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

Re: DRSSTC Primary Circuit Feedback Control



Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi,

At 09:27 PM 12/6/2004, Jimmy wrote:

>
> Yes!!  The models show that for sure.  If the primary and secondary are not
> tuned, then the primary voltage just keeps rising pretty much till
> something blows up!!  A spark gap might "still" be a good idea ;-))

As long as you limit burst time, you should be fine ;-)

Oh!! Good idea *;-))


> The models (not real confident in the numbers..) say the IGBTs will do
> about 40 watts of heat as is.  But with "perfect" zero cross switching,
> that goes down to 1.2 watts!!

Huh? What frequency and switching times? My simulations showed about a
factor of 2 increase in loss when hard switching about 80% of peak
current. A factor of 30 is hard to believe.

Looses at "zero current" are real low. Just the primary RMS current then... But not sure I trust then numbers. IGBTs are not really linear so switching losses get pretty high if the current is high. That mJ loss stuff on the data sheets adds up... Also worrisome having 20,000 volts behind it!!!


 I also seem to be able to keep
> the peak primary currents at just 80 amps which is trivial for the
> IGBTs.  Almost could use FETs again...
>

80 amps? Where's the power? I don't think you'll get very big sparks
with only 80 amps. If we're going by the 'semi-cw' theory, then you
need a lot of power to push the sparks out.  If you use a toroid with
no breakout point, then it might work...

Hmmmmmm... I'll have to look into that... The models may be tricky there since they are not good at streamer power issues... Might not have enough "bang energy"... Thanks for point that out!! It helps if the primary and secondary wiring losses are real low there contrary to Dan's observations about adding losses... I have to find out where the "heat" is...



Steve wrote,

Hey Terry,

I tested primary current feedback quite a ways back with my first
DRSSTC.  It still made about 48" sparks from about 1000W input power,
good performance.

Did you notice if the IGBTs ran any cooler vs. spark length? But the switching time has to be just right which might not be right, even with primary feedback, if one has any delay in the circuit.


Interestingly, secondary feedback gives a nice
linear looking ringup while primary feedback is almost a "stepped"
ring up, probably the phase relationship between the 2 LCs causing
that effect.

Yep. They dual coupled LC system normally wants to have that dual frequency thing. But with secondary feedback, the IGBTs have to "force" the system, which increases power dissipation on them. The dual frequency should not hurt anything and should take a lot of stress of the IGBTs...


Anyway, both ways seemed to work well, and ive even been
considering using primary feedback on my larger DRSSTC, but that might
not happen for awhile.  Anyway, im certain that you would have to
build in an overcurrent detection circuit on the primary side in this
situation to prevent things from running away.

A spark gap and resistor across the primary cap ;-) But every switching power supply out there has "real" over current protection... Probably a reason for that ;-)


I use overcurrent
detection on my big coil and it triggers on every ground strike ;-)
and also knocks off about 3 RF cycles!  Whats interesting is that it
can still feed the ground arcs, yet my controller is hacking off like
20% of the bang :-).

I nice area for computer models to play with since they can predict and tell one when things get bad.



Cant wait to build the CM600 coil :-))

:-)))))

Cheers,

        Terry



Steve