Ozone gas from brush sparking 2

[BobM3]

I was once told that the sparking between the commutator and the brushes of a motor produced ozone gas and this gas is electrically conductive. Every once in awhile I run across a badly sparking motor that flips circuit breakers (non-ground fault type). I was wondering if the gas might be conducting current to ground?

 

[jraef]

Hmmmm, 1/2 truth, 1/2 myth.
Electrical arcs do produce ozone by causing oxegen molecules (O2) to recombine into O3. It is corrosive and dangerous to breathe, but conductive? Never heard that. It is probably a myth brought on by the high levels of ozone you can smell hanging around shortly after a big electrical flash or a lightning strike. It is however an effect, not a cause. The badley sparking motor was probably ready to pop the circuit breakers because it was badley sparking. Any ozone smell that you detected at the same time was a result of that sparking, not the cause.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[jraef]

"...oxygen..."
Wish this forum had editing capabilities...

 

[UKpete]

What can happen in large dc motors where the commutation is poor (heavy sparking at the brushes) is that a flashover occurs, particularly at high speed. This is where an arc flashes between the brush arms around the surface of the commutator, and results in a high current surge from the supply. The heavily ionised air can cause the breakdown to earth between the brusharm connection and the frame where the airgap is smallest. This is an instantaneous and relatively violent phenomenon, not like a sustained leakage current. Sparking commutators without major flashover will not cause fault currents to earth.

The motor can often be restarted after the supply is restored (the current surge will cause a trip) and continue to operate, but it will be more likely to flashover again because of damage to the commutator, brushes and brushboxes. Frequently the motor will have to be taken out of service and have the commutator reskimmed and any damaged components replaced.

Ozone is not the same as ionised air, the person who told you this has got the two confused.

 

[BobM3]

Flashover sounds like it might be happening here. The circuit breaker sometimes trips during dynamic braking (large surge current, very visable arc). It's worse when lowering a large load (high speed at the motor and much energy to absorb). Cleaning the commutator helps for awhile. Resurfacing the commutator helps for a longer while.

When the load is being lowered (motor being powered) I can see arcs traveling about 1/5 of the distance around the commutator between the 2 brushes. It's not hard to imagine that the arcs extend from brush to brush during the dynamic braking.

I'm a bit confused about why the circuit breaker would trip during dynamic braking. The hot side of the AC supply is turned off (by a thyristor). I'm guessing the violent arcing and currents must be turning the thyristor on briefly. I'm going to mount a current probe off the thyristor leg and see if the current is passing through it.

 

[MikeHalloran]

Just an odd thought...

I'm wondering if flushing the brush area with an inert gas like argon during dynamic braking would improve the situation.



Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[UKpete]

Mike, whilst an inert gas will reduce the possibility of arcing, the problem with using it in a dc motor is that they are usually through ventilated for cooling, with large quantities of air. You could have a totally enclosed motor inert gas filled, with a secondary cooling circuit, but I have never heard of it. It would be expensive, and effectively you would just be masking the problem - a correctly designed, operated and maintained dc motor should never flash over in the first place.

Bob, you haven't mentioned the application of your motor, or the motor rating; when the load is being lowered, is this a sudden step change? I only have experience of series and separately-excited dc motors, not shunt or compound which may have different characteristics. But flashover is likely to occur under the following conditions:
1. sudden changes in motor conditions such as a voltage tap change, load change, or weak field step
2. operation at high speed
3. operation at very weak field (related to 2. above), particularly at moderate to high armature currents (referred to as low F/A ratio i.e. the field mmf compared to the armature mmf)
4. field failure (momentary or otherwise)
5. inappropriate choice of brush
6. poor brush contact force (weak springs, brushes worn too short, brushes jamming in boxes, brush pigtail connections snagging)
7. badly worn brushes (becoming loose in their boxes, or worn so short that the tamped connection is showing on the running face)
8. poor commutator condition (poor profile, incorrectly finished commutator surface), including poor mechanical stability of the commutator and localized electrical erosion of the commutator
9. excessive temperature on commutator due to overloading
10. excessive clearance between brush box and commutator, though not so likely

That's all I can think of at the moment, there will be others. If you are a good customer you may get assistance from a carbon-brush sales engineer - they have usually seen it all and can suggest a change in brush grade or spring. A less dense brush, split brushes and constant force (eg Tensator type) springs, a change in brush thickness - were helpful with problematic motors on occasions. Unfortunately there were also motors that were very difficult to get right , for reasons unknown, and the operator just had to accept a more intensive maintenance regime.

You should not be getting sparking around 1/5th of the commutator under any conditions. You can tell a lot by looking at the condition of the brushes and commutator, but if you can actually see that level of sparking you already know there is a problem. What is the commutator appearance? Is there evidence of brush vibration (evident from wear marks on the brush sides where they slide in the brush box).

Perhaps you could provide more details.

 

[BobM3]

UKPete -
The motor is a universal and used on our winches. At rated winch load the motor puts out about 1.3 hp, pulls 17 amps and turns at 11,000 RPM. When raising a load we push the motor well into saturation. "Armature Reaction" is significant and severely shifts the magnetic nuetral plane. Adding iron around the field helps to reduce the sparking at the commutator.

