OLTC - Re: Primary resonators with very low L/C ratio

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Terry Fritz" <teslalist-at-qwest-dot-net>

Hi Paul,

I think there are three "limits"...

At 06:57 AM 7/9/2003 +0100, you wrote:
>How do we construct a primary tuned circuit which has a very low L/C
>ratio?   We come up against this question when we try to take an
>OLTC and either raise the frequency or raise the power level.
>
>As a target for discussion, how would a bang size of 50 Joules,
>firing at 300V, and resonating at 100kHz be dealt with?

E = 1/2 x C x V^2    ==>  50 = 1/2  x C x 300^2   C == 1111uF

Assuming a BPS rate of say 100BPS gives an input power of 5kW.

Fo = 1 / ( 2 x pi x SQRT( L x C ))   ==>  100000 = 1 / ( 2 x pi x SQRT( L x 
1111E-6 ))

Lpri == 2.28nH !!

For the secondary...

Assuming the biggest topload that is reasonable is 50pF, we get an Lsec of:

100000 = 1 / ( 2 x pi x SQRT( L x 50E-12)    L ==  50.66uH

 From the above, the first limit we run into is how "small" of a primary 
inductance can we use in a real coil.  2.28nH is far below what a real 
primary would be.  If Steve had used one turn in his smaller OLTC, he would 
have gotten ~560nH.  My larger 3 parallel turn primary hit 
~450nH.  Antonio's program could figure it out far better (almost exactly!) 
but I would think ~~300nH is the practical limit for how small a primary 
inductor could be made.


100kHz is a pretty good frequency to shoot for so we now get a primary 
capacitance of 8.44uF.  If we want to run 50 Joules, we just have to up the 
firing voltage to 3442 volts.  However, things get messy over about 1000 
firing volts in that caps and IGBTs start to fall out of a reasonable range 
of what is available.  With 1200V IGBTS, we may have to limit the firing 
voltage to 1000 volts or 4.22 Joules/bang.  Hopefully, streamer length "is" 
proportional to the square root of input power in this case too.  So all we 
have to do to get to 5kW is up the BPS to 1185 Bangs Per Second.  Since 
OLTCs like high coupling and fast quenching (fast primary firing cycle 
duration), we can really push the BPS rate way up to get high power as long 
as we can charge the primary cap fast enough.  So the second limit is how 
high of a firing voltage we can tolerate.  Newer IGBTs may got to 
3000+volts, but right now on the "cheap used" market, 1200V IGBTs seem to 
all we can find.  So an upper limit of 1000 firing volts is our second limit.


If we are unable to get a giant top load capacitance, we may have to go to 
a larger secondary inductance like I did.  However, the larger inductor 
will reach a point were the losses get so high they steal the coil's power 
like mine did.  A better (not cardboard) form would help that, but there is 
still a limit to how large you can make the secondary inductance before the 
effective series resistance eats the power.  Higher Fo frequencies really 
help this!  I was trying to get away with 120BPS (AC resonant line 
charging) so to get power I needed larger primary caps.  Which lead to 
lower Fo, which lead to very large Lsec, which lead to High series losses 
in the secondary...  For higher power coils, resonate charging (even at 360 
BPS three phase) may not be an option.  To get very high BPS rates we seem 
to need a type of DC charger like Steve used.  My OLTC was a little too big 
for resonant charging and it's 120BPS limit.  But resonant AC charging does 
vasty reduce weight and complexity.  Probably great for a smaller or even 
"table top" OLTC...

BTW - Dave Sharp has suggested that adding ferrites to the center of the 
secondary could improve primary to secondary coupling, reduce the secondary 
loss, and maybe aid in tuning.  Unexplored territory there...

So you have a low limit of about 300nF for the primary inductance, a firing 
voltage high limit of about 1000 volts (but you can have high BPS), and an 
upper limit to how high of an inductance the secondary can be before the 
series losses get you.  Our L/C ratio for 100kHz is about 300nH / 8.44uF = 
0.0355 (H/F).

Cheers,

         Terry

>--
>Paul Nicholson
>--