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

Re: BIG resonance! - thoughts...

Hi Alex,

        The secondary winding of your transformer (inductor) and your capacitor
form a resonant LC circuit.  Depending on the exact values of the components,
this circuit will present a resonant rise limited only by the "Q" of the
circuit at 50 or 60 Hz (note that the inductance of the transformer's secondary
winding is in the range of many thousands of Henries and high current
transformers support much higher "Q"s).  The voltages and spark gap effects you
are seeing are showing this effect.   Of course, you need to be cautious of not
turning up the voltage too high or something will definitely blow!  (The rule
is that the part you love most will blow first :-))

        While it may be possible to damp the resonance with resistors, it is
probably far better to either increase or decrease the capacitance value or
simply not turn the variac up past a certain point.  However, it sounds like
you are so close to the resonant point that you should really change the
capacitance value.  Ideally, the capacitance would be selected so that, at full
input voltage, the capacitor/transformer combination give the proper rated
output voltage.  There are possibilities of adding resonant correction caps
(cheap and low Q) that alter the 60Hz resonance but do not act in the primary
circuit during gap firing to "fix" such problems, but this is definitely on the
"bleeding edge" of the art and not really the "best" way to go.  But in moments
of desperation...  Contact me if such a path is of interest...  Easy to do but
the various details of fault and actual conditions have never been modeled and
verified to be "nice", so I hate to recommend anything for use in none
experimental situations.

        You "can" run the system in the resonant charging mode, but it sounds
like you are now too far into the resonant region for this to be a good idea. 
A little resonance may not be too bad but if you are in the 6 to 8 times
region, you are risking a lot...  You are very lucky that your "quick" 45 volt
input test did not blow something!!!  When in doubt, it is far better to turn
the variac DOWN, not up! :-))  The currents in this case are also much higher
than expected causing even more problems.  They also follow the resonant rise
rule...

        Unfortunately, determining the "best" capacitance value is not simple
and requires a bit of fancy equipment, but there is perhaps an "easy way".  If
you can find some ordinary high voltage caps (~1000 volts) in the say 5nF range
(they do NOT have to be poly or anything fancy), you can simply try different
values of capacitance on the output of the transformer with your variac (or a
small low voltage transformer) and meter (a cheap one) until you find a
combination that gives you the proper input to output voltage ratio (make a
graph).  Once you know the "good" capacitor value, you can deal with
constructing yet another cap or two.  If you ran this cap dry, then perhaps the
value will change to a better one once oil is added...

        Unfortunately, with an unusual transformer, the typical or recommended
values of primary cap are probably not accurate and there is a bit of
controversy about what is "best" anyway.  It is often necessary to do testing
to determine what is the best capacitance for special case transformers...

        You can congratulate yourself on having safety gaps across the cap when
you first powered this up.  You saved yourself from seeing your new cap and/or
transformer frying instantly in front of your eyes in the first fraction of a
second into their first try...  You can easily see why people are always
blowing transformers and caps.  The damage that can be done by these resonant
effects is instant and very destructive!!  The energy that is stored is high
and dumping it into a weak spot in the cap or transformer is sure to do very
significant damage... especially to one's enthusiasm for this hobby :-(

        Digi Key's latest catalog (Apr-Jun) on page 324 has 1600 volt 4.7 nF
poly caps for about a buck each.  Ten or more of these would be perfect to make
a little resonant rise tester to find the "best" cap value for your
transformer.  Perhaps a little experimenting and design work could give a cheap
easy to make device so that we all could use to find the best cap values for
our specific systems...   Well worth the effort if it saves a freshly made cap
or nice neon from going BANG on the first flip of the switch or building a nice
cap that is not the "right" value...  

        For the new and often old coilers, resonant rise is definitely the
worst
enemy and the cause of the most blown equipment...

Of course, the many many unknowns in all this is what makes it all so much fun!

        Terry
        terryf-at-verinet-dot-com



At 05:47 PM 4/1/99 +0100, you wrote:
>Dear all,
>
>I will add to this after another test....
>
>I hooked up my cap across my NST again, but this time put my multimeter on
>the output of the NST. I turned up the voltage, and at only 1 or 2 percent
>of line voltage I was getting about 600-700 volts on the meter!!  I tried
>it without the cap and all I was getting was around 80-100 volts on the
>same setting... All I can think of is that the NST windings are resonating
>with my cap at 60Hz and causing this voltage rise of about 6 to 8 times!!
>Instead of shunts, these things have a small gap in the core which provides
>enough leakeage inductance to limit the current.
>
>In fact, I'm just doing the test again right now - with the cap connected I
>set the variac for an NST output of 800V. I removed the cap and I'm only
>getting about 85V. This must mean that I was actually getting about 10kV
>across my cap, no wonder I got the big sparks! this is a rise of nearly
>10x!  Help!!! does this mean my xfrmr is useless for Tesla work? I can't
>see that my RC filter (desingned for the Tesla frequency) is going to stop
>this!
>
>Please respond, I am at my wit's end as to what to do - can I fill the gap
>in my core and add limiting on the input to my NSTs? maybe this would
>reduce the inductance and stop the ringing. Help!
>
>
>Alex Crow wrote:
>
>> Dear All,
>>
>> I finished my first layered cap today, and decided to give it a test
>> run. I had oiled all the plates as I built it, and given it a quick duck
>> in a bath, but the testing was done 'dry' so to speak. I hooked the
>> thing to a variac and gapped the terminations about 1/4 inch apart for
>> safety. I turned up the voltage on the variac, and got really loud
>> sparks across the termination leads at only 20 volts in!
>>
>> Then I increased the gap to nearly half an inch, and got huge,,
>> deafening bangs at just about 32 volts input to my neon, which I worked
>> out was only just over 1300V on the output! I then gapped further, and
>> just past 35 volts in the gentle hiss from the cap lowered in pitch very
>> suddenly and got a lot louder - and no more sparks on the output.
>> I bumped up quickly to 45V in and still no spark, but the hum stayed the
>> same.
>>
>> To put it bluntly, this worries me. I have layered my plates with an
>> inch of gap on all sides - I was hoping this will be enough with 2 caps
>> in series across my 10kV NST. I have heard that others have been fine at
>> 15kV with 1.5 inches of minumum inter-plate distance at the edges. But
>> if I can only get up to a fifth of the voltage on the input at most in
>> this 'damp' state (admittedly there are air gaps at the edges) can I
>> expect them to take 5kV each when they are fully flooded with oil? Is
>> this just a case of me being premature and should I wait until I have
>> run them in in my deep-fat bath before I get too worried? i know this
>> sounds like corona discharge drawing extra current from the transformer
>> and thus limiting the output, but will my corona be five times less
>> severe when there's no air left in the thing? Maybe it's doing resonant
>> charging (0.04uF on a 10kV, 150ma short-circuit current neon) and I'm
>> getting a serious voltage rise)?
>>
>> This cap took me two days to build and I don't want to go through it all
>> again!
>>
>> Any help would be deeply appreciated.
>>
>> Alex Crow
>
>