ROTOR RSO REFLECTOMETER

TYPE TDR100 and TDR100RB

INSTRUCTION MANUAL

Issue 11 April 2014

Process Tomography Ltd.,

64, Courthill House,Water Lane, Wilmslow, Cheshire. SK9 5AJ. UK
Phone/Fax +44 - (0) - 1625 - 418722
Email: [email protected] Website: www.tomography.com

Copyright Process Tomography Ltd. and Convex Design Ltd. 2014

 CONTENTS

1. Principle of Operation

2. Operating Instructions

3. Interpretation of Results

4. Delay Line Test Unit

5. Use of Digital Oscilloscopes

6. References

7. Note on analogue oscilloscopes

APPENDIX

1 Additional information for TDR100RB version

2 Use of padded case

SAFETY WARNING

The use of this equipment on a rotor installed in an operational generator must

be carried out with the explicit permission and under the supervision of the
local plant operator. All local safety rules and procedures must be complied
with.

In particular, the equipment must only be connected to the generator rotor

after the field supply has been disconnected and isolated in accordance with
local safety regulations. Failure to comply with this instruction will damage the
equipment and may endanger both the the plant and the operator.

2
 FRONT VIEW OF TRDR100RB ROTOR REFLECTOMETER

REAR VIEW SHOWING MAINS INPUT AND BATTERY FUSE HOLDER

3
1. INTRODUCTION

1.1 GENERATOR ROTOR WINDING FAULTS

Figure 1.1 A typical generator field rotor (courtesy of GE Power Systems)

Large high-speed electrical generators use a rotating magnetic field produced by a

rotor in the form of a cylindrical electromagnet having either 2 or 4 magnetic poles*.
The rotor body is a solid steel forging containing radial slots for the coils which make
up the electromagnets (rotor windings). The turns of the coils are rectangular copper
bars insulated with an epoxy material and in a 2-pole rotor, there are typically 8 pairs
of slots for each pole of the electromagnet, with each slot containing up to 20
conductor turns. A cross-section of a typical radial slot (in this case, containing 15
turns of insulated copper bar) is shown in figure 1.2.

Figure 1.2 Cross-sectional view of a radial slot containing the rotor field
winding. (courtesy of GE Power Systems)

4
At the ends of the rotor body, the turns pass from the end of one slot to its equivalent
slot on the other side of the magnetic pole and are held in place in the end regions
by steel end rings. A direct current of typically 3000 amps flows through the rotor
winding to produce the magnetic field, which is at right-angles to the axis of rotation,
with clearly-defined north and south poles.

Figure 1.3. Examples of coil insulation breakdown (courtesy of GE Power

Systems)

A 2-pole rotor rotates at 3000/3600 rpm to produce a 50Hz or 60Hz alternating

voltage in the (3-phase) stator windings. The rotor windings experience large
centrifugal forces, which can damage the insulation, leading to either shorts between
the rotor winding and ground or between adjacent turns as shown in the figure
above. As the DC current is large and the short circuits will have finite resistance,
large quantities of heat can be generated at the fault location and this can cause
damage to the remaining insulation, resulting in severe damage to the rotor windings.
Short circuits can also cause magnetic imbalance, giving rise to increased vibration
levels.

Generator rotors are routinely tested to detect these types of fault, usually during
construction and also before and after routine generator maintenance. One standard
test method used is time-domain reflectometry. However, unlike the similar technique
used for testing transmission lines, a custom test instrument (reflectometer) is
required, because the rotor winding is a very imperfect transmission line and
produces a large number of reflections at each change in impedance between the
sections of conductors inside the radial slots and the sections in the cross-over end
regions.

* Note that the design of rotors for large hydro-electric generators is different from that
described above as they rotate at lower speeds and have multiple sets of magnetic poles.
However, the windings can be tested in the same way as for high-speed rotors.

5
1.2 OVERVIEW OF TEST METHOD

The method used to test rotor windings for earth faults or shorted turns relies on the
fact that the rotor winding is symmetrical. For example, a 2-pole rotor contains two
nominally-identical half-windings, one for the North pole and the other for the South
pole, both of which are connected in series. A four pole rotor is similarly symmetrical.

This symmetry property is used to compare the response of the 2 halves of the rotor
winding to a short voltage pulse applied between each slip-ring and the rotor body.
The pulses and any reflected signals are monitored at each end of the rotor winding
using an oscilloscope. If the rotor winding is fault-free, two identical waveforms will
be observed at each slip ring. However, if one half-winding contains a fault, the two
waveforms will differ. The test details are described below.

A (typically square wave) pulse (12V) is applied between one of the rotor slip rings
and ground and the transmitted pulse received at the remote end of the rotor and the
reflected pulse at the sending end are monitored using a dual-trace oscilloscope as
shown below.

A pair of adjustable matching resistors are used to test the rotor winding under
repeatable conditions and are normally set so that the pulse generator and
terminating resistor match the characteristic impedance of the rotor winding (typically
values in the range 30 - 1000 Ohms).

Figure 1.4 Rotor winding test method

The tests are carried out by applying pulses from each end of the rotor winding in
turn and the oscilloscope traces are recorded and compared. If the rotor is fault-free,
the oscilloscope traces will be identical.

In practice, the TDR100 test instrument has a switching circuit which applies pulses
alternately from each end of the rotor winding so that the waveforms are
automatically superimposed when viewed on a single channel oscilloscope.

Note: The easiest way to gain familiarity with the RSO test method is to use the
Reflectometer with the Demonstration Delay Line, which is supplied with the
equipment and as described in section 4. New users should follow the instructions
given in section 4 to acquaint themselves with the test method before attempting to
carry out the test on a real rotor winding. Further detailed information about the test
method is given later.

6
1.3. MEASUREMENT DETAILS

Electrical faults in generator rotors fall into two main categories, faults from the
winding to the rotor body ('earth faults') and faults between parts of the winding
('inter-winding faults'). The existence of an earth fault is detectable with a simple
multimeter. A single earth fault on a rotor is frequently tolerated and many generators
run in this condition (preferably with some form of alarm system to detect the onset
of a second earth fault). The existence of an inter-winding fault is not easily detected
by simple electrical methods. However, a rotor winding with a serious inter-turn fault
will frequently overheat, leading to more serious local damage, or display excessive
mechanical vibration and may have to be taken out of service.

The rotor Reflectometer uses a technique known as time domain reflectometry. The
application of this technique to testing rotors is known as the RSO (recurrent surge
oscillograph) method by power engineers in the U.K.

The method involves applying a D.C. voltage step between one end of the rotor
winding and the rotor body. The reflected wave at the input end of the winding and
the transmitted wave at the far end of the winding are monitored using two
oscilloscope channels. If the voltage step is applied from each end of the rotor
winding alternately, then two oscilloscope traces will be obtained which may be
superimposed on the oscilloscope screen. A healthy rotor winding will have two
identical traces. A rotor with a fault will have differing traces and the positions of the
fault may be deduced by scaling in the time domain.

Figure 1.5 Reflectometer operating principle

The basic Reflectometer system is shown in Fig. 1.5. A pulse generator supplying a
12V pulse of variable length at a repetition rate of up to 500Hz is connected via a
500Ω variable resistor to an electronic changeover switch S1 synchronised to the
pulse repetition rate. The changeover switch enables the rotor to be excited from
each end of the winding in turn, alternate pulses exciting the rotor from opposite
ends. The rotor is terminated in a second variable resistor R2 via the changeover
switch. The pulse generator, synchronous changeover switching network, matching
resistors and terminals are all contained within the Reflectometer unit. Two
oscilloscope channels monitor the voltage at the input end of the rotor (channel C1)
and at the output end (channel C2). The values of R1 and R2 are chosen to match,
approximately, the characteristic wave impedance of the rotor winding, to eliminate
reflections of the pulse at each end of the rotor.

7
The pulse repetition rate and pulse length are adjustable by means of three controls
on the front panel of the Reflectometer. The synchronous changeover switch first
excites the rotor at end 1 and the pulse propagates along the rotor winding, emerges
at end 2 and is absorbed by R2. The switch then operates and the next pulse excites
the winding at end 2, propagates through the winding to end 1 and is again absorbed
in R2. The changeover switch returns to the first condition and the sequence is
repeated continuously. Hence successive pulses from the pulse generator excite the
rotor from each end in turn and energy is always supplied via R1 and absorbed by
R2. The changeover switch is arranged to operate approximately half way in time
between successive pulses so that the operation of the changeover switch does not
adversely affect the leading edge of the pulse. A single channel oscilloscope
connected between S1a and earth as shown will therefore display two traces
corresponding to the signals applied to each end of the rotor.

1.3.1 Measured waveforms for a fault-free rotor winding

Input ends Output ends

Figure 1.6 Typical oscilloscope traces for a fault-free rotor winding

Figure 1.7 Simplified oscilloscope traces for fault-free rotor winding

8
The voltage waveforms at the input and output ends for a typical sound rotor are
shown in Fig. 1.6 and simplified versions of these traces are shown in Fig. 1.7. A
sound rotor will appear to be symmetrical with respect to either slip ring and
therefore, the two traces that either C1 or C2 monitor will be identical and will be
superimposed on the oscilloscope screen. The pulse will take a finite amount of time
(the transit time) to travel from the input end of the rotor to the output end. As a
result, the traces monitored by C2 will display zero voltage for this period (the transit
time) and the transit time may therefore be measured directly from the C2 traces.

1.3.2 Measured waveforms for a rotor winding with an earth fault

When an earth fault occurs part way along the winding, the traces that occur are
shown as measured, in Fig. 1.8 and in simplified form in Fig. 1.9.

Figure 1.8. Short circuit to rotor body at end of 5th coil slot
(16 coils in winding)

Figure 1.9 Simplified input end oscilloscope traces for rotor winding with earth
fault

9
At the short circuit to earth, the input pulse is reflected with reverse polarity and when
it returns to the input end, a decrease in voltage is observed. Assuming that the fault
is not exactly in the centre of the winding, the reflection will occur at different
positions for the two traces. The traces will therefore diverge as shown in Fig. 1.6.
The trace that is deflected first corresponds to the end nearest to the fault. A rough
estimate of the position of the fault may be found by noting the time to the fault as
indicated by the input trace ( t 2 seconds).

By linear scaling, the fault will be approximately t2 / ( 2t1 ) x 100% of the winding
from one end. However, the apparent propagation velocity of the pulse through the
rotor winding is not uniform and care must therefore be exercised in locating faults by
this method.

The best method for locating the position of the fault is outlined in Section 2 (c). This
involves removing the rotor from the generator and probing the winding in the end
region or down the radial cooling holes if these exist. The detection and location of
interwinding faults may be carried out in a similar manner.

1.3.3 Measured waveforms for a rotor winding with an interturn fault

If there is an interturn fault, the waveform at the slip ring nearest the fault is
characterised by a slight increase in voltage followed by a decrease down to a
minimum, followed by a slow voltage rise, as shown in the figure below.

Figure 1.10. Measured waveforms for a rotor winding with a single shorted
turn.

10
The shape of the traces for an interturn fault can be explained by considering the
effect on the incident pulse of a single shorted turn, as shown in the figure below.

Figure 1.11 Simple representation of a shorted turn

When the voltage step reaches the short circuit between the turns, it can take one of
3 paths in the forward direction, instead of a single path in the fault-free case. The
impedance that the voltage step sees looking in the forward direction will therefore
be Zo/3, where Zo is characteristic impedance of the winding

This causes the transmission line to appear unmatched at this point and a proportion
of the voltage will be reflected with opposite polarity to the incident voltage step,
leading to a decrease in voltage when observed at the input end.

However. the rest of the voltage step will propagate away from the short circuit and
two of the three paths available (round the shorted turn) will return the voltage step to
the point of the short circuit. Part of this voltage will then be returned to the input end
of the winding leading to an Increase in voltage. The part of the voltage step which
travels round the shorted turn will do so continuously, causing the energy to be
returned to the main rotor winding over an extended period of time.

Further details of the waveforms resulting from various fault conditions are shown
and discussed in section 3.

11
2. OPERATING INSTRUCTIONS

2.1 DETAILS OF EQUIPMENT SUPPLIED

The followings items are supplied with a standard TDR100 Reflectometer

measurement system:

TDR100/RB Rotor Reflectometer unit with mains lead

DL100 Demonstration Delay Line (inc. leads)

Terminal magnets (3)

5m 3-core test lead terminated in a connection module.

3 x 3m single core leads (Brown, Green, Blue) terminated in 4mm insulated

banana plugs (Red, Green, Blue) at one (connection module) end and insulated
crocodile clips at the other (rotor) end.

2 x 1.2m Coaxial oscilloscope leads

1m Delay line 4mm plug lead (red, blue, green)

This Instruction Manual

Padded Carrying Case

2.2 MEASUREMENT OPTIONS*

There are several modes in which the Reflectometer can be used:

1) Testing stationary rotor in generator.

2) Testing rotor at speed in generator.

3) Testing rotor when removed from generator.

The most straightforward case is when the rotor is at rest in the generator and the
test method for this will be described in detail. The other test modes are based on
this technique with suitable modifications. The measurement system connection
diagram is shown in Figure 2.1 and assumes the use of a basic two channel
analogue oscilloscope, although a single channel instrument will suffice if a 2-
channel oscilloscope is not available. Suitable low-cost instruments are listed in
section 7.

Note that it is also possible to use a digital oscilloscope under some circumstances
and the use of this type of oscilloscope is described in section 5.

* Please refer to Appendix 1 for additional information about the use of the
TDR100RB unit which contains internal rechargable batteries.

12
Note that the colours in parentheses eg (BROWN) refer to the connecting leads,
while the abbreviated colours eg R (RED) etc. refer to the 4mm banana plugs and
terminals.

Figure 2.1 Measurement system connection diagram and image

13
2.3 METHOD FOR TESTING A ROTOR AT REST WHILE INSTALLED IN THE
GENERATOR

SAFETY WARNING

The use of this equipment on a rotor installed in an operational generator must

be carried out with the explicit permission and under the supervision of the
local plant operator. All local safety rules and procedures must be complied
with.

In particular, the equipment must only be connected to the generator rotor

after the field supply has been disconnected and isolated in accordance with
local safety regulations. Failure to comply with this instruction will damage the
equipment and may endanger both the the plant and the operator.

2.3.1 Rotor Winding Connection Module and test leads

The connections between the Reflectometer and the Generator rotor winding are
made using the 3-core 5m Rotor test lead and Connection module supplied with
the equipment. This lead has 4mm insulated banana plugs at the Reflectometer
end and a Connection module terminated in 4mm insulated terminals at the rotor
end. Three 3m insulated single core leads connect this module to the rotor winding.
The Connection module and test leads are shown in figure figure 2.2 below.

Figure 2.2 Rotor test leads and connection module

14
Note: The connection module provides simple 1:1 connectivity between the banana
plugs at the reflectometer end and the output terminals on the module. This
arrangement has been used to allow connections to be made to the rotor windings of
large generators, where there may be significant distances between the slip rings
and the rotor shaft earthing point. It also allows damaged connecting leads to be
repaired or replaced easily on-site, or for customers to use their own connecting
leads if preferred.

The connections to the rotor slip rings and earthed shaft can be made using the
contact magnets supplied and shown in figure 2.3.

Figure 2.3 Contact magnets and keepers.

2.3.2 Setting up the test equipment

1. Isolate and make safe the generator stator winding.

2. Either isolate the field brushgear from the field supply (both sets of brushes), or
remove all of the brushes from each brushgear cage, ensuring that none of the
brushes touch the slip rings. For a brushless generator, isolate the generator field
winding from the rotating rectifier unit (both leads).

3. Connect the banana plugs of the individual 3m leads to the output terminals of the
Connection Module. Match the plug and terminal colours (red to red etc.). At this
stage, do not connect the other end of the 5m test lead to the Reflectometer
terminals.

4. Clean an area of rotor shaft adjacent to the slip rings with emery cloth, followed by
a degreasing solvent and wipe off with a clean rag. Remove the magnetic keeper
and attach one of the terminal magnets supplied (see figure 2.2 ) to the shaft at this
point.

5. Connect the crocodile clip of the green conductor of the 3m Green test
lead to the screw stud on this magnet.

15
6. Attach the crocodile clips of the brown and blue 3m test leads to each end of the
rotor field winding as follows:

i) If it has been possible to isolate the brushgear cages from the field supply,
then simply connect these leads to each brushgear cage assembly (clip the
crocodile clip on to one of the brush braids).

ii) If the brushes have been removed, clean a small area on each slip ring with
degreasing solvent and attach the two remaining magnets to the slip rings.
Attach the brown and blue leads to the terminal studs on these magnets
using the crocodile clips.

iii) For the case of a brushless generator, clip the brown and blue leads
directly to the up-shaft field winding leads after the isolating links have been
removed.

