Low current kills brush life? WTF?

[duffy]

Recently we have been having issues with motor life at work. They are brush-type PM DC motors used in airplane servos.

Lower voltage improved life, as you would expect. But then we lowered the voltage a little more, and lost a couple hundred k off the average life.

Someone at work suggested we may be seeing this effect -

If the current density exceeds this for long periods, the commutator will run hot, blacken and brush life will be reduced. If the current density is too low, the film will be stripped from the commutator and the commutator will begin to thread.

**broken link removed**

The effects on the commutator seem to verify the effect, but I am having a hard time wrapping my head around why a lower current density would ever be detrimental to brush life.

Anyone have any experience with this?

 

[Hero999]

Does the motor have copper brushes?

The only thing I can think of is that the lower speed might make it easier for the copper to weld itself to the commutator but that still doesn't explain it properly.

 

[shimniok]

From the whitepaper, it looks like "film" is a protective layer that forms/reforms/strips away on commutators and that low current density erodes the film resulting in quicker erosion and wear of the copper itself; good film reduces these effects.

Erosion can be the result of improper film on the commutator or a wear condition such as threading. ...

The condition of the commutator film directly affects friction and erosion and thus brush life. In order to achieve good brush life, the commutator must have good film.
...
When electric current is passed between the carbon and copper in the presence of water vapor, a microscopic layer of copper carbon composite or film, is formed. Good film is chocolate brown or burnished bronze to dark brown or black and uniform in color. It is not bright copper or burnt black copper. Consult a commutator color and appearance picture chart to determine the condition of the commutator. There is a condition known as false filming in which brush graphite deposits become cooked on the commutator resulting in an appearance similar to dark film. Oil can also leave a coating which resembles film. If this film can be easily wiped away, it's not the desired good commutator film!

Commutator filming is a continuous process. That is, the film is continuously being formed and stripped away. A good film is only 200 nano inch thick (0.000,000,2 inch or 0.000,005,08 mm). Thus the conditions required to build good film must always be present. Changes in current, humidity, etc. will affect the commutator film.

 

[duffy]

Yes, copper. What process would erode a surface film quicker at lower current densities than it would at higher?

 

[killivolt]

Yes, copper. What process would erode a surface film quicker at lower current densities than it would at higher?

White paper : Commutator surface speed.

The coefficient of friction between the brush and commutator increases approximately as the speed. Brush wear is proportional to the coefficient of friction. At higher speeds, above 5,000 or 6,000 fpm, greater brush pressure may be required, resulting in decreased brush life. At high field weakened speeds, commutation deteriorates, that is sparking increases. At higher speeds the film can be stripped from the commutators faster than it is being formed. If the motor runs at high speeds for only short periods of time, film can still be maintained.

Do you know what the speed is ?


kv

 

[dknguyen]

How are you lowering the voltage/current?

 

[Hero999]

Yes, copper. What process would erode a surface film quicker at lower current densities than it would at higher?

Perhaps the fact that the brushes are copper increases wear because the carbon film can't form?

I've always found cheap copper brushed motors to have a shorter life than more expensive carbon brushed motors.

 

[crutschow]

It's a servo, so do the motors have to drive a fixed load at a given speed? Have you measured the current? It's possible that at lower voltages the motor current may actually have to increase to provide the required output power.

 

[dknguyen]

It's a servo, so do the motors have to drive a fixed load at a given speed? Have you measured the current? It's possible that at lower voltages the motor current may actually have to increase to provide the required output power.

Yup. THat's what I was thinking (sort of). I was thinking more along the lines that if the actual voltage were not actually lowered but just PWM'd to lower the average voltage, the average current might decrease but the current peaks might increase since it's spinning slower so less BEMF to oppose the battery voltage while the actual battery voltage remains the same. Basically, it might be running closer to stall than the RPM might imply.

 

[Mr RB]

What anti-arc provisions do you have on the brushes?

 

[shimniok]

Yes, copper. What process would erode a surface film quicker at lower current densities than it would at higher?

