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

Re: [TCML] Re:Pros and cons



I just remembered something about the DRSSTC topology that might be
beneficial to you.  Theoretically, because the energy transfer is spread out
over time, you can achieve 4X the peak energy in the secondary as you ever
see in the primary (im talking peak instantaneous energy level).  This
occurs after 1 complete energy cycle of the double-resonant system.  The
second benefit of the DRSSTC is that the primary characteristic impedance
can be significantly higher than the single-shot pulse method considered
before (so likely more than 1 turn primary).  Now of course the down-side is
that you need 2 or 4 switches to make a half or full bridge inverter to
drive the thing, and now you are also more subject to turn-off losses in the
IGBT.  Luckily the DRSSTC exploits zero current switching.

Anyway, it may be worth a quick design to see if the DRSSTC has anything to
offer you (i usually find its the best silicon utilization of any other
topology i've looked at).  You have the unusual case that your load is
fairly well defined (compared to a spark), and because you intend on
suppressing spark formation, the factor of 4X can actually be exploited
(minus losses of course).  The 4X thing comes at a cost... it takes
comparatively longer for the secondary voltage to ring up (this can be
important for spark production, but probably not too critical for you).

For really quick design iterations, Antonio has a really handy design tool
for the DRSSTC:

http://www.coe.ufrj.br/~acmq/programs/

Antonio may want to comment more on this, and how you might set up the
simulation for your situation.

Steve

On Fri, Dec 4, 2009 at 7:34 PM, Steve Ward <steve.ward@xxxxxxxxx> wrote:

> Its obvious now that i did not inspect your proposed design :-)
>
> Yes, 15kA will be too much for those dies (dont bother testing, it will be
> a disappointment).  Im not sure how this higher voltage silicon behaves, but
> i found that a CM300 (1200V variety) module begins to de-saturate at about
> 5200A when applying 30Vgs, at which point destruction is immediate.  Aside
> from the sure destruction via de-saturation, the pulse heating will fatigue
> the thing eventually.  It should be somewhat easy for you to figure out the
> transient die heating based on a model of the Vce vs Ic.  If the delta T is
> more than 30 Celsius, then its probably not gonna last very long (i admit, i
> have no idea how long that is), but it sounds like 30 degree's is about the
> limit that many pulse semiconductor applications like to run up to.
>
> Why not parallel these modules?  It sounds like it would be pretty
> successful as the di/dt is low at turn on.
>
> But, come to think of it, im confused slightly about your requirements for
> energy.  I dont know what OSIV is, so maybe thats why im confused.  Anyway,
> where is the energy going if you are not producing sparks?  More
> importantly, what is the impedance of your load?  I suppose i shouldnt be
> questioning you about this as you surely know more than me about it, but im
> still curious.
>
> Steve
>
>
>
> On Fri, Dec 4, 2009 at 11:58 AM, <uhvsystems@xxxxxxxx> wrote:
>
>>
>> Steve,
>>
>> Calculated and simulated average currents in considered  tesla coil
>> primary circuit  are well below  max average current rating of the IGBT
>>  (=600 A).
>> Other dynamical characteristic satisfy as well.However,I'm not sure it is
>> safe to stress the thing repetitively by peak currents higher than 20x rated
>> one.
>> If you ask IGBT manufacturers or distributers what they think of that idea
>> you'll get an  answer in a style "Run it on your own risk".
>> Specification of the IGBT module we are talking about you can also see
>> online at:
>>
>> http://www.datasheetarchive.com/datasheet-pdf/019/DSA00326011.html
>>
>> Notice the guaranteed application only up to 1200 A.
>> It looks we will have to conduct some 15 kA impulse current tests and see
>> if the IGBT module can withstand that.
>> Needless to say ,it always remains possibility to put 2 IGBT modules of
>> the same type in series or parallel and lessen some of the design
>> problems,but I dislike that idea becouse of other things.
>>
>>
>>
>> Dex,
>>
>> I know what high currents are capable of.Not only electrothermaly but
>> electromechanicaly too.One more reason to pay special attention to design
>> and construction details of tesla coil primary circuit under considerations.
>> On a side note,it doesn't matter if I like sparks or not.It is simply that
>> the appearance of the spark on HV equipment usually means a bad thing:power
>> arcing.We try to avoid that all the time,except in the places where is
>> unavoidable or necessary (ex:ignitrons,contacts of HV circuit breakers
>> etc).
>>
>>
>> David,
>>
>> Making or ordering HF oscillator in 500-1000 kW range for our testing
>> purposes is a bad idea.Unexceptable both  application-wise and money- wise.
>>
>>
>> Regards,
>>
>> Fez Zaev
>>
>> Steve Ward <steve.ward@xxxxxxxxx> wrote:
>>
>> Fez,
>> It sounds to me like you know what you are doing, to the point that it
>> will
>> work.  the IGBT you have should be fine for the modulator approach you are
>> going for because the long turn-off tail is mitigated by turning off the
>> IGBT switch while the primary current is free-wheeling through the IGBT
>> anti-parallel diode (giving a huge window for zero-current turn off).
>>  Those
>> IGBTs would NOT be very suitable for DRSSTC operation, where an H-bridge
>> of
>> IGBTs must switch at the coils resonant frequency (its basically a big AC
>> drive to excite the primary system).
>> 155J is an *impressive* energy.  Something on that size would have the
>> capacity of producing sparks in the 20 foot range (if optimized for spark
>> production), though i realize that is not your goal.
>> Steve Ward
>>
>> Dex Dexter <dexterlabs@xxxxxxxxxxx> wrote :
>>
>> Fez,
>> 20 microfards and 4.4 kV will give bang size of aprox 193 J!
>> At F ~ 25 kHz you will have primary characteristic impedance
>> only 0.3 ohms and really high peak primary current ( 13 kA +).
>> I doubt anybody has ever worked with  tesla coil primary currents
>> of this magnitude.One thing is certain,your coil will not be a very
>> efficient one.I doubt it would be able to reach efficiency like 80%.
>> Pay your attention to this fact.
>> I hope IGBT HV module you refered to will be useful for the coil.
>> BTW,interesting standpoint about efficiency issue I must admit.
>> Reminds me of the view of coiler Peter Teren (aka Tesla downunder),
>> except in one big difference.He wants and likes sparks,you don't.
>> Dex
>>
>>
>> David Rieben <drieben@xxxxxxxxxxx> wrote:
>>
>> Sounds like you don't want a "Tesla coil" at all then, but instead a
>> large, high powered oscillator circuit, like for a multi-kilowatt radio
>> transmitter.?
>>
>> David
>>
>> _______________________________________________
>> Tesla mailing list
>> Tesla@xxxxxxxxxx
>> http://www.pupman.com/mailman/listinfo/tesla
>>
>
>
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla