Previous articles have
described the operation and construction of the cut out. To recap, it is simply
an electromagnetically operated switch having two windings. The shunt winding
has many turns of thin wire, is connected across the dynamo output and is used
to close the contacts when the dynamo rotates sufficiently quickly. The series
winding has few turns of heavy wire, is connected in line to the ammeter and
battery and is used to force the contacts open when the output of the dynamo
falls.
Recently, a
failed cut-out became available as a candidate for deeper investigation and
possible repair. It was a type CF1 from a Chummy, but the principle is exactly
the same for all common types on the Austin Seven. It was known to be faulty as
the shunt winding was open circuit, and not the anticipated 33-39 ohms. The
conclusion was that as it was already faulty, dismantling was no further loss;
if possible, a repair would be a bonus.
The cut-out was stripped
down so that the central bobbin could be accessed. Note that there is a mixture
of 2BA and 3BA nuts, using solder as a crude anti-shake technique.
The shunt winding is
underneath the series winding. Therefore, the series winding had to be removed
to examine it. The series winding was carefully unwrapped; not in order to save
the wire, as the insulation flakes off when it is bent, but in order to count
the turns and note the direction of winding. The series winding wire was
measured as being 1.65mm in diameter: this corresponds to 16SWG (Standard Wire
Gauge – Note: NOT American Wire Gauge). The turns were counted: 17 per layer, 3
layers which, allowing for a bit of bending at the start and finish, gives a
total of 50 turns. The bobbin was marked to show the direction of winding.
Having exposed the inner
shunt winding, the cotton retaining thread was removed and the unwinding began.
The inner wire is 0.22mm in diameter, approximating to 36SWG, (35SWG is nearer
but no longer easily obtainable). A quick calculation based on the wire
thickness gives the number of layers at about 36. Each layer has 140 – yes they
were counted! - 140 turns. This equates to a grand total of 5000 turns containing
over 100 metres of copper wire ready to spring off the bobbin. At this point, it
became clear that re-wind without a coil winding tool, or at least a lathe with
a low torque slipping clutch, is simply not an option to the home restorer.
However, a piece of very good fortune showed the break to be at the end of the
top layer. It was concluded that the magnetic field strength wouldn’t be reduced
by too much by simply removing the one damaged layer.
The damaged end of the
winding was insulated from the remaining layers by some thin PVC carpet binding
tape. The connecting lead to the terminal post was re-soldered, having first
removed the varnish with some 600 grit emery paper. The joint was covered by a
second piece of tape and the connecting leads reinsulated with heat shrink
sleeving.
New winding wire can be
obtained mail order from Maplins. As stated above, reuse is not possible due to
the insulation flaking off. Moreover, it is not possible to re-wind using kinked
wire – and 16SWG is very stiff and difficult to form once bent. A temporary
mandrel was made up from a coach bolt, so that the payoff reel could be held in
a vice. Starting at the stud mounting end, and leaving a tail long enough to
reach the A terminal, the wire was rewound, carefully ensuring that the
direction and number of turns was correct and that the windings were tightly
packed. This is essential if the correct number are to be rewound.
Even so, the
irregularity of the outer diameter of the shunt winding made this quite
difficult. The 16SWG has a tendency to spring back to its delivered diameter,
resulting in loose turns. These must be corrected as winding progresses. The
wire is quite stiff and requires considerable hand pressure to prevent it
unwinding as the grip on the bobbin is relaxed. No matter how carefully this is
done, the winding diameter inevitably increases due to imperfect packing. Once
the 50 turns had been rewound, the top end was cut and re-soldered to the
terminal near the armature. The start of the winding must be insulated using a
short piece of heat shrink sleeving to prevent chafing against the armature
support frame.
The completed re-wound
mechanism can then be returned to the insulating plate. Whilst dismantled, it is
a good idea to clean under the square headed screw as this forms one of the
contacts to the points. Finally, a blob of solder on each screw acts as an
electrically conducting locktite.
In conclusion, if an
open-circuit cut-out needs to be repaired, it may be possible to identify the
break in the shunt winding, repair it and re-wind the series coil. However, if
the break in the winding has not been located within the first four or five
layers, the winding should be entrusted to a professional having the tooling to
make a proper repair.
For expert re-winders, the
winding specification is:
1. First winding: 5000
turns 36SWG, 140 turns per layer, 36 layers. Start and finish at bottom of
bobbin. Wind anticlockwise as viewed from Armature. Start and finish to be
connected to flexible tails and identified. (Start = Earth, Finish = “D”)
2. Inter-winding
insulation: thin PVC or similar
3. Second winding: 50
turns, 16SWG, 17 turns per layer, 3 layers. Start at bottom leaving 10cm tail to
connect to “A”; finish at top and connect to armature contact. Wind
anticlockwise as viewed from Armature.
Why so many turns?
The shunt winding as an
impedance of about 36 Ohms, which means that at 6V approximately 160mA flows in
the winding. Thus it produces 0.16 x 5000 Ampere-Turns of magnetic attraction.
In order to counteract this, the force in the series winding must be equal but
opposite. As the winding has 100 times fewer turns, this would occur at 100
times more current, about -16 Amps. The reason the cut-out operates at a lower
reverse current, as seen on the ammeter, of somewhere between 5 and 10 Amps is
that the spring on the armature acts to assist the unlocking action; hence the
“drop out” current is significantly lower than the theoretical 16A.
This article,
written by Geoff Hardman, originally appeared in CA7C Seven Focus in Jan 2007
pp18-20.
