Re: Qiuck info before I buy

[Terry Fritz <twftesla-at-uswest-dot-net>]

Hi Dave,

        Here are my thoughts.

I try to dissipate 5%-10% of the coil's power in the resistors.  After
calculating and modeling such things for hours on end, this seems to be the
final results I always ended up with.  So you have 5000 watts of transformer
(300mA) and you would want about 250-500 watts in the resistors.  Since the RMS
current is 300mA the resistance can be found by rearranging ohm's law.

P = I^2 R  ---->   R = P / I^2

For 250 watts, R = 2.78K
for 500 watts, R = 5.56K

Since you probably have two legs on the circuit you would need two resistors of
1/2 the above value.

I would get four L225J1.0K-ND  220 watt 1K resistors from Digi-Key for about
$50  These big power resistors get super hot (like 300C!) when run at their
full rated power so doubling up on them lowers their temperature to a sane
value.  The high temperature on the resistors has surprised many.  These are
more like space heater elements than those little 1/4 watt things! ;-))  Two on
each leg gives 4K total.  Mount them so they can run hot.  Of course, voltage
stand off is not an issue with four of these BIG resistors.

For the caps, you can work from the cutoff frequencies and some common sense. 
At 120BPS, your transformers can charge a 37.5 Joule cap to full voltage if
everything is just right.  At a 21000 volt firing voltage, that give a cap
value of 170nF.  Applying that same 5 to 10% rule to caps you could use a cap
value of 8 - 17nF.  10nF is a nice value. so we can look at that.  That is a
very loose rule but it seems to get close for starters.

Fc = 1 / (2 x pi x R x C)  so the cutoff frequency of the thing is about 4 kHz
which seems low.  However, each 5nF 2K leg is 15.9kHz which is very good!  But
is it "too" good?  Every time the gap fires, the filter caps get discharged
into the resistors.  This added heating should not be significant so we had
better check.

The power dissipated by discharging two 5nF caps at 10500 volts 120 times a
second is 66 watts which is fine.  

P = 1/2 x C x V^2 x BPS

Seven 33nF 1600 volt poly caps in the MMC style on each leg to ground would
give the capacitance needed (10M drain resistors across them is a good idea
too).  The RMS current is about 100mA so the current through the filter caps is
not a concern here.  It is interesting that they see about 2 amp high frequency
spikes which is far beyond what the neons would want to see!

So you could parallel all the transformers with 5nF from each leg to ground. 
Then two 1K resistors are in series with each leg to the main gap.  Of course,
a safety gap in parallel with the main gap and MOV strings across the legs to
ground are recommended as well.  I would also individually fuse the input of
each transformer to prevent then from nuking if they saturate just wrong.  PFC
caps would really help the variac and such buy cutting the input current from
maybe 70 amps to only 40 amps input.  If the neons are 120volt, the input
current becomes a bit more of an issue compared to 220V transformer systems.  I
assumed you are connecting this all together as in the diagram at:

http://users.better-dot-org/tfritz/NSTFilt.jpg

You may want to get higher current MOVs although they get pricey at higher
current levels.

you could just make individual filters for each transformer but with five, I
think it is more economical to do just one big one.

There is a thought that you could make an RCRC filter with your four resistors
and two sets of caps but I have not studied such a thing.

I hope my thinking out loud and skipping a few things with arm waving was not
too confusing.  This can really get to be a very messy subjects and I tried to
hit the main thoughts without getting too messy.  Within certain guidelines,
the values are not very critical at all.  It is more a mater of checking to be
sure nothing is wildly off the mark.

In general only the power dissipation in the resistors chages with increasing
current.  The component values are really fairly stable regardless of how many
transformers are used.  Thus, I sort of knew the answer before I started ;-))

Cheers,

        Terry




At 01:06 PM 7/15/00 -0400, you wrote: 
>
>     Since everyone seems to say that wound toroid chokes are not really safe
> for NSTs anymore, I want to make a RC filter for my NSTs.  Can anyone
> recommend a good value for the resistors and for the capacitors for
> 15kV/300mA, or better yet, how I can just figure it out for myself for all
> different setups?
>  
>     Thanks very much to everyone for their help - happy coiling - Dave