Note: Do not connect the other end of the 5m lead to the Reflectometer at this
stage.

7. Using an electrical test meter, measure the rotor winding resistance between
the red and blue banana plugs at the Reflectometer end of the 5m rotor test lead.
This should be typically less than one ohm, including the resistance of the leads. If
the measured resistance is greater than one ohm, check the
contact resistance between the clip ends of the brown and blue leads and the field
winding. If magnets are being used, remove them and reclean the slip ring and
magnet faces if necessary. Record the measured winding resistance.

8. Check the contact resistance of the earth magnet to the rotor shaft by
measuring the resistance between the green banana plug at the Reflectometer
end and a point on the rotor shaft near the magnet. If the resistance exceeds one
ohm, reclean the shaft and the magnet face and repeat until a low contact resistance
is obtained.

9. Using an electrical test meter, measure the insulation resistance of the rotor
between either one of the the red or blue banana plugs and the green banana plug.
A healthy rotor will have an insulation resistance in excess of 1MΩ, although if the
winding is damp, this may be reduced to 10KΩ or less. Record the insulation
resistance.

10 Now connect the red and blue plugs of the 5m test lead to the same colour
terminals on the Reflectometer front panel (slip ring 1 and 2 terminals) and connect
the green plug to the green earth terminal on the Reflectometer (see figure 2.4).

16
 Figure 2.4 Reflectometer front panel controls (TDR100RB)

11. Connect the coaxial BNC leads to the oscilloscope input channels

Figure 2.5. Oscilloscope Lead

12. Connect the coaxial BNC Oscilloscope leads, shown in figure figure 2.5 from
the oscilloscope input channel type BNC terminals to the reflectometer
osciloscope BNC terminals as follows:

Connect one coaxial lead between the oscilloscope channel 1 input and the
reflectometer oscilloscope input ends connector. Turn the BNC connector on the
coaxial lead clockwise to lock it into position.

Similarly, connect the other coaxial lead between the oscilloscope channel 2 input
and the reflectometer oscilloscope output ends connector.

The system connectivity should now be as shown in figures 2.1 and 2.6.

17
 Figure 2.6. Measurement system connectivity

2.3.2 Testing the rotor winding

1. Set the controls on the Reflectometer initially as follows:

R1= R2 = 100Ω

PULSE FREQUENCY : Fully clockwise

PULSE WIDTH SWITCH : Centre position
PULSE WIDTH POTENTIOMETER : Mid scale
PULSE OUTPUT SWITCH : Auto
R1/2 TOGGLE SWITCH : Up (0 - 500 Ohms)

2. Set the oscilloscope controls initially as follows:

DISPLAY : Channel 1
VERTICAL SENSITIVITY : 2V/CM (Both channels)
TRIGGER CONTROLS
- MODE : Normal
- SOURCE : Channel 1
- LEVEL : Positive
- SLOPE : Positive
- COUPLING : D.C.
- TIME BASE : 20 µsec/CM

18
3. Switch on the Oscilloscope and the Reflectometer. Adjust the oscilloscope trace
position and triggering controls until a stable voltage step is displayed on the
oscilloscope screen. This is the waveform at the input ends of the rotor winding and
should resemble one of the traces shown in Fig. 2.7 (a) to (c).

Further detailed information about these waveforms is given in section 3.

Figure 2.7 Typical oscilloscope waveforms

4. Switch the oscilloscope to display channel 2 or, if a single channel oscilloscope

only is available, connect the oscilloscope input lead to the ‘OSCILLOSCOPE
OUTPUT END’ terminal of the Reflectometer. Adjust the oscilloscope controls until
a trace similar to that shown in Fig. 2.7(d) is obtained. This trace shows the
voltage step received at each remote end of the rotor winding. Adjust the
oscilloscope timebase switch (and, if necessary, the Reflectometer pulse width
controls) until the single pass transit time [t1 in Fig. 3.1 (d)] occupies less than half of
the trace width. Record this transit time in micro-seconds.

19
5. Switch the oscilloscope back to display channel 1 and adjust the pulse width
controls on the Reflectometer unit so that the trailing edge of the voltage step can
be seen at the far right of the trace as shown in Figs. 2.7 (a) to (c). This is a
convenient way of displaying the zero voltage level. Now adjust R2 so that there is
no reflected signal from the ends of the winding.

Fig. 2.7 shows three cases:

a) R2 matched (no reflection)

b) R2 too large (positive reflection)

c) R2 too small (negative reflection)

The reflection is seen at the input ends t2 seconds after the start of the voltage step
where t2 is approximately twice the single pass transit time (t1).

Note that the range of the Matching Resistors R1 and R2 can be increased using
the R1/R2 switch.

If the switch is in the UP position, the range is 0 - 500 Ohms.

If the switch is in the DOWN position, the range is 500 - 1000 Ohms.

If the toggle switch is down when correctly matched, add 500Ω to the dial reading to
obtain the value of R1/R2.

6. Having matched the output ends correctly, note the value of R2 (adding 500Ω if
necessary), which is the average characteristic wave impedance of the rotor
winding. Now set R1 = R2 to complete the matching at the input ends. This
eliminates the possibility of multiple reflections from one end of the rotor to the other.
It may now be necessary to adjust the oscilloscope vertical sensitivity controls to
optimise the trace size relative to that of the oscilloscope screen.

7. If the rotor winding is perfect, two perfectly superimposed traces will be

displayed on the screen. If this occurs then the rotor winding can be safely assumed
to be fault-free. To verify the existence of the two traces, push one of the trace
identifier buttons, when one of the traces should be displaced vertically towards the
zero voltage level, showing the two traces. If only one trace is displayed check that
the PULSE OUTPUT mode switch is in the AUTO position.

