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

Re: NST Filter and Circuit Configuration?

Hi Ed, Travis, and All,

	I thought I should mention something about the stress on the filter
resistors in RCR  filters to clarify why they can easily fail if the
elements cannot dissipate heat quickly as in most wire wound types.  This
will illustrate my concerns:
	Assuming one has a 1000pF cap and a 5K resistor on each leg of the 15kv
transformer,  The cap will charge to around 10kV at the point were the gap
fires and shorts.  With a 30ma neon the resistors will dissipate about 5
watts just from the AC charging current (I^2R) which is certainly not an
issue.  However, when the gap fires and shorts, all the energy in the
filter cap is dissipated in the 5k resistor.  In this case, each firing
will dissipate 50 mJ (1/2 CV^2) of energy in the resistor or 6 watts at
120BPS.  At the gap firing, the full 10kV of the cap is placed across the
filter resistor.  This can lead to the windings arcing and failing in some
types of resistors.  So the resistors have to be able to withstand 10kV
pulses.  The real killer is the instantaneous energy dissipated in the
resistor.  10000 volts across 5k ohms gives and instant current of 2 amps
at an instantaneous power of 20000 watts!!  If you divide 50mJ by 20000
watts you get a rough idea of the duration of the pulse which is 2.5uS in
this case.
	So, to make a long story short, the resistors will have to withstand 10000
volt, 2 amp, 20000watt pulses that last 2.5uS at a rate of 120BPS.  The
average power is around 6 watts.  Apparently, the resistors I mentioned are
in close enough contact to the ceramic core and ceramic coatings to allow
the elements to handle these pulses.  Some wire wounds that have poor pulse
dissipation (fiber core surrounded by sand) will fail quickly.  Increasing
the current or the size of the filter caps will increase the power
dissipated and the pulse energy.
	RCR filters are very effective at NST protection but the resistors due
have this stress problem which the Ohmite resistors I mentioned seem to
withstand well.  I have also used common mode chokes to reduce this problem
but these add complexity to the circuit which is hard to predict on all
cases.  I hope this illustrates that filter design is not straight forward
and there is much to be considered.  However, RCR filters are my first
choice since they are easy and they provide very good protection.  They do
dissipate some power and the resistors need to be tough.  This can all get
very complex when resonant charging, the NST output impedance, etc. are
considered but the RCR circuit is pretty trustworthy in most situations.

Hope this made things clearer and not more cloudy. :-))

	Terry
	terryf-at-verinet-dot-com




At 05:35 PM 12/16/98 -0800, you wrote:
>Tesla List wrote:
>> 
>> Original Poster: Travis Tabbal <bigboss-at-inquo-dot-net>
>> 
>> Thanks for the comments Terry. I had located one you listed has having
>> good filtering capabilities, however, the new circuit you refferenced
>> seems a better choice.
>> 
>> In regard to the ohmite resistors you mentioned. Digi-key only lists low
>> values for them. 500 ohm seems the highest, your circuit recomended 5k.
>> They also list them as wirewound. Am I missing something?
>> 
>> Do the capicitors need to be anything special? I found some at a local
>> supply that are rated at .02 mfd 15KV and 47pf 25KV. The first is a small
>> white tube and the later looks like a ceramic disc. That's about it for
>> high voltage stuff. The ceramic disc looks like the leads are close enough
>> to arc though.. Do you have a source for either part?
>> 
>> Travis
>
>	Wire-wound resistors should be fine.  C&H Sales, in Pasadena, CA has
>(or at least had a few months ago) a large supply of 5k, 50 watt WW
>resistors for a couple of bucks each..  Can't find their web address
>right now, but you should be able to get it through a net search.
>
>Ed
>
>
References: