.

Define low currents.

It's been a long time, but I believe the current in question was less than -10% of I_{D}max. at a V_{GS} of 15 volts (full on).

 

[Phrak]

The FETs are neat, and I'm guessing your using something like hard drive motors. Can you say anything about how big the system can be ?
As stewartcs suggested something mechanical might be needed to absorb that much wattage. Are you looking for ideas ?

I'm sure you guys must be right. I'm comming around to combined mechanical and dynamic braking.

I don't have a great deal of diameter to do mechanical breaking.
The total available volume for braking is a cylinder 2" in diameter by 5 inches long with the wheel shaft passing through the center. In addition it could have as much as 1 inch protrusions off two sides of the cylinder over it's entire length.

Any suggestions would be greatly appreciated.

and I am still clueless about FETs...

-deCraig

 

[stewartcs]

Hey, stewartcs.
Why do I want a dead short?

Wrong why. I was asking why not load it beyond the resistance of the windings. In effect, I was asking why not use a resistor bank to dynamically stop it...Not why do you want manage a dead short.

CS

 

[stewartcs]

It's been a long time, but I believe the current in question was less than -10% of I_{D}max. at a V_{GS} of 15 volts (full on).

I seem to remember it being around 3-5 amps. Not 100% sure though.

CS

 

[stewartcs]

I'm sure you guys must be right. I'm comming around to combined mechanical and dynamic braking.

I don't have a great deal of diameter to do mechanical breaking.
The total available volume for braking is a cylinder 2" in diameter by 5 inches long with the wheel shaft passing through the center. In addition it could have as much as 1 inch protrusions off two sides of the cylinder over it's entire length.

Any suggestions would be greatly appreciated.

and I am still clueless about FETs...

-deCraig

There is always Plugging, but that can possibly over-stress the motor.

CS

 

[NoTime]

Here are some FET circuits that may or may not be useful.
http://www.discovercircuits.com/F/fet1.htm

Most dynamic braking systems use resistor banks to dissipate the energy or in the case of regenerative braking charging the battery.

 

[Phrak]

CS, the rate of deceleration is inversely proportional to the total series resistance. The series resistance is the motor resistance + load resistance + FET channel resistance.

To be precise, Power_of_deceleration = (E_motor)^2 R_total.

E_motor is the voltage generated by the motor obtained by operating the motor as a generator.

So you add an external load, and your rate of decelertion decrease. I'd love to dump the heat elsewhere but it makes for a very soft brake unfortunately.

I hope this answers your question.

 

[stewartcs]

CS, the rate of deceleration is inversely proportional to the total series resistance. The series resistance is the motor resistance + load resistance + FET channel resistance.

To be precise, Power_of_deceleration = (E_motor)^2 R_total.

E_motor is the voltage generated by the motor obtained by operating the motor as a generator.

So you add an external load, and your rate of decelertion decrease. I'd love to dump the heat elsewhere but it makes for a very soft brake unfortunately.

I hope this answers your question.

I wasn't asking a question, it was rhetorical.

If you use a resistor bank as a dynamic brake the resitor bank is generally sized to provide an armature current that will approximate 150 to 300 percent rated current.

The motor behaves as a generator and feeds current to the resistor, dissipating heat at a rate equal to $$I^2R$$.

You'll need a mechanical brake once the motor has slowed dynamically to completely stop and hold it still.

If you need to stop it VERY quickly, then Plugging will do the trick (with the Caveat I mentioned previously). Just remember to use a Plugging resistor in series with the armature circuit to limit the current.

CS

 

[Phrak]

I wasn't asking a question, it was rhetorical.

If you use a resistor bank as a dynamic brake the resitor bank is generally sized to provide an armature current that will approximate 150 to 300 percent rated current.

The motor behaves as a generator and feeds current to the resistor, dissipating heat at a rate equal to $$I^2R$$.

You'll need a mechanical brake once the motor has slowed dynamically to completely stop and hold it still.

If you need to stop it VERY quickly, then Plugging will do the trick (with the Caveat I mentioned previously). Just remember to use a Plugging resistor in series with the armature circuit to limit the current.

CS

OK, so you know basic motor theory.

I ask a simple question about a FET a practiced and experienced engineer in DC motor controls should be capable of knowing.

I'm new to DC motors. Everything you tell me anyone can pick up and assemble off the internet given the motivation and where-with-all.

I'm familiar with the exponential deceleration decay inherent in dynamic braking. I've considered nearly every combination of dynamic, regenerative, mechanical and 'plugging' as you call it, within innumerable design constraints from the physical to practical to manufactuarability and costing, thru consumer appeal, and on and on and on.

I've been doing engineering for other-peoples-problems for over twenty years. This one is mine.

