1. Introduction to Detecting and Testing for Shorted Turns
Inductors are made up of a length of wire, usually wound around a core.
The core is usually some type of magnetic material such as iron or ferrite,
however air cores are also sometimes used.
The wire, is referred to as the “winding”, and this winding is made up of a
number of turns.
Generally, inductors are a single winding, and transformers mostly have
multiple windings, (there are special cases, such as “Auto
transformers” where only a single winding exists) and in many transformers
different wire diameters can be used in the various windings.
Inductors and/or transformers that are wound using a large number of turns
and/or using very fine wire, require a method of detecting the existence of
shorted turns, and also the ability to stress the winding in order to detect
imperfections or potential weaknesses in the winding insulation that could
be a weak point over time.
Winding imperfections can normally be attributed to damaged enamel caused by
physical damage during manufacture of the transformer, or imperfections
during the manufacture of the raw wire itself.
These imperfections can result in shorted turns during normal operation if
undetected at point of manufacture.
The high temperatures resulting from the increased current at the short will
fairly quickly cause copper melting and create a low resistance spot weld.
This low resistance short will then completely short out one turn, affecting
the performance of the winding and hence the whole transformer.
2. Detecting shorted turns
Shorted turns and potential weak areas can be detected with Voltechs AT
testers by two test methods:
SURGE or impulse testing (SURG) - suitable for fine wire, or high voltage
windings.
STRESS WATTS testing. (STRW / STRX) - suitable for line voltage windings.
In both the following cases we will discuss the effects of stressing the
primary, but remember that via basic induction of voltages across all
windings, you will be testing the longevity of ALL windings on the
transformer.
As such you should always stress test the winding with the greatest number
of turns, as this will ensure that you are not inducing more than the
generated voltage on any winding, and hence protect the unit.
2.1 SURGE or impulse testing (100 V - 5 kV DC)
As there is no universally defined method, or measurement parameter for this
type of test, a perfect example component is required for comparison
testing.
The perfect component will benchmark the measured result will be used as a
comparison value.
The voltage level and the number of pulses required is dependent on the
total amount of stress needed on the component’s winding.
As an example, in the event of a lightning trike, a mains-driven transformer
could experience spikes of up to 2kV from the raw-mains supply, therefore
three pulses at a voltage level of 3kV should adequately test and stress the
windings for inter-turn insulation imperfections.
Each injected high voltage pulse will produce a defined characteristic decay
time of the transient voltage.
Poor insulation and/or shorted turns will dissipate some of the energy,
resulting in shorter decay times.
FIGURE 1 (decay versus time of one pulse from a surge test, Left=Good
part, Right = Bad part).
The AT Series “SURGE” test provides a high voltage surge test from 100 V to
5 kV and a choice of 1 to 99 impulses
The test signal is generated by discharging a capacitor into the winding of
the part under test, and then measuring the length of the resonant
relationship between the capacitor (in the AT) and the inductor (UUT)
If multiple pulses are requested by the test program, then once the AT
detects that the resonant pulse has reached zero, it will recharge the
capacitor, and discharge again, for the next pulse.
This takes around 100-200ms between the end of one pulse decay, and the
beginning of the next pulse decay
There is no user-defined time parameter for the impulse and subsequent
measurement, as the decay rate is dependent on the relationship between the
AT Surge generator and the part under test.
The results returned by the AT are presented as a volt-second measurement
(ie the area under the decay graph).
If the transformer is faulty the measured result will be a smaller value
than that of the perfect transformer, as losses will cause a shorter decay
time, and result in a smaller area under the graph.
The SURGE method is preferable over the later STRESS WATT method, as the
higher stress voltages available give better sensitivity to a single
adjacent winding failure. Of course, use of SURGE also requires that the
design of the part can withstand such high pulses even when correctly
manufactured.
Summary
When using this test as a characterizing metric of a transformer design,
parts likely to suffer an early failure can be detected by assessing the
length of resonance against that of the perfect reference part used to
define the limits of the test.
Any parts with hard interwinding shorts, or weak areas (for example in the
enamel coating) will flash over under the stress of the voltage pulse, and
so can be detected and removed from production for rework or scrapping.
2.2 STRESS WATTS testing (1-270 V AC)
A transformer will still draw some current and consume power when testing a
transformer at no load with the secondary open circuit.
This power consumption is measured in watts and is the power absorbed by a
coil subjected to an alternating current.
Typically, the current draw due to core loss (eddy currents and hysteresis)
is only a few percent of the normal load hence is usually negligible.
Watts testing (WATT) is usually operated at the transformers full line
voltage and operating frequency.
FIGURE 2 - WATT test primary 220 V @ 50 Hz, TR 5:1, secondary is 44 V @
50 Hz
However, it is also very common and desirable to “stress” the transformer
above its normal operating voltage to give some margin of quality assurance.
This stress testing (unlike normal WATT testing) should also be performed
over an extended and fixed duration, as weaknesses may not show under
instantaneous conditions.
During this stress period, any instantaneous dramatic increase in measured
power would indicate that an inter-turn insulation winding fault or shorted
turn was present as a larger amount of current would be consumed through the
defect
FIGURE 3 - STRESS WATT test primary 440 V @ 100 Hz, TR 5:1, secondary is
88 V @ 100 Hz
Faraday’s law shows that providing the voltage and frequency is increased
proportionally core loss should remain roughly the same. Therefore, a stress
watt test (STRW) can be performed at twice the voltage rating and twice the
frequency rating of the transformer.
As we have proportionally increased the voltage and frequency from figure 2
to figure 3, the core loss will remain the same allowing the windings to be
stressed at a greater voltage than used in normal operation.
The flux density (B) in the core will remain the same
B ~ V / (f * A * N)
N = The number of turns
A = The cross-sectional area of the core
V = Voltage applied.
f = Frequency applied
Practical notes
In reality, you will find that the core losses do increase with frequency
(core losses are a function of flux density AND frequency) even though we
have kept flux density the same, hence the STRW may read higher, but the
result will still be repeatable and characteristic. You can mitigate the
core losses by doubling the F again, so for a 100 V, 50 Hz transformer, you
may find that 200 V, 200 Hz is more suitable than 110 V/100 Hz.
Line supply transformers typically have a 240V winding, with a tap to give 2
x 120V windings.
To double the voltage on the 240V winding would require 480V, which is
beyond the 270V capacity of the STRW test.
Here we suggest either;
a) Testing the 120 V winding (if you have one) individually at 120 V
(WATT) for normal operation and then at 240 V for the stress (STRW) test.
This will in turn induce 480 V across the 240 V winding, without the need to
supply the 480 V
or
b) Testing a lower voltage secondary, at twice its operating voltage.
Similarly, this would induce 480 V on the primary, but as the primary nodes
would not be used in the test, the 5 KV isolation on open test nodes will
protect the AT tester.
Summary.
The AT5600 and AT3600 provide a stress watt test (STRW) from 1 V to 270 V @
20 Hz to 1500 Hz to detect potential faults in the inter-turn insulation of
a winding.
The user must also specify a dwell time for the test from 0.5 s to 180 s,
over which the power is continually monitored.
The test results are presented in Watts.
Where voltage and current levels require extending, please use Voltech’s AC
Interface fixture with the AT.
This allows use of either an external step up transformer, or an AC power
source to generate higher voltage (up to 600 V) and current (up to 10 A)
The test signals, measurement, and pass-fail criteria are still
automatically controlled by the AT using the 4 X tests; MAGX, WATX, STRX
and VOCX.