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

Re: Air Permeability & a question

Hi Harry,

> Has anyone tried using an RF trap instead of a choke?  Would a pair of
> parallel tuned circuits, tuned to the tank frequency, work in place of the
> chokes?  These would have a high impedance at the tank frequency
> and a low impedance at 60 cycles.
> 
> There is probably some painfully obvious reason why people don't do this.
> Your comments and critique would be appreciated.

I thought about that a little more.  Are you referring to what I would call
a resonant band-stop or band-pass filter?  That must be what you mean.

If you are talking about these, I have never heard of anyone trying it.  
I have not tried doing any calculation for one before.  Ideally, a band-pass
filter that was capable of passing the 60 Hz component only would be ideal.
I did some quick calulations.  With a .007 uF capacitor, we would require
a 1 H choke.  Now for this to work, the choke would need to have an iron
core to keep it "small".  Even so, it would be large.  We would need to
avoid core saturation.  An iron core can take about 8,000-10,000 gauss.
Without doing any more calculations, it can see the difficulty.  The cap
would need to be quite large, or the choke would need to be quite large.
the cap would get expensive to be able to take the voltage and have a large
enough value.  The choke would need to get large to be able to get enough
henries out of it and not burn out due to the high voltage gradient between
turns.  A .007 uF cap of a fairly high Q would cost you $100 or so.  The
choke would probably weigh a hundred pounds or so.  If I assume a 4" square
cross section on the choke core (to make it fairly large) and an initial
permeability of say about 60, it would require 100,000 turns of wire to get
to 1 H.  OUCH!  I would hate to wind that thing.  Even with quadruple the
capacitance, we are still at 25,000 turns.  The 25,000 turns gets us to a
flux density of 2182 in this theoretical core.  Taking it to to about 8,000
gauss, we are down to 6811 turns.  That would be close to the maximum flux
density for a core of this nature.  still it is a tremendous number of
turns.  Even if we managed to make it (value of ~68 mH, and it didn't burn
out), it would require a capacitor of 103 uF at operating voltage RMS.
That of course is outlandishly expensive.  So, we would need to go to a large
choke again.  Keep in mind that I puuled a lot of the above figures on the
choke core out of the air and this was meant just for the sake of this
theoretical calacutlation.  THis may not truly represent real world avialable
components.  To me, this method just looks way too expensive, and large.

Now the band-stop filter is a different matter altogether.  Now we would be
talking about a filter that is trying to stop a specific frequency and not
pass that cor saturating 60 Hz.  I can see one problem right off the bat.
It is only going to be good for that exact frequency band.  If the band
could be made wide enough though, it might just do the job.  Then there is
the matter of attenuation percentage.  And what about the same self-resonant
 problems we experience with "standard" filter arrangements?

You might want to try to do some calculations to see what you come up with.
This is an intereting avenue.  If someone would like to double check my
figures above on the band-pass arrangement, I would appreciate it.  I have
been known to make a few errors.  It is an interesting excercise.  Dig out
those old electronics texts and enjoy!

Scott Myers