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

Re: Arnold Toroidal Ferrite?? Cores

Hi Ed,

Good to hear from you.

> Re:Combination of 60Hz and 100kHz signals
> in Choke Design.
> 
> The formula you have for Flux Desnisty computation
> if just a modified form of Faraday's Law of
> Induction: V=NA dB/dt where we assume a sinusoidal
> form for both V and B, ie B=B0sin(2pi ft) and V0 is
> 90degrees out fo phase from B so that
> V=V0cos(2pi ft). This gives the formula:
> 
> V0=NA x B0 x 2pi x f
> 
> Or its rearranged form:
> 
>         V0
> B0= ---------------------
>         2pi x f x N X A
> 
> Where   V0 is the peak voltage
>         B0 is the peak flux density
>         f is the excitation frequency
>         N number of turns of wire
>         A core area
> 
> The factor 4.44 which appears in the popular
> formula comes from multiplying 2pi by 0.707 so
> that V0 can be expressed in RMS voltage.
> 
> Key point:
> So, it is perfectly legitimate to "add" the voltages
> of various frequency components to obtain a total
> Bmax=B0. [If you want, you can go back to the
> deriviative form to prove it to yourself.]
> 
> Ah, but one hitch:
> How do you know what are reasonable voltages for
> each frequany component? [60Hz and 100kHz]

Your right.  I don't.
 
> I don't think your guess of 5000 Vrms at 60 Hz is
> reasonable. If your chokes have 2mH of inductance=
> .75ohm impedance at 60 Hz. If you had 5000 Vrms
> across this choke at 60Hz, it would be pulling
> 6600 amps! Most likely, you'd have at most a couple
> volts across the choke at 60Hz. So I don't think
> the 60hz contribution is important in the flux
> density calculation.

This was my contention originally with the techs at Pyroferric.  That is that
the voltage drop was what we were really concerned with.  Since the reactance
is so very low at at 60 Hz at these inductance values, the drop is very small
at the 60 Hz supply voltage.  The techs said that when the voltage swings to
negative from positive, and vice versa, that the flux does does so to the tune
of the total sign change in the circuit.  I didn't really "see" their logic,
and I told them so.  They admitted they were also quite shaky in this arena.
You would think that the manufacturer of the cores would know a little more
about them than they seemed to.  They also admitted that they had little
experience in filtering in such a demanding application with such high voltages.
I appreciated their candor.  They suggested a test, since they didn't really
know for sure in a circuit with both a 60 Hz and a RF component at high voltage.

> I think the real question is:
> How much of the oscillation voltage at 100kHz (or
> whatever) appears across the chokes and how much
> appears across the the supply transformer.

What about the bypass caps?

> If this were at DC, it would be a simple matter of
> an inductive voltage divider. The neon/pig has much
> more inductance than the 2 or 10mH of the choke, so
> most of the voltage should appear across it - not
> the choke.
> However, I don't think we have a good feel for what
> the actual high frequency impedance of a neon/pig is.
> There is so much interturn capacitance in a big
> high voltage transformer that I would expect its
> high frequency impedance to be near that of the
> chokes people often use (to apparently good effect).

I do have figures on a 12 KV potential transformer.  It was about a
4 KVA unit, if memory serves.  It has an inductance value of 59 H.
With 59 H -at- 100 KHz, we get a reactance of 37 MegaOhms.

This is where I don't see the logic of even considering the transformer
impedance into the picture.  After all, there is a capacitor(s) directly
after the choke that has only several hundred ohms impedance that goes
directly to a RF ground.  I am looking at the problem as that the RF is
being produced in the tank and we are trying to keep it from reaching the
transformer altogether.  While it is true that the transformer has much more
impedance at the operating frequency than the chokes, the filter caps that go
to ground that are between the chokes and the transformer are practically
a dead short at these RF frequencies, when compared to the transformer.  The
reactance of a 1 nF cap -at- 100 KHz is 1592 Ohms.  The reactance of a 1 nF cap -at-
200 Khz is 796 Ohms.  These figures should be much lower than the reactance of
any HV transformer at these frequencies.  Please correct me if I am wrong on
this point.

Now it may be that I am way off base.  I am trying to keep this RF
out of the transformer windings as much as possible.  I don't want the
transformer being part of the filter circuit.  I have heard talk of considering
the transformer's impedance before.  The bypass caps are basically in parallel with
each leg of the transformer.  Since the bypass cap is there, doesn't the impedance
value at that point fall below that of the cap?  It seems like a reactances in
parallel problem.

Comments?
Scott Myers