Cyndi Nyberg 
Former EASA Technical Support Specialist 

In the April 2007 issue of CURRENTS, we covered surge testing anomalies, speci.cally for AC windings. The surge test can be used for DC windings as well. It can be a useful tool for evaluating armatures and some DC fields. 

A note of caution:  If a winding does not have a minimum insulation resis­tance per ANSI/EASA AR100-2006, it is not safe to apply an overpotential test (surge or high potential). 
Surge testing shunt .elds may not provide meaningful results if the surge pulse decays too quickly — if it dissipates through only the .rst few hundred turns. To obtain a test voltage high enough to test every turn would require too high a voltage. That high voltage would overstress the ground-wall insulation. 

Surge testing armatures 
DC motor or generator armatures can be tested with a surge tester. Al­though the winding of an armature is different than that of an AC stator or wound rotor, it will still be symmetrical if the winding is good. 

Surge comparison testing of an armature, similar to the high-frequency bar-to-bar test, compares two sections of the armature winding, as shown in Figure 1. The surge pulse is sent through one section of the winding, and then an equal pulse is sent through another. The setup has three brushes, or probes. One is common and is positioned in the middle of the other two. The surge pulse is sent from the common to each of the outside probes. The waveforms resulting from the two surge pulses are then superimposed on an oscil­loscope screen and, if they are equal, only a single trace appears. If a second trace can be seen, as shown in Figure 2, there is some inequality in the two sections being tested. This would indicate a winding fault such as a ground, short or open; or the winding has unequal turns and the bars spanned need to be changed to obtain an equal number of turns across the span; or an equalized armature does not have the same number of equalizers in the two sections being compared. 

As stated above, the surge test sends a pulse from the common to the two outside probes. However, the surge pulse also travels from the common all the way around the armature (CW) to the outside of the “left” probe, and from the common around in the opposite direction (CCW) to the “right” probe. Although the strength of the surge pulse will dis­sipate through the armature winding, a short in the armature will still show up on the surge trace. To prove this, try shorting two bars outside the area under test; the trace will show a separation. 

Some surge tester manufacturers offer a high-current booster attachment to drive the surge pulse through the ar­mature windings. One manufacturer's surge tester has a capacitor in the back of the unit with a separate lead position for testing armatures; that supplies the required current for the test. 

Voltage level 
The bar-to-bar test voltage for the surge test should be about 350-500 volts per bar. However, to avoid over-stressing the groundwall insulation, the maximum total surge test voltage should be 1500 volts for armatures rated less than 500 volts, and 2000 volts for armatures rated 500 volts and above. The relatively high voltage peak of the surge makes it quite effective for probing the winding for faults like cracked insulation or ther­mally degraded insulation or insuf.cient voltage creepage clearances. 

Surge comparison test example 
Each set of bar-to-bar test probes must span an equal number of bars for a meaningful test. Otherwise the two compared sections will not be equal. Further, the total number of turns being compared must be equal. If the armature has an unequal turn sequence, then the surge pattern in the oscilloscope may show separa­tion. It may be necessary to adjust the number of bars spanned to obtain a good pattern, as long as the minimum voltage per turn is maintained. 

Also, when an armature is equal­ized, the sides under test must contain an equal number of equalizers. In either case (unequal turns or equaliz­ers) the result is a pattern that changes at regular intervals. Again, adjustment of the number of spanned bars may be necessary. 

Paschen’s law requirements 
The test voltage must also meet the requirements of Paschen's Law (i.e., at least 350 volts per turn) without exceeding the test voltage for the ground insulation. Paschen's Law states that, at a constant temperature, the breakdown voltage is a function only of the product of the distance between parallel plane electrodes. 

So how does this relate to surge testing? Stated a little differently, Paschen's Law means that an applied voltage cannot bridge the gap of two .at plates (i.e., shorted turns) with a certain gap between them until the voltage is raised above a minimum level. This minimum is 350 volts per turn. A surge test, at the proper voltage level, is the testing method that can do this between the turns.
Example: For a 500-volt armature: 
Surge test voltage = (2 x rated voltage) + 1000 = (2 x 500) + 1000 = 2000 volts 

This voltage applies to a new winding as well as one that has been in service, since this is considered a non-destructive test. 

Since we need to exceed Paschen's Law, determine the number of com­mutator bars spanned on each side of the test probes. The volts per bar will be the minimum 350 volts per turn, dividedby the number of turns in the winding. If the number of turns is not known, then assume that it has one turn. Using our 2000 volts, and 350 volts per bar for a 1 turn coil: 

Only 5 bars should be spanned on each side of the probes. Figure 3 shows the setup for this example. The surge pulse is sent from the center through the 5 bars, and then through the rest of the wind-Figure 3. Armature surge comparison test. ing. The armature is rotated slowly by hand, so that all sections of the armature are compared to the others. 

Series fields and interpoles 
While the surge test may not be a useful tool for evaluating shunt .elds, series .elds and interpoles can be surge tested. The maximum surge test voltage should not exceed the AC high potential test voltage — twice the rated (armature) volt­age plus 1000 volts for most DC machines. Use the surge tester to compare each series .eld or inter-pole to the one on each side of it. That way, each one is tested twice. If each pattern on the scope is iden­tical, then the .elds or interpoles are not shorted. 

Synchronous rotors 
Depending on the number of turns in each rotor pole, the same issues arise with surge testing as would be present with shunt fields.

Further information on testing of DC machines can be found in the “Fundamentals of DC Opera­tion and Repair Tips” manual. 

Categories: Alternators/generators, DC motors, Armatures, Motor testing

Tags: Armatures Motor testing Surge testing

Rate this article:

4.7

Print