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

Re: Terry's DRSSTC - 6000 BPS >:o)



Original poster: Jimmy Hynes <jphynes@xxxxxxxxx>

On 5/5/05, Tesla list <tesla@xxxxxxxxxx> wrote:
> Original poster: Steve Conner <steve@xxxxxxxxxxxx>
>
>
> >Or as your 'mum' would say, fools seldom differ :P
>
> :P
>
>
> >Your scope shot is missing a cycle when it's pretty much reached
> >equilibrium
>
> Yes I was testing it at very low power with the current limit set so it
> only just kicked in. There wasn't even a secondary in place, I was only
> using a primary coil.
>
>
> >What is the time constant of a streamer anyway? This could be found by
> >finding the 'knee' in the bps/spark length curve.
>
> It seems to be about 1/70 to 1/100 of a second. The literature on spark
> discharges has the relaxation time of ionisation in air at about 1/70 sec.
> IIRC. So that's why our coils all work best around the 100bps kind of zone
> and anything more than 200 is just a waste of power. If you're aiming for
> maxium length/power input that is: higher bps can be fun for other reasons :P
>
Hmmmm... I didn't know the ions actually stayed around so long. I
thought it was just really hot air.
>
> >I know the time constant is really 1/e
>
> I think that's how relaxation time is defined- it's the time for 1/e (68%)
> of the ions and electrons to recombine into neutral atoms.
>
>
> >the little IGBTs are a bit more fussy about early switching than
> >the bricks. The hard switching causes huge spikes that kill them when
> >out of tune.
>
> I think it's because the little IGBTs have faster turn-on times. When you
> go out of tune in the "low" direction, the bridge sees a capacitive load,
> and you get horrific spikes as the IGBT turning "ON" causes forced recovery
> of its neighbour's antiparallel diode.
>


Actually, it's not. In my case at least, it's the turn off that is
bad. The ufast diodes are good enough for that. When it shuts off when
the current is still flowing, the inductance in the rails causes some
bad spikes (maybe it's the lytics fault, not the IGBTs?). It's nice
that way, so delay in the feedback circuit doesn't hurt much.

> *To explain recovery- The diode is essentially a short circuit to the rail
> causing massive "shoot through" until all the recovery charge has been
> passed, then it suddenly stops conducting. Fast recovery diodes have a low
> recovery charge, and soft recovery diodes have a gradual tailing off
> instead of a sudden stop. Diodes have no "recovery time" as such- it
> depends on the current.
>
> You might be able to make the spikes a less fierce by using bigger gate
> resistors to slow the turn-on.

Tried that, but still unacceptable

> But I guess the real geek cred is in making
> a controller that keeps it in tune all the time, or at least stays slightly
> too high so the load is always inductive.

Gonna do that... kinda... except capacitive.

> Gary Johnson added a choke across his inverter output to draw an inductive
> current that helped cancel out any capacitive current. This would make it
> more forgiving of mistuning. There might be some mileage in resurrecting
> this idea?
>
> Steve Conner
>
>