You and me both!
..a

This is what Faulhaber (who make coreless motors) say:

One of the factors limiting brush and commutator life is the inductance of the motor armature. During commutation, when current flows through a particular coil winding there is storage of energy in the form of a magnetic field. When the motor commutates and the current flow is switched to another winding, the magnetic field collapses and the resulting discharge of energy causes an arc between the commutator and brush. This arcing accelerates electro-erosion and decreases motor life. One could, theoretically, reduce the armature inductance of the motor windings by decreasing the number of turns in each armature segment. This lowers the torque constant of the motor, however, which increases the motor current for a given torque and, therefore, increases the current density at the brush-commutator interface. This, obviously, is not a viable option. Several manufacturers of coreless DC motors now offer certain motor series with a capacitor disk mounted to the commutator. Each winding is connected in parallel with a small capacitor. The collapse of the magnetic field during commutation then serves to charge the capacitor rather than creating an arc between brush and commutator. This technique, while increasing the electrical time constant of the motor somewhat, is extremely effective in increasing motor service life.

Operating conditions other than torque and speed also affect service life. The application may require frequent starting and stopping or reversals of direction. Both situations result in periods of high current density and a resulting shortening of service life. A similar effect is seen in applications where pulse width modulated (PWM) drives are used. If the PWM frequency is too low, the motor is constantly accelerating and decelerating with an accompanying increase in current density. As a general rule, PWM frequencies of 20 KHz or higher are recommended for ironless core motors.

http://www.micromo.com/03application.../tutorial4.asp

Hello KillerWatt:

Let me present my best understanding of a possible mechanism to explain part throttle brush burnout in coreless motors. If I oversimplify please excuse me, I mean no offsense, since there is a wide range of technical background here on ezone.

Lets start with the motor equation, Vin = (Km * RPM) + ( I * Rm) where Vin is the motor input voltage, Km is the volt per RPM motor constant ( I am using Km here with units of volts/RPM and not RPM/volt as typically used, Km is only the reciprocal of RPM per volt), I is the motor current (I am assuming Io, the no load current, is 0, as it makes no difference if I is only the load current or the load current plus Io) , and Rm is the motor resistance. What it says is that any given input voltage will be offset by the back EMF (Km * RPM) plus the voltage drop across the motor resistance (I * Rm). What this equation also indicates is that, for a given input voltage and RPM, the motor will draw enough current to cause a voltage drop across its internal resistance to equal the input voltage minus the back EMF, since Km and Rm are fixed motor parameters. This is the reason a stalled motor has such a high current draw, since the back EMF is 0 when the motor RPM is 0. When viewed as a unit, the motor follows these rules and Vin can safely be replaced by Vavg. Vavg being the average voltage as supplied by an ESC using PWM.

Now lets look at a single armature pole. I will suggest that the equation now becomes Vin = (Km * RPM) + (I * Rm) + (Lm * dI/dt). The first part of the equation matches the motor equation and the last term (Lm * dI/dt) adds in the effect of motor inductance, where Lm is the motor inductance and dI/dt is the rate of change of the motor current. As for the motor as a whole, the input voltage on a pole must be offset by the sum total of the back EMF plus the voltage drop across the armature resistance plus the inductance effect. But in this case Vin can only be replaced Vavg if the ESC is switching fast enough to pulse each pole several times as it passes the brushes.

With those definitions out of the way, lets look at a coreless motor. First lets set up a hypothetical system so we can use some math.

Vin = Battery voltage = 10 volts
Km = Back EMF constant = 1 volt per 1000 RPM
Rm = Motor resistance = 0.1 Ohms

With no load and full battery voltage of 10 volts the motor speed is 10,000 RPM. At 1 volt per 1000 RPM the motor must spin 10,000 to make a back EMF of 10 volts since there is no load current to cause an additional voltage drop across the motor resistance. If we add a load (propeller) to the motor and draw 5 amps current the motor speed becomes 9,500 RPM. Where 9,500 RPM makes a back EMF of 9.5 volts plus the 0.5 volts from (I * Rm) or 5 amps * 0.1 Ohms = 0.5 volts.

