Re: Solid State - progress.

To: tesla-at-pupman-dot-com
Subject: Re: Solid State - progress.
From: harriman <harriman-at-metrolink-dot-net>
Date: Mon, 17 Jun 1996 23:56:09 -0400

>Paul
>Thanks for the schematic - interesting- but won't the primary
>detune the LC circuit and certainly bleed current and power
>out of it. I certainly think that these ideas are worth poking
>around with - if instead of brute force driving we can get some
>kind of resonator in the circuit then we should be able to get to
>much higher efficencies.

The LC is the resonator. 
It is true that it is detuned by the effects of the tesla
coil drawing power from it. But there will still be a q greater
than one if you set it up right. It requires a lot of current
from the push-pull driver when there is less load on the tesla coil
output. It works kina like a single stub tuner. It impeadance matches
you coil to your driver. 
 The primary of the tesla coil must be closely coupled with the secondary. 
I have tried it, it does work but requires
low resitance caps (mine melted after a about 10seconds).
It also solves the harmonic problem that most switchers have.
Harmonics from the switcher set up some mean nodes.  
That can burn up your secondary. I'm not ready to give up on it yet.

     (vin)
       |
       )
       )  
       ) L
       )
       |        R
       |         (Load)
Vout   +-----/\/\/\/-----GND
       |
       |
     -----
     -----  C                           
       |
       |
       |
     GND           

w=2*pie*freq
if  w = 1 / ((L*C)^.5)        

thus L = 1 / (C*w*w)
      
Vout/Vin=R * ((C/L)^.5)     

sub in L and you get

Vout/Vin=R*((C*C*w*w)^.5) = R * C * w = resistance * cap * 2 * 3.14 * (freq
in hz)

so if you run at 150Khz and R is say 100 ohms and cap is .5 mfd

you get vout/vin of 7.5 

or if your coil looks like 13.33 ohms  your  Vout/Vin=1

What does your primary look like at the resonante freq ?
 


>>try this for snubbing.
>>Im this is only a guess. 50% less power into snubber
>                  diode       cap  
>  drain_of_fet------|>|---+----|(---- ground 
>                          |
>                          |
>                          \
>                          /
>                resistor  \
>                          /
>                          |
>                          |ground
>
>But won't current be flowing through the diode and resistor
>virtualy all the time. I put the resistor accross the diode
>to discharge the capacitor during switch on, then it can
>charge through the diode at switch off.     
>

Yes this is true. Hmmm. It was just a guess.

->>How about if resistor is tied to VCC?

The resitance should be adjusted such that the voltage
across the cap is about 2Vin. 

power diss= V*I
if R is tied to Vin
and V across R is Vin.
->> power diss= V*I=Vin*Vin/R
and R must be such that V across R is Vin 
I=current disspated due to spike.  
->> R=Vin/I
->> power diss= V*I=Vin*Vin/(Vin/I)=Vin*I
Vin=100
off time 5%
current during off time=current during on time
say 15amps drive current
15*.05=.75amps
power diss= 100*.75=75 watts

If you can get your off time down to 1%
power diss= 100*.15=15 watts


 It looks like is would catch spikes better than a series R-C 
snubber. It also allows you to push your fets to the very close to the 
edge of Vmax. Might be worth the extra power loss to get more V out of fets? 
 It would not work well for PWM or forward converters. I like the
quazi-resonante setup better for the forward converters. PWM is a
great way to regulate but not so good for tesla coils because of the off time. 
If you want to regulate the output of a tesla coil modulate Vin at some lower
frequency.
   
Paul Harriman EE