[Prev][Next][Index][Thread]

Multi-layered Secondary


----------
From:  Alfred C. Erpel [SMTP:aerpel-at-op-dot-net]
Sent:  Sunday, March 22, 1998 7:03 PM
To:  Tesla List
Subject:  RE: Multi-layered Secondary

>
> Hi Ed,
> Just getting to old list mail !!
> Try cutting open a video monitor flyback with a bandsaw - you will
> find the sectioned stacked wiring you suggest below.
> Dale
>
> ....snip follows
>
> "Another thought - it is possible to imagine a multi-layer coil layered
> the 'other way', effectively a stack of concentric spirals going
> alternately in and out, e.g. for 3 layers,the turn numbers being
> something like :
>
>    | |
> 987| |789
> 456| |654
> 321|_|123
>
> It would however require a lot of ingenuity to figure out how to
> actually wind such a beast - any takers?


	 What you are trying to do is essentially stack pancake coils in series
that have been manufactured out of one continuous wire. The way to do this
is to  wind your coil on a solid bar and separate each pancake with an
insulator that will withstand the voltage differential that gets built up
between the ID (inside diameter) and OD (outside diameter) of each coil.
Drill holes down the length of the solid bar parallel and angled at the
pitch necessary to make the wire emerge at the base of the next pancake.
Also, the dielectric strength of the solid bar must be such that the wires
passing through it and parallel to each other are sufficiently insulated to
prevent arc over through the material.
	 Thread the entire length of the finished wound coil back and forth through
each hole leaving the amount of wire for each pancake spooled or folded in a
U shape, depending on its length (what is practical). Now before doing any
coil winding, you will have a solid bar with insulators separated by air
gaps which are equal to width of the wire you are winding, with a loop of
wire available that is exactly the amount required for one pancake stored at
the location of each pancake.  You now rotate the entire system except for
the loop you wish to wind into a pancake. Thus the rotating entire system
winds the disengaged loop into a flat coil. Then you disengage the next loop
of coil, and wind the entire system (except the disengaged loop) in the
opposite direction and form the next flat coil. And follow this method
(alternately winding in opposite directions) until you have wound all the
loose wire. It is not intuitive, but draw it on a piece of paper and you
will see that the pancakes all are wound in the same direction.
		This method is not conducive to verbal instructions, nor is it easy to
accomplish.  If you are not a machinist and reasonable tool designer, forget
it.  However, after you spend about 10 hours designing this fixture, and 45
hours building it, you could wind coils like this very easily, and in not
much more time than a single layer simple helical coil. There is more to
this that I haven't mentioned, like applying the correct enamel that will
flow down between coils and through capillary attraction fill gaps,
essential, possible, and even easy if you do it right, but if you do it
wrong, you will irreparably loose many hours of work.  I don't think the
extra effort is worth the additional gain in inductance achieved over your
method suggested below.


> Keeping the spirals going the same way on each layer (i.e. always
> outward) would be easier to wind, but less neat due to the 'flyback'
> of the wire back towards the centre on each spiral, reducing the
> number of total turns possible."

	(following dimensions in inches)
	I am currently winding a coil exactly like this.
	It is much easier than the method above  (but still requires special
skills/equipment). It requires 4 different types of spacers/insulators.  Two
from .010 polyester (Mylar) and two from .030 polycarbonate (Lexan). The
center line pitch spacing of the pancakes is .081. It is being wound on a .5
diameter ferrite core. Each pancake has an ID of .64 with 12 turns of .028
diameter wire (22 gauge).  There are 70 layers of pancakes. I just completed
the form today (Sunday, my kids think this form with no copper in it is
major cool, my wife is just puzzled) and hope to wind the copper onto it
this week. It is approximately 1.5 OD x 5.7 high.  Anticipated inductance is
approximately .1 Henry. 3.5 ohms resistance. Awesome Q.  Something unknown
(to me) is how much distributive capacitance this type of coil will display.
At 50 picofarads, total capacitance (with top load), its frequency will be
71 kHz.  This works out to a theoretical Q of greater than 12,000!  (2 x pi
x freq x inductance / resistance )
	I am planning for this coil to be run on my desktop, and hopefully battery
powered. And I still have not nailed the exact configuration of the power
supply feeding the primary.   Right now I am trying to adapt the charging
circuit from a disposable flash camera as a power supply for this coil; I
have wound a small transformer and built a small spark gap to kick the 300
volts DC up to 9000 volts but I will probably need help on this (just tried
it this weekend with unclear results).   I will post more on my small coil
as it develops.

		I am aware that many people on this list espouse that polyester is a no-no
in Tesla Coil applications.  I understand how this is true in an
electrostatic application such as a capacitor.  But I don't agree that it
will cause an efficiency problem as an insulator in an electromagnetic
application.  If I am wrong, I will learn it the hard way.

	I have plans for a much larger coils that I am developing using stacked
pancakes wound from .005 x .400 copper strip. This cross section is
equivalent to .055 (16 ga.) wire in resistance.  Also, this thickness
thwarts the skin effect, which makes it much better than 16 ga. wire.



Regards,
Alfred Erpel