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

Re: Re: [TCML] SSTC full bridge control system question



Steve,
Before, I was allowing the simulation to run for several milliseconds,
and I observed the envelopes the current and voltage reach a steady
state amplitude without noticing any peaks or notches.  My K was
always in the neighborhood of 0.14-0.3 so it shouldn't have taken more
than 7 cycles to see a peak, if there was one to see.

-Mike

On Tue, May 4, 2010 at 10:40 AM, Steve Ward <steve.ward@xxxxxxxxx> wrote:
> Mike, how long are you running the transient analysis?  I think in general
> you should not have a really hard time exciting the "beating" mode for at
> least one transfer cycle (about 7 cycles at K = 0.15), but after some time
> the system tends to stick with one frequency or the other, and not stay in
> the middle because its not stable.  The tuning can be very sensitive, so if
> you are changing the primary L by 1uH increments, it might miss this tuning
> point that produces the exact result.
>
> In my quasi-CW system (which is double resonant), i could "bias" the system
> to lock to the upper or lower resonant frequency just based on tuning.
> Because the coupling was k=.5 or so, the transient behavior (which shows up
> as beating) was usually impossible to see because it was over very quickly.
> I found that somehow the system worked better running at the upper
> resonance, but that seemed to be the only system where this held true.  For
> my "transient" mode DRSSTCs, i always found that tuning low gave more
> favorable results.
>
> Steve
>
> On Tue, May 4, 2010 at 12:29 AM, Michael Twieg <mdt24@xxxxxxxx> wrote:
>
>> All right, I'll try to apply your methodology to our existing
>> secondary parameters and see if I can get anything promising.  I've
>> tried already a couple times and can't get it to resonate at the
>> center frequency; it always goes for the lower frequency.  And to be
>> honest, that's what I would expect to happen.  There are generally
>> going to be three frequencies at which the phase of primary is current
>> is zero (which means zero crossing can work there), but the middle one
>> is always the largest impedance, and I would expect the other two to
>> dominate.  As I mentioned before, pulling some of your example numbers
>> and throwing them into LTspice did result in it resonating at the
>> center frequency, but I haven't gotten anything else to do that.  Any
>> idea what causes the controller to select the middle frequency?
>>
>> -Mike
>>
>> On Mon, May 3, 2010 at 9:09 PM, Antonio Carlos M. de Queiroz
>> <acmdq@xxxxxxxxxx> wrote:
>> > Michael Twieg wrote:
>> >>
>> >> That was a fascinating read Antonio.  I was able to replicate your
>> >> results accurately, even in my zero-crossing control model.  I didn't
>> >> think that complete beating in the signals was possible with
>> >> zero-crossing switching, since the beating should require a reversal
>> >> of phase.  In my previous simulations, the system was always excited
>> >> to the primary resonant frequencies, and never the "middle" frequency
>> >> as in you explanation.  Would this be because my own attempts didn't
>> >> meat your criteria for the spacings/ratios between the resonant and
>> >> middle frequencies?  I'm still a little cloudy on what you mean by
>> >> "odd" and "double odd" differences.
>> >
>> > For complete beats a set of particular relations among the inductances,
>> > capacitances, and
>> > coupling coefficient must exist. Nothing very critical, as the element
>> > values used in a normal
>> > SGTC are usually pretty close to the tuning required for an SSTC working
>> in
>> > "notched" mode,
>> > in one of the many possible modes (but not identical). I define the
>> "mode"
>> > by three numbers that
>> > define the ratio of the two resonance frequencies of the system and the
>> > excitation frequency (the excitation
>> > between the resonances). In the best design, the numbers are odd and have
>> > "double odd" difference,
>> > as 1:3:5 (difference=2, two times 1, that is odd). Usual systems will be
>> > around mode 37:39:41,
>> > as in the example in my page. Irregular modes are also possible, with
>> > differences between the
>> > odd numbers being 6 (2x3) or 10 (2x5), but there is no advantage in using
>> > them.
>> > Of course, streamer loading affects what is actually obtained, but the
>> > influence is similar to
>> > what occurs in a SGTC, that works with similar waveforms (actually, a
>> > magnifier works with
>> > more similar waveforms).
>> > It is really interesting that zero-current switching is natural in SSTCs
>> > operating in this way.
>> >
>> > Antonio Carlos M. de Queiroz
>> >
>> >
>> > _______________________________________________
>> > Tesla mailing list
>> > Tesla@xxxxxxxxxx
>> > http://www.pupman.com/mailman/listinfo/tesla
>> >
>> _______________________________________________
>> Tesla mailing list
>> Tesla@xxxxxxxxxx
>> http://www.pupman.com/mailman/listinfo/tesla
>>
> _______________________________________________
> Tesla mailing list
> Tesla@xxxxxxxxxx
> http://www.pupman.com/mailman/listinfo/tesla
>
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla