RE: SSTC Question from a newbie.

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

Original poster: "Mike Panetta by way of Terry Fritz <twftesla-at-qwest-dot-net>" <ahuitzot-at-mindspring-dot-com>

On Fri, 2002-07-12 at 07:07, Tesla list wrote:
> Original poster: "Mccauley, Daniel H by way of Terry Fritz
<twftesla-at-qwest-dot-net>" <daniel.h.mccauley-at-lmco-dot-com>
> 
> 
> 
> > 2. On the SSTC schematics I have seen so far, everyone uses a xfmer gate
> > drive.  Is there any reason noone has used a gate driver chip to direct
> > drive the gates? 
> > 
> Yes there is a reason.  You want to isolate your output circuit from your
> control circuitry.  Remember, some of your FET gates are swinging at high
> voltages.  
> 

Ok.  I understand that.  I thought maybe that people were using them to
isolate the driver circuitry if a streamer hit the primary.  Thats why I
could not understand how an xfmer would help (I was thinking about
protecting against tens or hundreds of kvolts, not just 1 kvolt).

If thats the case, then I ask again, why not just use one of the IRF
driver chips with a current multiplying arrangement (if necessary)? 
They isolate up to 1200V.  I was thinking something like the IR2112 or
any of the other IR21xx high and low side driver chips (not the half
bridge drivers though as the deadtime would be too high I think).  If I
had to, I would even build a small simple high side gate drive supply to
use with these chips if the charge pump were not enough.  IRF even has
an app note on how to do just that (along with the current multiplying
arrangement).  Maybe the real issue with these particular chips would be
that they cannot switch fast enough?  At what frequency do the fet gates
usually need to switch at?  Say 100-300KHz depending on the resonance
frequency of the coil?

> 
> >  If the output current is not enough for the frequency
> > (the IR parts I was thinking about only go up to half an amp I think),
> > there are schematics that show how to multiply the output current up to
> > 6A.  If isolation is a problem (I cant believe that the wire people are
> > using on these xfmers can withstand any voltage, maybe 600V at max...)
> > 
> I use 1000V Teflon coated wire.  You do really need too much withstand
> voltage.  For most 240VAC circuits, you only would need about 600V.

Yeah, now I see.  Again like I said above, I thought the driver xfmer
was to protect the controller/driver against strikebacks, possibly
saving the controller but obviously there is no way to save the output
FETS ;)

> 
> > then maybe using one of these gate drive chips with some form of optical
> > isolation would be in order.  Maybe fiberoptic cable?
> > 
> > 3.  Do any of these circuits work at voltages as low as 12V up to say a
> > max of 48V?  I would like to run the thing off a lower input voltage
> > while I am first learning. 
> > 
> Absolutely.  Either input DC or use a step-down transformer to get your
> 12-48VAC or use a variac.

I would want CW operation so I would be using streight filtered full
wave rectified DC.

> 

> > 
> 6-12" arcs should be extremely easy to obtain if you build a standard
> H-Bridge circuit.  The biggest problem will be dissipating the heat of of
> the FETs.  My earlier SSTC's using IRF460 FETs got extremely hot even at
> 120VAC input.  I later changed over to 1000V IXFN24N100 FETs which worked
> incredibly great except that they cost about $70.00 apiece.

For low voltage operation I was thinking about using some of the 55V
100+A fets.  They have a low Rds ON, so I do not think heating would be
too much of a problem, unless my primary current was excessivly high.

> 
> 
> > 4. (This question may show just how little I know about tesla coils ;) I
> > was wondering if it may be advantagious to place the SSTC primary inside
> > the secondary instead of outside.  Does this make sense?
> > 
> > Would it be possible to use a
> > conventional flat primary (external of course) with an SSTC driver? 
> > What would the pros/cons of this configuration be?
> > 
>       Wouldn't really work that well.  With a SSTC you want very close
> coupling.  Pancake coils can be used with conventional tesla coils since
> they are run at very high voltages (12kV +) Thus the fields created are very
> large even at the distances from the pancake coil to the secondary.  SSTC
> coils are run at very low voltages (<1000V) therefore need to be tightly
> coupled to the secondary.  The reason tightly coupled coils aren't used with
> conventional coils is because of flash-over reasons among others.

But I thought magnetic flux was generated by the current flowing through
the inductor not the voltage.  I understand that by using a really high
voltage you can get an even higher current (I=E/R etc...) so I was
thinking I could just lower the resistance of my primary to get a higher
current? 

> 
> 
> If you are interested in building a SSTC, i have two kits available:

Thanks for the offer, but as I want this to be a learning experience
(even if it means I have to blow some components along the way), a kit
is not really what I want.

I have seen your website (the ebay post is actually what helped spark
intrest), and I have done much research on the web in the last few days,
which is actually what braught me to the point of where I thought I
could ask some questions.  I have been to several SSTC web sites already
(yours included), so basicly these questions I am asking because they
were not asked or mentioned in any way on any of the web pages I looked
at.  I am specificly very intrested in why noone has tried the direct
fet drive method yet (almost all of the SSTC drivers I have seen use the
drive xfmers, the only thing I have seen that uses direct fet drive is
Thor, and its not a coil, its basicly just a powersupply).   My second
intrest I guess would be why noone has used IGBT's yet (I think I can
answer that one myself though...).  I am intimately familure with
allmost all of the appnotes and technical papers on the IRF web page
about driving fets, as I have already read them when I was desiging an
H-Bridge driver for a motor controller.  So I really have no questions
in that aspect.

Thanks for the help,
Mike