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

Re: Solid Toroid Beneficial?

Hi Gary,

At 04:03 PM 03/12/2000 -0600, you wrote:
>I have been working on a solid state drive for a Tesla coil for many years.
>Standard wisdom has been that sparks will be inferior to those produced by
>conventional Tesla coils (with a spark gap), but explanations given always
>seemed a bit vague to me.  I have been trying to combine theory and
>experiment in such a way that they will be readily understood by electrical
>engineers, even if my sparks are not as long as those of conventional TCs.
>I think I am getting close to understanding what is going on, but hit
>something the other day that surprised me.  Perhaps someone can help me past
>the mental block.
>
>I am testing a variety of coils and toroids, but for purposes of discussion,
>I will describe one coil and two toroids.  The coil is 14 ga magnet wire
>spacewound on a polyethylene form, diameter 39.64 cm, winding length 116.62
>cm, 387 turns.  One toroid is half spun aluminum, 6 inches by 24 inches. I
>assume it was made by starting with a flat disc of aluminum perhaps 40
>inches in diameter, and turning the edge down and back in until the outer
>portion looks like a toroid.  I set it on the coil so the bottom lip is
>level with the top turn and only an inch or so away.  It would definitely
>form a shorted turn at the top of the coil.  The other toroid is made from
>about 15 lengths of quarter inch copper tubing spaced about a quarter inch
>apart.  The center is open, so there is no place for eddy currents to flow.
>Size and capacitance are almost identical to the half spun toroid, so
>resonant frequency is basically the same.
>
>The coil is driven directly at the base by a square wave from an IGBT
>inverter. Think of an extra coil being driven by a function generator.  The
>power rating is on the order of 50 kW for short periods.  I have actually
>seen a peak power of 33 kW at the point of spark initiation.  It has the
>capability of CW operation, but because of thermal and circuit breaker
>constraints I mostly use disruptive operation.  Power is applied for say 4
>to 10 ms for 1 to 10 times per second.  Sparks look like standard disruptive
>TC sparks, white, thick, up to 3 ft long in air.  A coil this size with a
>properly sized top load should produce sparks at least 10 ft long.  The key
>is the peak power applied.  A conventional TC driven by a pig with an
>average power of 10 kW would be supplying peak powers to the coil of at
>least 100 kW during discharge.  Other things being equal, a peak power of
>100 kW will always produce a longer spark than 33 kW.  It might be possible
>to build solid state drives capable of pushing hundreds of kW into a TC, but
>it would be a challenge!  So my inverter is not going to win any spark
>length competition, but it is nice for making measurements on coils.

200000 watt solid state inverters can be made from IGBTs the size of a
house brick.  They run about $1000 each...  Commercial 200kW sine wave RF
supplies use input wires about an inch think!  480 volts three phase at
about 300 amps for input power...  They could easily run a Tesla coil but
their price tag is prohibitive I would think.  I think that type of
equipment runs in the $1/watt range.

>
>I measure the top voltage in two ways.  One is a fiber optic system (good
>for any voltage) and the other is a capacitive voltage divider good up to
>100 kV.  The latter is used to calibrate the former.
>
>Finally, I am ready to ask my question.  If I am not mistaken, the standard
>wisdom among Tesla coilers is that shorted turns are bad.  Eddy current
>losses will lower the efficiency of the TC.  However, the current and
>magnetic field decrease as one goes to the top of a TC, so a shorted turn at
>the top does not decrease performance a noticeable amount.  The 'big dawgs'
>do not cut their beautiful toroids with the hope of getting a few more
>inches of spark.  But I am measuring the opposite effect.  The shorted turn
>toroid actually improves performance.  I am testing well below spark
>breakout, with input power in the range of 100 to 800 Watts.  At fixed input
>voltage, the shorted turn toroid will have a top voltage about 10 to 15
>percent higher than that of the open turn toroid.  The input current will be
>10 to 15 percent higher as well, for the shorted turn toroid.  The trends
>are consistent: more current = more input power = higher top voltage =
>longer spark (if voltage is high enough for breakout).  The surprise is that
>this occurs for the shorted turn toroid rather than the open turn toroid.
>Has anyone else seen this effect?  Does the effect show up in PSPICE models?
>Is there a 'simple' explanation?

Apparently, the Q is higher for the aluminum toroid case (lower loss).  I
don't quite understand how the copper toroid is made.  If it is 15 rings of
tubing arranged in loops above the primary, I would think they would act
like 15 shorted turns of copper tubing??  One would think the aluminum
toroid would always be worse, but the key may be that it is aluminum and
not copper.  Copper will have very high RF conductivity.  However, aluminum
will have a thick oxide layer at it's surface that may be providing
substantial resistance and thus make the shorted turn resistive so that the
current flow is far less and thus the loss is less.  It may also be due to
the small radius of the 1/4 inch tubing getting small but not visible
corona discharge on it's surface and thus providing the loss.  It is also
possible that the way the copper torroid is constructed is somehow lossy.
If there are any ferrous metals in it or the connections to the tubing are
resistive to RF, then that may account for the loss.  Basically, you are
looking for RF loss factors in the toroids.  The solid center of the
aluminum toroid may be providing a better low resistance path to the
capacitance on the outer edge of the toroid.  Hard to say without seeing it
but I bet something along these lines is causing the loss.

>
>I have noticed my name on the list lately in regard to some early results I
>published about input impedance.  Anything I have written about Tesla coils
>is in the public domain.

Neat!  We were interested in your "Input Impedance of a Tesla Coil" paper.
We wanted to post it at a web site so others could see it but we were
concerned about copyrights.

Cheers,

	Terry


>
>Gary Johnson
>