Chuck Yung
EASA Technical Support Specialist
When an armature is rewound, there is always a slim chance that it may be connected incorrectly. If two coil leads are switched, or if the error results in an armature where each coil closes on itself, normal tests will detect the problem. The trouble arises when the misconnection results in a uniform winding. When that happens, the result may be—in effect—an accidental redesign for a different voltage.
The number of parallel circuits in an armature winding can be changed simply by shifting the top lead position. As with a 3-phase winding, doubling the circuits halves the design voltage. If a wave-wound armature is rewound with the same data but connected lap simplex, or if a lap simplex armature is connected lap duplex, the circuits have been doubled. The same is true of a wave simplex armature reconnected wave duplex.
Since the winding is uniform, it will pass all the normal electrical tests: surge test, bar-to-bar, and growler. The first indication of a problem occurs when the assembled motor is test-run. The symptom: it runs either “way too fast” or “way too slow.” Because armature speed is proportional to the applied armature voltage, a misconnected armature is likely to run at a predictable change in speed. When misconnection changes the number of armature circuits, the speed change is roughly proportional to the change in circuits.
Only one bar off
For example, if the top leads of a lap simplex armature are shifted off by only one commutator bar (commutator pitch of 1-3 rather than 1-2), the resulting lap duplex armature will run twice as fast at rated voltage (Table 1, Row 1). Miss the correct position of the top leads by a single bar in the other direction, and the armature changes from progressive to retrogressive (Table 1, Row 6), or vice-versa. In that case, the armature polarity is reversed, so the only correction needed is to interchange the brush holder leads.
Sometimes the placement of random wound armature leads can be confusing. If an armature has an odd number of slots and bars, there is a good chance it is wave wound. An even number of slots and bars is more likely to be lap wound. There are many exceptions to this, but it’s a good quick check. It’s a lot easier to figure out the correct winding style before we put the wrong one in the armature.
A wave wound armature, misconnected lap simplex will run twice as fast as its rated speed.
The speed ratio also doubles from lap simplex to lap duplex. At no-load, a shunt-wound motor will usually run approximately 10% above nameplate speed. So the no-load speed resulting from an accidental circuit change will fall between 2 and 21/2 times rated speed. Of course, the reverse is true of an error that halves the circuits. Never assume that a lap wound armature is connected lap simplex.
In effect, a lap simplex armature misconnected as a lap duplex has been “redesigned” for half the nameplate voltage. When operated at the original nameplate voltage (twice its “redesigned” voltage), it runs about twice as fast as the nameplate base speed. If the error is duplex to simplex, the speed will be roughly half of base speed.
To confirm that the armature has been misconnected, it may be necessary to lift the top leads and ‘light out’ the ends of a single coil. This can be done using an ohmmeter, a test-light, or any other device set up for this purpose (an audible ringer, for example.) A faster alternative, which requires the use of an accurate low-resistance ohmmeter, is to measure and compare the resistance between bars 1-2, 1-3, and so forth. If there are two separate circuits (even-numbered bars in one circuit, and odd-numbered bars in another), the armature is connected duplex. If the resistance between bars 1 and 3 is lower than between bars 1 and 2, then the armature is connected lap duplex. If the data card indicates it should be connected lap simplex, then the winder probably missed the placement of the top leads.
Locating a grounded coil
When a ground occurs during the process of rewinding an armature, visual inspection is rarely enough to find the problem. To pinpoint the location of a grounded coil in an armature, there are a few methods to simplify the task. We’ll start with the least destructive method first.
If the ground fault resistance is low enough to measure with an ohmmeter, measure the resistance from each bar to ground. The bar with the lowest ground resistance is connected to the grounded coil.
Since the grounded coil is at ground potential, another method for finding it is to measure the voltage from each bar to ground. For this test, use a growler to energize the armature winding, and measure the voltage between each bar and ground. When measuring the voltage from the shaft to each bar, the closer to the grounded bar, the lower the voltage will be. If the grounded coil is a multi-turn coil, the actual voltage won’t be zero unless the grounded turn is the one connected to the bar, but it will be lower than the remaining bars.
The third method, which will damage the ground insulation, is to use an AC hipot to apply continuous voltage through the ground until visible smoke identifies the location. Before applying hipot voltage, try using a 110-volt test lamp connected to ground and to the commutator. It may pass enough current to cause the fault location to emit smoke. While this method damages the slot insulation at the site of the ground fault, the insulation was already damaged— or it would not be grounded. Magnet wire enamel is also likely to be further damaged by the hipot. Nonetheless, it is usually possible to re-insulate the conductors and repair the ground insulation.