It's in the whitepaper that was posted.

 

[duffy]

The current is proportional to the voltage. At 5V, it is an amp, at 3.6V it is 680ma. There is no compensation by the PWM circuit in this Futaba S3154 servo for torque, no current detection. It's a simple H-bridge driver, the only feedback is the pot. Decreasing the voltage proportionally decreases the motor current, brush current, speed and internal temperature.

The motors drive a fixed load (a steering assembly in a model car) on a static life tester and a dynamic load on a test track, there is good correlation between the two. We test 12 on the static tester and 4 on the track continuously to failure. Our sample size is not huge, but we have burned out several dozen so far and have some confidence in our estimations of the mean time between failures. Servos are disassembled and inspected to verify the failure mode is burned out brushes.

The speed is a function of power through the device. It operates at maximum speed until the error signal nears zero. The maximum speed is proportion to power, which is proportional to voltage. Operating at lower voltages drives the device at lower speeds.

The motor is not cheap, it has an integrated MOV on the rotor (that black ring) which extends motor life considerably over ones that do not have it.

Low voltage motors do not typically use carbon brushes, copper brushes perform better with low voltages and high current, carbon brushes are better for high voltage and low current.

 

[duffy]

It's in the whitepaper that was posted.

What, "If the current density is too low, the film will be striped [sic] from the commutator" - ?

That's not describing a process.

 

[shimniok]

What, "If the current density is too low, the film will be striped [sic] from the commutator" - ? That's hardly describing a process.

Seriously. It is in the whitepaper. Read the whitepaper.

 

[Hero999]

I read the whitepaper but it doesn't mention copper brushes.

The carbon copper film theory doesn't make any sense when the motor in question has copper brushes.

 

[shimniok]

Note to self: read the whitepaper! Sorry ... totally missed something.

So they are talking about carbon brushes and copper commuator. Not copper brushes. They're talking about film forming on the copper commutator.

So... I'm a fool and I have no answer

 

[dknguyen]

Are copper brushes even used? Copper isn't exactly a precious metal, furthermore it oxidizes a lot.

 

[killivolt]

I read the whitepaper but it doesn't mention copper brushes.

The carbon copper film theory doesn't make any sense when the motor in question has copper brushes.

Your right I missed that. SA45 carbon brushes are all I see too !


Are the copper ones special replacements?

Or order spec based on need of application?


kv

 

[Hero999]

As mentioned before, copper brushes are used more on low voltage motors.

Low voltage motors do not typically use carbon brushes, copper brushes perform better with low voltages and high current, carbon brushes are better for high voltage and low current.

 

[duffy]

Yes, copper brushes, typical for low voltage motors. I don't have the option to change the brushes or the motor, it's in a servo.

I'm stuck with this particular servo. We designed an image recognition system that locates the cars on the track and predicts the path. Other servos we tried have drift with temperature that throws off that prediction (other servos use RC oscillators, this one uses a crystal, that seems to be the thing). The temperature changes quite a bit depending on how heavily the game is played.

Life is important because operators expect a certain ROI (return on investment) with coin-op games, basically that it pays for itself in a few months with no problems. It's a $20,000 game, 25 laps for $1, 4 player stations (so $1 to $4 per game, depending on how many people are playing) - anyway, it works out to about a half-million cycles we need out of that servo. We're close. We had the mean time between failures over that when we dropped the servo voltage from 5V to 3.7V. Now we're under it again. I lowered the servo voltage to 3.5V. Could lowering that voltage by .2V really cost us some 200k off the life? I'm looking for other explanations - we didn't chart humidity with our life tests, perhaps when we did the earlier tests in the winter, when the humidity was lower...

My question was if anyone had experience or knew anything about this strange notion that lower currents can somehow reduce MTBF on motors. It fits with the tenuous data I have, but doesn't make a lot of sense. Doesn't sound like anyone has any information on this, (figured it was a long shot) but I appreciate the replies.