DC Motor Control - Heat Issue

Have you looked at any of the large size drum core inductors? Renco pioneered them, years ago, but they are made by many companies, today, and many deliver at prices way below Renco.

Look at the size 13 inductors at the bottom of this catalog.

I read in sci.electronics.design that Andrew wrote (in ) about 'DC Motor Control - Heat Issue', on Tue, 18 Jan 2005:

There is never any accounting for company policy! You have a problem to solve and policy pre-emptively rules out a likely low-cost solution. Go find a less donkey-brained employer.

I read in sci.electronics.design that Mark Jones wrote (in ) about 'DC Motor Control - Heat Issue', on Tue, 18 Jan 2005:

Billions of audio transformers use laminated silicon iron cores. But the stuff used for microwave transformers isn't likely to be very good at 8 kHz; it may, however, be good enough. Generally, one would choose grain- oriented silicon-iron 0.3 mm thick maximum for 8 kHz, but this inductor carries 15 A DC and therefore needs a large air-gap. This means that the iron quality need not be so critical.

Rich,

Why do you suggest so low? Why are some people saying to try higher than where I am at, and some are saying low? There is no set "right way" I suppose?

The motor driver that the motor company sells and recommends runs at

I found a formula on the web for calculating winding losses that is as follows:

Pr = (RT^2Vs^2)/(48L^2) = (RVs^2)/(48L^2) * (1)/(Fs^2)

where Pr is the loss (in Watts), R is the winding resistance (in Ohms), T is the period (in seconds), Vs is the supply voltage (in volts), L is the inductance (in Henry's), and Fs is the PWM frequency (in Hz).

Assuming this formula to be true for the amount of power lost in the windings (generating heat), it can be seen that there are two ways to reduce this power loss (heat): by raising the PWM frequency, or by raising the inductance. This is the reason that I want to do one or the other, however, Rich, you indicate that a lower frequency might solve my heat issues? It seems that it would be generating more power loss, i.e., heat? I'm getting more and more confused!

Any further thoughts?

I might also add that the formula I put up is based on 50% (worst case) PWM signal!

Mark,

First of all, thanks for bearing with me!

Secondly, the motor definitely is too hot. The housing of the motor reached 225 degrees F after 3 hours of operation at 50% duty cycle. This 225F is roughly 107C, which translates to roughly 203C in the rotor, where the maximum rating for the motor is 155C.

The 3 hour test run consisted of 15 seconds drawing roughly 3 amps, 15 seconds drawing roughly 1 amp. This continued with roughly a 1 second pause in between each cycle. The test setup was an arm with a 15lb. weight on the end. All torque specs for this setup were calculated, and the torque necessary to lift the arm is well below the maximum continuous rating for this motor.

The motor is a high quallity brushed DC motor designed for continuous use at 6A, 24V, max. The specs can be found here:

It the red column, 148867 motor.

motor is very low, and to use it with a PWM signal, the inductance needs to be increase (by a DC choke). Maxon (the motor company) sells some chokes, and indicates that they are necessary to be used with the motor driver that they also sell and recommend for this motor.

I believe that with the driver that I am currently using, I need this choke. With the inductance I currently have (no choke, just the motor's 80uH), I am generating 9.71W due to the current ripple (via the formula I posted earlier, assuming it is correcct). Dissipating this

Raising the inductance by 470uH (with a DC choke), raises total inductance to 550uH, which changes the power generated by the current ripple to 0.21W, reduced by a factor of over 46x! The new temperature rise is roughly 2.6C, this is down from 123C! The new total temperature rise of the rotor is 56.6C, making the rotor roughly 81C, which is much more reasonable).

Hopefully my calculations are correct, and adding this DC motor choke to increasea inductance will decrease the current ripple power loss and in turn decrease the heat generated in the motor. I will let you know how it goes.

Ok good, your drawing doesn't show it but these MOSFETs have built-in reverse recovery diodes. On page 2 of the datasheet is shows the diode is rated for 290A pulsed current / 100nS / 2.0Vf, that should survive anything this motor can throw at it.

