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

Re: Neon XFRMR Protection

Hello David,

>  Well, I have been following the posts concerning RF filtering and have come
> up with a design for my resonator,.......... 
> ..............Now for the filter. I have 8 T-250-B
> toroids and 2 Arnald toroids from Burt. Now I plan on winding them with 50-60
> turns of 20-22 guage single strand hookup wire rated at 600 volts and placing
> 5 on each leg for about 3 mH (?).

You don't need to use so many cores.  You are in no danger of flux saturation.
I currently am using these same cores with approx. 115 turns in 2 layers for
a measured value of approx. 3 mH each.  I am using 2 in series per leg.  If I
were to use only one per leg, I would still be at less than 10% of the maximum
flux density for these cores.  I use 24 gauge hookup wire, and that is plenty
big for the amperage.  I am planning on winding a few more double layer units,
but with magnet wire.  I should be able to get the value up to 6 mH a piece.
Will they burn out with the magnet wire?  They shouldn't, but I'll let you know
after I run them.  Save your cores.  Wind more wire per core for higher
inductance values.  Just use plenty a dielectric between layers.  I used
standard electrical tape.

> Next I will place 2 doornobs rated at 4000
> pF-at-30KV in series on each leg to the RF ground for a total of 2000 pF-at-60KV.
> The chokes will also be in series with 2 Kohm-at-100(?) watt  resisters. And
> there will of course be the safety gap to the RF ground.

With 2000 pF per leg, you will get approx. 5.5 mA of leakage current -at- 60 Hz
-at- 7.5 KV, per leg, for a total of 11 mA.  That seems acceptable for the power
input you have available.

The resistors should dampen out the resonance nicely.  Here is something that
many may be interested in;  I have just looked at the schematic of a commercial
line filter recently and noticed that there was a resistor across the 2 legs of
the filter, between the 4 chokes in the filter.  I couldn't help but wonder, is
that filter there to eliminate/reduce resonance within the filter?  I had no
other explaination.  There was no value indicated.  I have some 400 megaohm
high voltage resistors that I could try in my high voltage RF filter to see if
that works to reduce the resonance.  If it does, the power wasting 2.5 Kohm
series filters could be eliminated.  Thoughts?

> So now if anyone can see any gaping holes in this setup please show no
> mercy in the criticism, as so the RF does. I am looking for about 90%
> reduction in the RF back to the neon and am willing to lose some power threw
> the Caps. Also, if I am sigifigantly off in estimating the inductance value
> of the chokes, please anybody who has actually measured or knows better, let
> me know.

In your message, you don't indicate what your expected operating frequency is.
I would imagine that you can get to the 90% with the above arrangement,
depending upon your frequency.  Below 150 KHz, you are going to need a larger
choke or large cap value.  Since the cap adds more leakage, I expect you will choose
the larger choke value. 

This bring another point to mind.  Ed Harris and I have had some conversations
on the effect of increasing frequency on the reactive value of a neon
transformer's secondary on increasing higher frequency.  He measured some values
on actual neon transforemers.  Here were his results, as previously posted:

--------------
Ed's measurements:

Freq.                   Secondary Impedance
                        Primary open    Primary Shorted
------------            ------------    ---------------
100Hz                   1.2Mohm         .38Mohm
1kHz                    600kohm         same
10kHz                   35kohm          same
100kHz                  4.2kohm         same
300kHz                  1.8kohm         same

* The impedances are the total of  the reactive and resistive
componets. However, "eye balling" the scope traces suggests that the
current is always lagging the voltage by about 90 deg, so most of the
impedance is effectively inductive.

** I used a 1volt excitation for all measurements.

*** The open and shorted primary measurements show that the mutual
inductance between the primary and secondary has no effect of the secondary
impedance above about 1kHz.

--------------

If these measurements are correct, it would suggest that an additional choke
is neccessary just after the spark gap, and before the "traditional" choke/cap
filter  arrangement.  Ed has asked others if they can see why he measurement
techniques may be flawed.  No one responded, yet.  Adding this choke effectively
increases the impedance of the transformer at operating frequency.  This keeps
the capaciotr passing the lion's share of the remaining RF.  This does go against
normal thinking.  The neon's impedance value should go up with frequency, not
down!  Odd isn't it?  Ed offers this explaination:

------------------

Ed says:

I suggest that there are two reasons why the impedance is so low at high
frequencies: (1) The inter-turn capacitance in the big secondaries of these
kinds of power transformers is so great that it effectively shorts out
much of the secondary inductance as the frequency is raised. (2) It is well
known that a big 60Hz iron core cannot respond to high frequencies due to
eddy current screening. So the core "dissappears" at high frequencies and
therefore the inductance of the secondary may decrease by a factor of
around 1000 (approx permeability of silicon steel).

--------------------

Richard Hull indicates that the seconday of these transformers has a very large
inductance value.  I too have measured them with an LCR meter and it is in
Henrys.  Since the core "disappears" (I think we can agree on that), we are
effectively left with an air core inductor at operating frequency.  Still, with
that many turns, we should have a high inductance, as pointed out by Richard.
That brings us back to Ed's point 1, the interturn capacitance.  I have not
measured it, but it certainly seems plausable.  If the theory on interturn
capacitance is true, it would certainly explain the burn-outs we have all
experienced.  Since I can find no fault in the method that Ed used to measure
the reactances of the transformers at various frequencies, I must assume them
to be correct.  As soon as I get this old scope running correctly, I am going
to run these tests myself.  Thoughts?

> Also, I am not sure about the current rating of the resisters as
> there is a potential for 900 watts on each leg, so in theory they should be
> rated at 900 watts?

No.  A resistor only needs to be sized for the power it will dissipate.  Use
the formula P=I^2xR.  Since you know you have 2000 Kohm per leg and 109 mA
current flow (120 mA - 11 mA leakage), it's just plug and chug.  The answer
is size them for at least 24 Watts.  I would use at least 50 watt units to
be safe.  If you intend on increasing power at a later date, I would go with
100-200 watt units.

Scott Myers