8. If two perfectly superimposed traces are not obtained, there may be a fault in the
rotor winding. Section 3 explains in detail the trace shapes to be expected for
various types of faults.

20
9. Record the traces by photographing the oscilloscope display using a digital
camera (with the flash turned off), then edit the image using an image editing
program such as the freeware Irfanview, available from: http://www.irfanview.com/

An example of a screenshot obtained using this method is shown below.

Figure 2.8 Screenshot of traces for Delay line with simulated interturn fault.

2.3.2 ADDITIONAL OPERATING MODES

In addition to operation in Auto mode, it is possible to apply pulses to either slip ring
1 or slip ring 2 only. This option can be useful when using a digital oscilloscope to
monitor the traces as described in section 5. The operating mode is controlled by the
3-way Pulse Operation switch as follows:

Position 'END1' - pulses are injected into slip ring 1 only.

Position 'END2' - pulses are injected into slip ring 2 only.

Position 'AUTO' - normal operating mode.

NOTES

1. Mode switch

For normal rotor testing ensure that the switch is in the 'AUTO’ position, as the single
trace produced when the switch is in the 'END1’ of 'END2’ position will not indicate a
winding fault.

When in the 'AUTO’ position always check that two traces are present by using the
'Trace Identify’ buttons. If only one trace is shown when one button is pressed,
check and adjust the triggering of the oscilloscope (particularly when using a digital
oscilloscope).

21
2. Operation of the case handle.

The case handle can be rotated by pushing in the black buttons at each handle pivot
point as shown below:

Figure 2.9 Handle pivot buttons (one on each side)

Depress both buttons to allow rotation of handle

22
2.4 METHOD FOR TESTING ROTOR AT SPEED

CAUTION

Any testing carried out on a rotating rotor must be carried out with extreme
care and with the explicit permission and under the supervision of the local
plant operator. All local safety rules and procedures must be complied with.

1. The Reflectometer may be used to test an unexcited rotor at speed. This is

particularly useful for detecting and locating faults that are speed-dependent.

The most useful information is obtained if the test is conducted either while the rotor
is being run up to speed from rest, or while it is run down to rest from synchronous
speed. The method is essentially the same as for testing a stationary rotor except of
course that it is necessary to make contact with moving slip rings and shaft earth
connections.

2. If the brushgear cages can be isolated from the field supply, then connections can
be made to the slip rings via the brushgear. However, in most cases it is not
possible to isolate the cages. In these circumstances it is necessary to remove all of
the brushes from the cage and install some insulated brushes that have been
previously prepared. In anticipation of the test, three brushes per slip ring should be
removed from the cages and marked so that they may be reinserted in the positions
from which they have been removed. The brushes should be machined
undersize, and strips of Tufnol or other suitable insulating material glued to the faces
of the brushes using an epoxy resin adhesive. The insulated brushes should then
be re-machined down to the correct size for reinsertion.

Experience has shown that it is necessary to use brushes that have been in service
in the machine and which have been passing current. This technique does not work
if new brushes are used because these give very poor contact with the slipring. The
insulated brushes are installed in the machine prior to the test, and the brushes in
each cage are commoned taking care not to let the brush braids touch the cage
assembly. Connections are then made from these insulated brushes to the
Reflectometer.

3. It is necessary to make a separate earth connection to the rotor shaft. The most
effective method appears to be to clean an area of the shaft and to hold a short
length of heavy duty stranded earthing cable against the shaft. This may be done by
stripping off the last few centimetres of insulation from the cable and taping the cable
to a short length of dowel so that the temporary earth brush may be held safely in
contact with the shaft. It has been found by experience that it is not satisfactory to
use an existing shaft earth brush for this test because of the large amount of
electrical noise generated by these devices.

4. With the modifications mentioned above, the test may be carried out as the rotor
speed is increased or decreased. The equipment should be set to monitor the traces
at the input ends of the rotor and should be watched carefully for any changes in one
trace which will indicate a speed dependent fault.

23
2.5 METHOD FOR TESTING ROTOR WHEN REMOVED FROM GENERATOR

1. The test method is basically the same as that outlined in Section A. However, with
the rotor removed from the machine, further tests may be possible.

2. If the rotor end rings are removed, the shape of the traces obtained may differ
considerably from that for a rotor with the end rings in-situ. Moreover, because the
windings can expand radially in the absence of the end ring, two slightly different
traces may be obtained for a rotor that is known to be fault-free, because the
expansion of the end region windings may not be uniform. In general, the effect of
removing the end rings increases the characteristic impedance of the rotor
considerably.

3. If a winding fault has been detected in the rotor, and the end rings have been
removed, it is possible to find the approximate location of the fault by putting a similar
fault onto the other half winding of the rotor and moving the position of this deliberate
fault until two identical traces are obtained. This may be done by using insulated
probes.

If an earth fault is suspected, then one of the probes should be earthed to the rotor
body using a short flexible lead, and the end winding should be probed until the
application of this fault causes similar traces to appear. The faulted coil may be
found by touching the probe onto the outer turn of each coil in the end region of the
winding. When the coil which causes the traces to almost coincide has been
located, the faulted turn may be located by moving the probe radially down this coil in
the end winding region and making contact with the sides of the conductors, which
are not usually insulated. When the turn has been located which causes the traces
to coincide (or nearly so) its coil number and turn number (found by counting turns
down from the outside of the winding) should be noted. The fault lies in the
equivalent coil in the other half winding.

4. It is possible to use this same technique without removing the end rings if the
rotor contains radial cooling holes that run next to the conductor slots. In this case,
the winding may be probed directly.