Hmm here's how the "freewheeling" or "recovery diode" works. Ignoring a bunch of details including PWM frequency, imagine the PWM pulse has just gone low and the motor stops being energized. A lot of energy has went into creating the motor's strong magnetic field. This begins to collapse. Due to induction principles this forces a potentially kilo-volt and micro-second EMF pulse back at the driver. (The faster the magnetic field can collapse, the higher the reverse voltage potential produced.)

At about -2V, the diode conduction begins, limiting the voltage rise and protecting the MOSFETs from being destroyed by excessive reverse-bias. Inside the motor, the current flowing through the motor and MOSFET diode does not drop to zero instantaneously, thus there is still a partial magnetic field maintained.

On the next PWM rising cycle, the entire system is forward-biased again and the motor's magnetic field grows again. Logically, the less the PWM duty cycle, the less overall power is going to be delivered and retained (or "freewheeled") by the motor and diode - effectively powering it less.

Ignoring a bunch of other factors, here are some alternate things that can cause a motor to become too hot:

1. The motor is underrated for the voltage/current you are applying to it. Some motors, steppers especially, can be over-driven in terms of voltage for a specific purpose. A 12v stepper could be powered by a

1. Missing/unspecified components. Does the motor in question require a built-in cooling fan/housing/heatsink, but had it removed to fit in the space provided? (Or removed by the seller, unbeknownst to you?) Is the motor rated only for 10°C temperature rise at 20% duty cycle? Some motors are designed to provide extreme power in a small space for short durations, while others are designed for continuous duty.
2. Is the motor a compatible type? Remember above that this PWM driving technique retains some magnetism in the motor between pulses. If the motor were a "synchronous", 2-phase, split capacitor, or any other non-DC, single-phase motor then driving it this way may be incorrect. (See
   formatting link
   for examples of two different drivers for both Brushed and Brushless DC motors.)
3. Is "hot" a relative term, or is your motor going out-of-spec? Some motors just operate at higher temp rise than others. My gut feeling is, if you can burn yourself by touching the motor, then it's too hot regardless of rating. Some are designed to operate quite warmly, however. Even so, excess heat wears out bearing grease and causes thermal stresses.

I don't intend to sound scoff or pedantic, merely trying to cover all the bases and/or point out something overlooked. It is very difficult to diagnose a problem verbally... ask anyone who has done telephone tech support!

Regards, Mark Jones

-- "I think the state of the universe at the moment of conceptulization determines part of how an entity further interacts with the rest of its reality, and hence our experience with it." MCJ

It sounds lke you may benefit, significantly by adding an inductance considerable lower than 470 uHy. Did you see my post suggesting a type of low cost choke?

Are they driving the same load? If one motor is slower than the other, then the slower motor will put some drag on the system and the other motor will have to work harder.

In general, there are additional Iron Core losses with PWM for motor control.

Then... if you are gearing the motor shaft speed down to generate more torque also expect to lose (dependent on gear system) up to 30% (or more).

How about gease? Is the motor shaft geased up with high temp grease? Friction is a heat generator.

What method do you use to current sense? Do you have the free-wheeling diode across the motor? Lack of this device will also cause the motor to ring.

Yes, each 10 degree C rise in temperature yields are reduction of life by 50%. Keep them as cool as possible.

pdrunen

It may seem silly to mention this, but adding a series choke without adding a suitable capacitor across the brushed-motor terminals won't work very well. The motor needs to see a low impedance source at the commutating frequencies.

Your intention was, after all, simply to remove driver ripple effects at 7KHz.

RL

That low, really? Well if a lot of heat is being generated, energy is being lost, perhaps moreso in this case.

I've seen a few 3000hp 3-phase AC motors which ISTR were rated at 92% effecient. Entirely different topology though. Even so, they could only be started once every 30 minutes, since temp rise during startup was so great.