I've fielded dozens of questions overall on this board. I get zero back.
I should be asking where the serious engineering forums are. Do you know?

Practical experience would tell me if I could run MOSFETS back to back with conduction through the channels, rather than the reverse body diode. Practical experience would tell me the tried and true architecture of dymanic braking circutry.

 

[RonL]

I should be asking where the serious engineering forums are. Do you know?

Try this one, if you don't already know about them.


http://www.eng-tips.com/

Ron

 

[stewartcs]

I'm sure you guys must be right. I'm comming around to combined mechanical and dynamic braking.

I don't have a great deal of diameter to do mechanical breaking.
The total available volume for braking is a cylinder 2" in diameter by 5 inches long with the wheel shaft passing through the center. In addition it could have as much as 1 inch protrusions off two sides of the cylinder over it's entire length.

Any suggestions would be greatly appreciated.

and I am still clueless about FETs...
-deCraig

If you are indeed an EE and have, for the past twenty years, been fielding dozens of questions, I'm certain you wouldn't be "clueless" about FETs.

You seemed to be asking for alternate suggestions, that's why I gave you one. If you don't like it, then fine, don't use it.

OK, so you know basic motor theory.

I ask a simple question about a FET a practiced and experienced engineer in DC motor controls should be capable of knowing.

I'm new to DC motors. Everything you tell me anyone can pick up and assemble off the internet given the motivation and where-with-all.

I'm familiar with the exponential deceleration decay inherent in dynamic braking. I've considered nearly every combination of dynamic, regenerative, mechanical and 'plugging' as you call it, within innumerable design constraints from the physical to practical to manufactuarability and costing, thru consumer appeal, and on and on and on.

I've been doing engineering for other-peoples-problems for over twenty years. This one is mine.

I've fielded dozens of questions overall on this board. I get zero back.
I should be asking where the serious engineering forums are. Do you know?

Practical experience would tell me if I could run MOSFETS back to back with conduction through the channels, rather than the reverse body diode. Practical experience would tell me the tried and true architecture of dymanic braking circutry.

How is it that you are new to DC motors yet are so familiar with them?? FYI, that's a rhetorical quetion too.

CS

 

[NoTime]

Practical experience would tell me the tried and true architecture of dymanic braking circutry.

There is no tried and true.
Dynamic braking is a niche process, generally inefficient, often expensive and normally quite specific to the application.
The most common use is for rotating saw blades, very little stored energy.
The mechanism is simple.
Just a DPST switch that shorts or applies power to the motor.
The thermal mass of the motor is sufficient to constrain the temperature increase.

From what you have previously stated, your application seems to be a poor candidate for dynamic braking.
If you can make it work for you, then good, but it will be a major effort on your part.

 

[Phrak]

Thanks, NoTime.

I'll be looking into it Ron.

 

[zeitghost]

To brake a 3ph DC, the best I've come up with, apparently, is to place a pair of power FETS, back-to-back across each pair of leads.

I would like to manage a hard short without additional load beyond the motor winding resistance, then duty cycle to get controlled braking.

As far as having the (positive) load current in a NPN MOSFET directed from source to drain, I have only heard of this done once, and that was at low currents.

Is this at all feasible to do this up near the rated current, or am I completely wrong about this?

Why?



Define low currents.


CS

Hey, stewartcs.

Why do I want a dead short?

I'll be using the motor to both power and brake a vehicle. I can't obtain an acceptable deacceleration rate otherwise. .. Maybe I could supplement with dissipating up to 20% with an external load, but not much.



It's been a long time, but I believe the current in question was less than -10% of I_{D}max. at a V_{GS} of 15 volts (full on).

I seem to remember it being around 3-5 amps. Not 100% sure though.

CS

There was an ancient Siliconix application note about the inverse connection of PowerMos in a synchronous rectifier... the on channel resistance was lower than the forward drop of the intrinsic diode, so made for greater efficiency. Whether the same thing can be achieved in this application is beyond me, though considering the decrease in Rds(on) and increase in Id(on) over the last 20 years, anything is possible.

 

[zeitghost]

Try this one, if you don't already know about them.


http://www.eng-tips.com/

Ron

A good site.

The depth of water sensing thread is to be particularly recommended.

 

[Phrak]

For what it's worth, after an exhaustive search, I discover that the FET implimentation of an SSR (solid state switch) is often constructed of twoP-channel MOSFET with their sourced connected.

If the little drawings associated with the data sheets are an accurate representation, they incorporate the same architechure as power FETs with the substrate shorted to the source.

According to the IR technical representative, and gleening information from the SSR data sheets, power FETs conduct just as well under reverse source-drain bias as forward conduction.

The lowest R_ds currently offered by IR is an incredibly low one milliohm if your application is under 24 volts V_ds peak. So I'm good to go.