When lowering a load the motor draws very little current. The load helps drive the motor to a very high speed (25,000 RPM). Dynamic braking is initiated by turning off the applied power and making a series connection with the armature (polarity reversed), field and a dynamic brake resistor. The current quickly ramps to a peak and then decays over about 100 ms. It's a great brake for decelerating the load. Very harsh though. When watching the winch it does indeed look like a "step" function.

The commutation deteriorates (as judged by the sparking when lifting a load) when running with large loads (high currents) on our winches . When the commutation problem gets to a certain point we start having problems with the dynamic braking. The braking will start fine with the current quickly ramping up to its peak and then the current decay will start. After maybe 5 to 20 ms the current in the brake resistor drops off suddenly and gets vary erratic. It will last from 5 to 10 ms and then recovers. There is a loud harsh sound from the motor when this happens. Sometimes the breaker flips. I had thought that the alternistors I was using in the circuit for turning off the power and reversing the polarity of the armature were turning on due to the electrical noise and dv/dt. Current probes have since shown that they are staying off. I believe the disturbance during the braking and the loud noise is the flashover you mentioned earlier.


I had thought this problem was with my new control and the alternistors I was using. I hadn't seen blown breakers or heard harsh sounding brake events when using our existing control (with switch contacts). I'm going to put our exisiting control on this particular winch and see if the breaker still breaks and the brake current still goes whacko.

I know our motor needs work with commutation. My focus right now is the new control and making sure it is not causing the problem or being damaged by the motor. Pointing out the flashover problem has helped quite a bit and has allowed me to open up to checking other possibilities.

 

[UKpete]

Bob, yes it's a different type of motor to what I had imagined, my comments above assumed a larger machine. It sounds too small to have interpoles.

As you motor in one direction and generate in the other direction then the brush neutral is shifted to the same position in both cases (for a particular value of armature current). So I suppose that if possible it may be worth moving the brushes back against the motoring rotational direction to get minimum sparking in both motoring and generating modes.

I'm not familiar with electronic controllers and alternistors (I didn't even know what they were until just now) so I wouldn't like to speculate what is going on with your current variations during braking. I wonder if it would be possible to bring the braking in earlier to prevent it reaching such a high speed, very high speeds don't give very good commutation. Good luck with it anyway.

 

[BobM3]

I am going to try to keep the speed lower. Do you know of any good commutation books? Practically, what is it about a worn commutator the causes the sparking to get worse? Does the brush bounce? Contact resistance change? DI/dt goes higher?

 

[UKpete]

There are a lot of texts on commutation but they are all rather old and out of print, you would need to go to a good technical library. I only know the books by UK authors.

There are two main types of commutator: moulded and built-up. The former have copper segments moulded into a plastic or composite base, these are cheap to produce but only used in small motors because they don't have a high current capability. The built-up commutators are used in larger machines and are constructed from copper segments with mica separators, held in place against centrifugal forces by steel vee-rings insulated with micanite. The manufacturing process is skilled and lengthy, and requires special tooling.

The following comments largely apply to the built-up type of commutator, as I don't have experience of small dc motors. A worn commutator will cause sparking if there are any undulations in the profile of the commutator. It isn't always obvious why an undulation may occur in a particular place on a commutator, it may be mechanical instability of one or more of the bars, or a bearing problem, or a problem with one of the armature coils. A commutator does grow dynamically at speed, and it's profile can also change dynamically.

Once an undulation starts to form (it may be less than 20µm and only one or two bars wide), sparking will occur and there will be erosion of the commutator film at that point. Depending on the type of motor and the application, the pit will slowly but steadily deepen and widen, it may end up being up to 0.5mm deep and spread over a tenth of the circumference of the commutator. It is an instability that some designs of dc motors are more prone to than others, even if the materials and manufacturing processes are the same. The brushes do bounce and the affected bars will loose their uniform healthy brown appearance and start to show black burn marks, possibly only along one edge of the bars, possibly with characteristic tiny bright scratches that are erosion tracks caused by sparking. The brushes will wear quickly and may spit from the trailing edge with audible crackling, the brush faces become smoked particularly along the trailing edge, and the brush sides will show wear where they have been vibrating up and down in the brush boxes. Eventually the motor will flash over.

If they have reached that point, as you have discovered they can only really be put right by removing the armature and re-skimming the commutator to restore the profile.

If the motor then goes back into service and the problem re-occurs, a replacement motor may work. If it is occuring across a number of identical motors, perhaps the first thing to rule out is mechanical instability of the commutator - if you have a commutator that has a poor profile on the brush track but is ok off the brush track (i.e. at the outer end of the commutator) then you know that the commutator is mechanically stable. Profile can be measured either by special commutator profile instruments (mechanical or electronic), or with less accuracy by a clock gauge with a mushroom head (the standard tip will tend to jam and bounce over the gaps between bars).

The motor could also be operating outside its design limits, in terms of speed and current.