5. The position of an inter-turn fault may be located by using two probes connected
via a length of flexible lead. In this case, adjacent turns of the opposite half-winding
are shorted together to locate the fault. An alternative method is to make up a
special two-pronged probe to apply the shorts between adjacent turns. An example
of this type of probe is shown below.

Figure 2.10 Probe for locating shorted turns

24
3. INTERPRETATION OF RESULTS

3.1 Fault-free, Healthy Rotor Winding

Fig. 3.1 shows the results to be expected when a healthy rotor is tested.

Figure 3.1. Typical waveforms for a fault-free rotor winding

Fig. 3.1 (d) shows the voltage step received at the far end of the rotor, and t1 is the
time for it to propagate through the rotor. This is termed the single pass transit time
and is typically 20-100µS.

Fig. 3.1 (a) shows the traces at the input ends of the rotor and displays two perfectly
superimposed traces (except in the region after the end of the voltage step, which is
of no relevance to the test). This indicates a healthy rotor winding.

25
Fig. 3.1 (c) shows the input end traces when the output ends are terminated in a
near short circuit (R2 = O). The top of the voltage step is seen to start to decrease in
amplitude after a time t2 seconds, which is the time for the pulse to pass through the
winding once and then back again.

Similarly Fig. 3.1 (b) shows the input end traces when the output ends are terminated
in an open circuit. In this case, the top of the voltage step starts to increase after t2
seconds.

It will be seen that the input and output matching resistors have no effect on the
individual traces apart from changing their shape. It is not possible to produce two
different traces for a healthy rotor by maladjustment of the Reflectometer or
oscilloscope controls and hence the possibility of misinterpreting traces caused by
operator error is greatly reduced. Always check for two traces by pressing the trace
identify switches.

3.2 Trace indicates a Faulty Winding

There are several basic modes of failure of a rotor winding and some example traces
are shown in figure 3.2.

Figure 3.2 Typical Traces for various winding faults

26
a) Fault between winding and rotor body (earth fault)

A full or partial fault may occur between the winding and the rotor body. An example
of the type of trace obtained in this case is shown in Fig. 3.2 (a). The voltage step
which is injected from the end nearest the fault is seen to increase to a peak at the
fault and then to decay rapidly. The voltage step injected from the end furthest from
the fault increases and decays some time later, as the voltage step injected from the
remote end takes a longer time to reach the position of the fault. It should be noted
that the sharpness and rate of decay of the second peak is considerably less than
that for the first peak. This is an example of the general rule that the resolution of
the measurement for this test is greatest near the point of injection of the voltage
step (i.e. at the slip rings) and decreases as the position of the fault moves further
into the winding.

The approximate location of the fault may be found by comparing the time from the
start of the voltage step to the peak of the trace with that to the end of the winding
when R2 is set equal to zero as shown in Fig. 3.2 (c). This must have been done
prior to the fault occurring and highlights the need to record the traces of 3 (a) and 3
(b) for the rotor while it is in a fault-free state (ideally when it is first commissioned)
to act as a reference 'fingerprint’ for future use.

b) Interturn Fault

Fig. 3.2 (b) shows the type of trace obtained when a short circuit between adjacent
turns occurs. The trace corresponding to the end nearest the fault increases slightly
then decreases and finally increases to meet up again with the trace injected from
the other slip ring end. The faulty trace is that which gives rise to the lower part of
the first major loop shown in Fig. 3.2 (b). The approximate location of the fault may
be found by similar means to that described for the earth fault case. It should be
noted that the test is particularly sensitive and will detect a relatively high resistance
(a few ohms) interturn fault that may, in practice, not carry current in an operational
state. Further tests involving measuring the voltage drop across adjacent turns by
passing a large direct current through the total winding must be carried out to
determine whether the fault is current-carrying or not.

c) High resistance joint in rotor

The effect of a high resistance joint on the winding is shown in Fig. 3.2 (c). The
apparent characteristic impedance of the end of the winding nearest the fault is
increased relative to that of the other end of the winding. Hence the upper trace
shown in Fig. 3.2 (c) corresponds to the end of the winding nearest the fault. It is
possible that the fault may be caused by a high resistance joint at one slip ring. In
this case, the fault may be confirmed by placing a variable 0 - 500Ω resistor in series
with the lead to the other slip ring and adjusting this variable resistor. If it is possible
to make the traces coincide by these means, then the fault occurs very close to the
first slip ring.

27
d) Interwinding fault

It is possible for faults to occur between rotor slot coils and the upshaft leads which
connect the slip rings to the ends of the rotor winding. Fig. 3.2 (d) shows an example
of a fault of this type, in which the upshaft lead had shorted to the fifth coil in the
winding, effectively shorting out the first five coils in a total winding of sixteen coils.
The lower trace corresponds to the slip ring nearest the shorted coils.

e) Other causes of non-identical traces

Apart from these common faults, there are circumstances where, in applying the
test, problems occur which may indicate that a good rotor is faulty. These may be
caused by:

i) Poor contact between the slip rings and the test leads. If this occurs, the
traces will resemble Fig. 3.2 (c), the difference between the traces being
dependent on the magnitude of the contact resistance. If this occurs, re-
check the contact between the test leads and the slip rings before assuming
that the rotor winding is faulty.

ii) Poor contact between the earth lead and the rotor shaft. Again, the cure is
to re-check the contact resistance.

iii) The characteristic impedance of both rotor half windings are not identical.
On the face of it, this seems most improbable. However, it may be caused by
a previous repair to one half winding using insulation different from that used
during manufacture. Moreover, when the end rings are removed, the end
windings expand radially in a non-uniform manner, causing the impedances to
differ. Consequently, two slightly different traces are nearly always obtained
when one or both end rings are removed.