Now look at the motor at 50% duty cycle (half throttle more or less) and no load. Vin becomes Vavg which is 5 volts, and the motor speed is 5,000 RPM. Again, 5,000 RPM times 1 volt per RPM gives 5 volts which is the average input voltage. Similarly with a 5 amp load the motor speed is now 4,500 RPM.

From the armature's point of view the above calculations only apply if Vin can be replaced by Vavg as was the case for the motor. For this to be true the ESC mush be switching fast enough so that each pole is energized several times as it passes the brushes. Now the problem occurs, the motor is operating at part throttle and no load with Vin of 10 volts and Vavg of 5 volts. The motor RPM is 5000 but the armature pole that is not modulated as the passes the brushes sees Vin, not Vavg. The armature equation (disregarding the inductance effect) is 10 volts = ( 1 volt per RPM * 5000 or 5 volts) + (I * 0.1 Ohm). The motor current I, must now rise to 50 Amps to produce a voltage drop of 5 volts across the motor resistance of 0.1 Ohm. Remember that all this happens as a single pole passes the brushes so motor inertia keeps the speed almost constant. There is the current spike and the reason for the brush burning. Interestingly enough this analysis predicts that the problem will occur even at no load. Addition of a load doesn't do much except change the numbers a little.

The above analysis omits, simplifies, and glosses over several possible details in an attempt to convey the concept behind my explaination of the part throttle problem of coreless motors. None of these should have an effect on the validity of the concept. The inductance term in the armature equation was included since the same part throttle brush problem exists in the standard brushed motors also and they don't burn out. I assume that the higher inductance of the standard motors limits the rise of the motor current and the problem arises in coreless motors because of their low inductance.



DNA: FET technology keeps improving so we can build faster ESCs. A 9 pole motor at 20,000 RPM gives almost the same numbers as a 5 pole motor at 40,000 RPM.


The above is only my opinion you may choose to believe it or not. And please be gentle with the flamethrowers, I am out of asbestos.

Kevin

Great explanation. So it is the back emf that limits the currents and burning of the brushes at full throttle (self limiting to a degree). Very well put.

hey guys.. thanks for all the Heavy motor info...i'm still digesting all of that emf stuff....btw, anybody know of a very light weight 5 amp, ultra hi-rate ESC, suitable for a 9 segment core-less motor............thanks ..............kw

Schulze Slim-105bek 5amp 100khz.

Not available in US yet that I can find.

Thanks everyone...

I am aware of the increased power dissipation due to higher switching frequency but am not really concerned about this as my speed controls still run cool even at 3KHz.
What I was really wondering is if there was any frequency that resulted in less loss of range at part throttle or if different frequencies made any difference in motor life.

After testing one of my speed controls running at 3KHz in a model this weekend I can’t say noticed any difference in range at part throttle or anything else.


Regards…


Paul.

I have been runing a coreless motor (Orion Elite) on a 3k ESC now with no problems as of yet. I only have 5 hours (about 20+ 15 minute flights) on the motor so far with 90% of that time spent at 50-70% throttle. The current is generally in the 2-3 amp range. This motor has the nice benefit of having easy access to the brushes. I looked at them yesterday under a magnifying glass and noticed only slight wear on the fingers and no pitting or discoloration. The comm looked nearly brand new. I predict This motor is going to soon become a favorite hear on the Ezone (thank you DNA for making us aware of this motor!) with its speed 300 power @ 280 current levels and will become even more popular if standard ESC's don't dramatically shorten its life. I will post if it fails prematurely. I think the high Freq. ESC is a valid argument on ultra high rpm coreless motors but since the vast majority of motors will be spending there time around the 10-25k partial rpm mark, IMHO the standard 3K rate is just fine. And in the Orion Elite case, as long as 10 or so hours can be had from a set of brushes, none of this really matters as Orion will be releasing replaceable brushes soon. Read below for more info.

http://www.team-orion.com/cgi/ultima...c&f=2&t=000258