Locating a shorted coil
When testing confirms that a newly rewound armature is shorted, it is natural to first suspect the solder connection at the commutator risers. A little excess solder is all it takes to inadvertently bridge bars. The growler induces enough current that solder shorts of this nature are often blown clear (think of this like a small fuse carrying too much current). If the bridge of excess solder is large, or if the short is located elsewhere, it is necessary to pinpoint the actual short. With the growler and an AC voltmeter, measure the induced voltage between adjacent bars. Shorted turns will have the lowest voltage potential, so the voltage between the bars connected to those turns will be zero.
If the shorted turns are connected to adjacent bars, the short is probably located in one of these areas:
- At the back of the risers
- In the V-ring area
- The leads just behind the risers
- In the slot, if the shorted coils share a slot
If the growler method failed to clear the solder shorts, use a good light and a magnifying glass to inspect between the bars and behind the risers. A great armature winder’s tool is a dental pick. Use it to scrape the mica along the back of the suspect risers. Use a hacksaw blade to scrape the mica surface between the bars.
Use growler, hacksaw blade
If the armature is not equalized*, use a growler and hacksaw blade to identify the slots involved, and mark them. The growler will energize the armature winding and the hacksaw will vibrate when it is over slots with shorted coils. If three slots ‘growl’, check the coil pitch against the slot positions. You’ll probably find that the shorted coils share one of those three slots. Raise the top coil side from the slot and retest. If the short disappeared, inspect and re-insulate the coil.
If the shorted coils do not share a slot, trace the coil leads to determine the areas where the coil leads cross, and inspect those areas. Gently lift the top leads away from the bottoms and retest.
If that cleared the short, re-insulate those leads.If all else fails, tightly band the commutator brush surface, and remove the outer v-ring. Inspect the mica cone, and the underside of the 3o angle under the bars, for excess carbon or flux that could be the source of the short. Note: Bar-to- bar testing of the commutator before rewinding an armature is important. If the v-ring is being pulled off now it means it wasn’t tested when it should have been – prior to rewinding.
One armature problem that cannot be detected by routine tests is the “cold solder” joint, or high- resistance connection. The procedure for checking an armature is similar to that for a 3-phase stator. Measure and compare the resistance. The use of a digital low-resistance ohmmeter (DLRO) or Ductor allows comparison of resistance between adjacent bars (or adjacent bars in each path, if du- plex-connected). As with a 3-phase winding, resistance should be identical within 5%. Be sure to factor in odd turns and equalizers, when present. This test is also useful for detecting open equalizers, when that fault is suspected.
At a minimum, by quartering the commutator and comparing the resistance through each quarter, it is possible to quickly determine whether or not the armature has poor connections between the risers and coils.
Do it right the first time
Now that you’ve experienced the hassle of locating and repairing that shorted/grounded armature, it’s a good time to review the critical points in armature winding. First, the commutator should be tested after the armature has been stripped and cleaned. If it needs to be refilled, find that out before inserting all the new coils. Before winding the armature, test the commutator again.
Testing the commutator should include a hipot test at the armature test voltage, and bar-to-bar test light at either 110 or 220 volts AC.
After the bottom leads are installed, visually check them over. If color-coded sleeving was used, it takes only a minute to verify that nothing has been reversed. Make sure the leads are evenly spaced with no sloppy areas. Double-check the correct placement of the first top lead installed, then check it again. Once the top leads are all inserted, double-check the sleeving color sequence.
Solder or TIG weld the armature leads to the risers, then repeat all the electrical tests.
Megohmmeter the windings, then band the armature and test it again. After the banding material is cured, but before the armature is impregnated (dipped or VPI’d), test it again.
* Lap wound armatures are often equalized, especially those 10 hp or larger. Each equalizer connects bars of equal electrical potential (180 degrees opposite for 4 poles, 120 degrees apart for 6 poles), so the growler and hacksaw blade test responds as if the entire armature is shorted. Use an ac voltmeter to measure the induced voltage between adjacent bars. A good armature will have a clearly discernable pattern consistent with the turns per coil (odd turns) or equalizer spacing (if equalized).
Print