28
4. USE OF REFLECTOMETER DELAY LINE TEST UNIT

Figure 4.1 Delay line type DL100

The delay line unit is used to check that the Reflectometer is operating correctly and
is also an aid to demonstrating and understanding the test method. The unit consists
of a 10 section lumped component delay line. The characteristic impedance of the
delay line is 100Ω and the propagation time for a single pass through the unit is
11µS. The junctions between each section of the delay line are connected to a
series of 2mm sockets, enabling external connection to these points. The input and
output ends of the unit are connected to 4mm sockets.

Figure 4.2 1m delay line connection lead

In use, the Delay line is connected to the Reflectometer using the 1m 3-core test
lead supplied and shown above in figure 4.2. At the Delay line end, the red banana
plug is connected to the red input terminal on the delay line, the blue banana plug
is connected to the blue output terminal and the green banana plug is connected
to the green common terminal.

29
At the Reflectometer end, the red banana plug is connected to the red (Rotor slip
ring 1) input terminal on the front panel, the blue banana plug is connected to the
blue (Rotor slip ring 2) output terminal and the green banana plug is connected to
the green ground terminal. The overall connection arrangement is shown below.

Figure 4.3 Use of Delay line test unit *

The oscilloscope controls should be set up as described in Section 2 of this manual.

Set the oscilloscope vertical sensitivity to 1V/cm, the horizontal sensitivity to 5µS/cm
and both Reflectometer matching resistors to 100Ω.

With the oscilloscope monitoring the input ends of the delay line, adjust the pulse
width so that the display resembles that shown in Fig. 4.4 (a). The effect of
mismatching the output ends of the delay line may now be demonstrated by
changing the output matching resistor, R2 on the Reflectometer. Fig. 4.4 (c) shows
the traces obtained when R2=0. The double pass transit time is seen to be 22µS
and Fig. 4.4 (d) shows the traces obtained with R2=500Ω.

The effect of a simulated earth fault may be demonstrated by shorting one of the
delay line junctions to earth using the 2mm plug lead supplied.. Fig. 4.5 (a) shows
the result of shorting junction 4 to earth.

Similarly the effect of a shorted turn may be demonstrated by shorting out one or
more delay line sections as shown in Fig. 4.5 (b).

NOTES

The impedance and single pass transit time of the unit supplied may vary from the
figures quoted above.

* The delay line common terminals and the common lead plug are now GREEN
instead of black.
30
Figure 4.4 Delay line oscilloscope traces with no faults applied

Figure 4.5 Typical delay line traces with faults applied

31
5. USE OF A DIGITAL OSCILLOSCOPE

In principle, it is possible to use a digital oscilloscope instead of an analogue

instrument to display the reflectometer waveforms. However, most budget digital
oscilloscopes have limitations which may prevent them from being used effectively
with the TDR100 range of reflectometers. This section describes some possible
ways for coping with these limitations.

5.1 OVERVIEW

One of the main uses of the TDR100 measurement system is to display live
waveforms during a winding repair, so that the winding state can be monitored
continuously and the onset or clearance of faults seen immediately. To do this, the
oscilloscope must display simultaneously the two waveforms corresponding to pulse
injection from each end of the rotor winding.

The TDR100 produces alternating waveforms at the input to one oscilloscope

channel corresponding to the pulses injected from each end of the rotor winding. To
view them correctly, the oscilloscope must trigger on the leading edge of each
alternating pulse waveform. Most analogue oscilloscopes can do this routinely.
However, many digital oscilloscopes have difficulty displaying signals where
successive waveforms are different.

Digital oscilloscopes often use some form of waveform averaging to reduce the noise
on the displayed traces but if this is used, only one of the waveforms can be shown
at any single time. Although averaging can usually be turned off, the result is
waveforms with a high level of superimposed noise. Even with averaging disabled,
the osciloscope triggering may still be unable to track the alternating waveforms
produced by the TDR100 equipment.

A number of options are available to overcome some of these problems and these
are described in the following sections.

The instructions given are based on the use of the Tektronix TDS 1000/2000
oscilloscope equipment range and should be modified accordingly for other
oscilloscope models.

5.2 OPTION 1. USE OF AUTO MODE

It may be possible to use the TDR100 Auto mode in the same way as described for
an analogue oscilloscope, by careful adjustment of the Frequency control on the
TDR100 unit. We suggest that familiarity with the technique is gained using the
demonstration delay line initially, as described below:

Apply a shorting link between terminals 5 and 6 of the delay line to simulate an
interturn fault and connect the delay line to the TDR100 system as described in
section 4.

Connect channel 1 of the oscilloscope to the TDR100 input end oscilloscope

terminals ((yellow and black) and set the reflectometer controls as described in
section 4. Then set channel 1 to display the input end waveforms, as shown in figure
5.1.

32
 Figure 5.1 Digital oscilloscope waveforms at input ends using delay line with
simulated inter-turn fault

The following comments (applicable to Tektronix TDS 1000/2000 series) may be

helpful in setting up the oscilloscope initially:

Press the Acquire button and select Peak Detect from the ACQUIREscreen.

If necessary, disable all other channels by pressing the appropriate channel select
buttons until all traces other than that for channel 1 are no longer displayed.

Move the channel 1 baseline to +1 horizontal graticule line from the bottom of the
screen using the channel 1 offset control knob.

Adjust the maximum waveform amplitude to + 7 graticule lines from the bottom of
the screen using the channel 1 gain control knob.

Expand the horizontal time scale to show all of the waveform using the horizontal
sec/div control knob.

Move the start of the waveform to the +1 vetrical graticule line using the horizontal
position control knob.

Adjust the frequency control on the TDR 100 unit until 2 waveforms are displayed
as shown above.

Using this technique, the 2 input end waveforms can be displayed and compared
directly and continuously. However, in the case of the TDS oscilloscope range, these
2 waveforms cannot be saved directly, as the save image function apears to save
one of the individual waveforms, not the displayed image. It is also not possible to
reduce the image noise by averaging in this option.

33
5.3 OPTION 2. USING THE SINGLE END INJECTION MODE TO CAPTURE THE
WAVEFORMS

Overview

As an alternative to alternating end pulse injection, the TDR100 unit can inject
pulses at one end of the rotor winding only.

If this is done sequentially at each slip ring, it is then possible to use the digital
oscilloscope to capture and store these two waveforms in internal memory and then
compare them.

The advantage of this method is that it is possible to save the two measured
waveforms as .bmp files directly to a USB memory stick. It is also possible to reduce
the noise levels in the images using the averaging option

This can be carried out as follows:

Equipment Set-up

Connect the TDR oscilloscope input ends lead to the Oscilloscope input
channel 1.

Press the ACQUIRE button on the oscilloscope and set the screen options to
Average 128 samples.

Capture and save the Slip ring 1 waveforms

Set the TDR100 mode switch to END1 (Inject pulse at slip ring 1 only).

Press the SAVE/RECALL button on the oscilloscope and set the screen options to:

Action: Save waveform

Save to: Ref
Source: Channel 1
To: Ref A

Press the (on-screen) Save button

The slip ring 1 end input waveform will be saved to Reference A in the oscilloscope
memory.

34
Capture and save the Slip ring 2 waveforms

Set the TDR100 mode switch to END2 (Inject pulse at slip ring 2 only).

Press the SAVE/RECALL button on the oscilloscope and set the screen options to:

Action: Save waveform

Save to: Ref
Source: Channel 1
To: Ref B

Press the (on-screen) Save button

The slip ring 2 end input waveform will be saved to Reference B in the oscilloscope
memory.

View the 2 waveforms

Press the Channel 1 menu button to switch off the live channel 1 display.

Press the (white) REF MENU button on the oscilloscope

Use the (on-screen) buttons to turn on (display) the stored Ref A and Ref B
waveforms as shown in the figure below:.

Figure 5.2 Captured Screen image showing the two reference waveforms.

Save the waveforms

Press the SAVE/RECALL button and select the Save image option using the on-
screen button.

Insert a USB memory stick into the oscilloscope USB socket.

Press the (on-screen) Save button to store the image as a bit-map (.bmp) file.

35
6. REFERENCES

Additional information on testing generator rotors can be found in the following

documents:

1. GE Generator Rotor Design, Operational Issue, and Refurbishment Options,

Ronald J. Zawoysky, Karl C. Tornroos, GER 4212, 08/01, GE Power Systems.
Schenectady, NY

2. Rotor winding short detection, Wood, J.W. and Hindmarch, R.T., IEE
Proceedings, Vol 133, Pt. B, No. 3, May 1986, pp 181-190.

Copies of both of these papers are included on the documentation CD supplied with
the equipment.

7. NOTE ON ANALOGUE OSCILLOSCOPES

The following model has been tested with the TDR100 unit

Isotech ISR 622 (20MHz)

A similar higher performance mode (with a brighter trace) is also suitable

Isotech ISR 6051 (50MHz)

36
 APPENDIX 1

TDR100RB OPERATION

The TDR100RB is similar in operation to the TDR100. The main difference is that
the TDR100RB contains a maintenance-free 12 volt 2.5AH rechargeable Nickel
Cadmium battery pack, giving over 15 hours continuous use under average
operating conditions. There is also an additional switch on the front panel to turn on
the instrument for both battery and mains operation. Note that the equipment can be
used while the battery is being recharged.

To recharge the battery, connect the Reflectometer to a suitable 240V or 110V

50Hz mains supply via the rear mounted IEC socket and turn on the Mains switch
on the rear panel. The green 'CHARGE’ indicator on the front panel will light showing
that the unit is on charge.

To turn on the instrument to carry out a rotor test, switch on the Supply ON switch
on the front panel. The red Supply ON indicator will light showing that the unit is in
operational mode. The TDR100B unit can be operated either on its internal batteries
only, or with a mains supply connected to the instrument.

With the front panel 'SUPPLY’ switch set to the 'OFF’ position the battery can be fully
charged from a fully discharged state in 14 hours.

NOTES ON BATTERY MAINTENANCE

The internal battery pack may be damaged if it is allowed to discharge fully. If the
unit is not used for long periods, we suggest it is given an overnight charge once per
month..

The internal battery-charging circuit operates as a trickle charger and should not
damage the battery if the Reflectometer is left connected to a mains power supply
for long periods. However, to maximise battery life, the mains supply should be
disconnecetd from the unit after a maximum charging period of no more than 24
hours.

37
 APPENDIX 2

USE OF PADDED CASE

The TDR100 units are supplied with a padded case containing pouches for the
instruction manual, leads, magnets and oscilloscope adapters.

The case can be used in one of 2 modes, transit and storage.

In transit mode, the unit is inserted into the case with the handle vertical and the top
cover passed through the handle and attached to the velcro fastening as shown in
figures A2.1 a and b below. The unit can then be carried by the handle.

(a) Handle vertical (b) Cover passed through handle

Figure A2.1 Padded case in transit mode

For storage purposes, the handle can be tilted forwards and the cover passed over it,
as shown in figure 2.2 below. This protects the handle from knocks.

(a) Handle forwards (b) Cover over handle

Figure A2.2 Padded case in storage mode

38