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

Re: More thoughts on RSG dwell time

At 08:23 AM 4/30/99 -0400, you wrote:
>>From my previous post:
>>I'm not convinced that electrode dwell time has too much bearing on
>>quench performance.  In your case, the rim of an 8" disk rotating at 1500
>>RPM will travel at 628 inches per second.  Let's shoot for a quench time
>>of 100 usec.  The disk rim would travel 0.0628" in this time.  Unless
>>your electrodes are each less than 0.03" wide, the dwell time will exceed
>>the target quench time.
>
>Warning - armchair anaylsis follows...
>
>It occured to me after I sent this that even this analysis is optimistic,
>it doesn't take into account the fact that the electrodes will arc well
>before they are lined up.  If the voltage on the cap just prior to
>electrode alignment would normally jump a 0.30" static gap with nice,
>smooth, large radius-of-curvature electrodes, it will certainly do so on
>the RSG electrodes.  If the RSG has two pairs of electrodes, this means
>that it will fire sometime prior to being 0.15" away from alignment.  If
>we assume that it will quench in 100 usec, this will have occured and be
>over BEFORE electrode alignment.  Electrode dwell time would appear to to
>totally irrelavant.  I suggest that efforts would be better spent on
>electrodes with large radii of curvature to minimize the distance at
>which arcing occurs.
>
>Gary Lau
>Waltham, MA USA


Hi Gary,

I wonder what factors really are "important" here:

We have,

Quenching - Using a very nice high speed, thin electrode, rotary, and multi
gap (16 in series) spark gap on loan to me, I can sometimes quench at the
high current part of the primary waveform as opposed to the zero current
crossing.  I used this to explore some of the claims made by the Corum's
last year.  At 5000 RPM with 1/8 inch electrodes, the electrodes "align for
180 uS" (1800 inches per second).   However, the first notch occurs at 60uS
in the coil.  These quench experiments never quenched on the first notch.
This gap is very high speed.  5000 RPM with 16 series gaps on a 100KHz
coil, if this won't quench on the first notch noting will!!  I can only
fully support your statement that gap dwell time does not affect quenching
and is "totally irrelevant".  The arc is far faster than any mechanical
system can reasonably be expected to deal with.  However, with thin gaps
traveling past each other at 100 MPH there is tremendous air flow such that
if the gap DID want to quench, it would have all the help one could expect
to give it.  I think the 16 gaps and the high air flow are all that allow
full current quenching with this gap.  The dwell time seems very irrelevant
indeed.  First notch quenching can be easily achieved with multi-gap static
gaps which do seem to work well but also have higher resistance...


Gap Resistance - I usually measure about three ohms of primary resistance
in my coil judging from ring down times and such.  Assuming all of that
resistance is in the gap (it isn't), the best I could do is to reduce that
resistance to near zero and see how that would affect the coil's
performance.  In real life I can't do that but my computer models can with
do this with ease and also remain in the spirit of armchair analysis :-))
With three ohms of primary resistance the peak output voltage is 296959 volts.
with 0.1 ohms of primary resistance the peak output voltage is 379994
volts.  Almost a 28% increase in output voltage.  So the old rule of making
the primary system resistance as low as possible still stands.  Of course,
really, how much lower can one get it?  You can gold plate big copper
tubing and have big electrodes that can get very close before the arc
occurs, but all the power is still in that big white hot arc.  You can do
some things to reduce the resistance but the effort to do so starts to get
pretty high.  There is a chance that the difference in a "good" and "bad"
gap has to do with the gap resistance and power dissipation.  However, what
controls this is still a bit of a mystery.  If we KNEW this was the key,
then we could specifically design to reduce this resistance at the expense
of other less important parameters...  Like quenching???


Timing - In all the things I have played with, spark gap timing gives the
most dramatic results.  Output arcs really like a nice steady smooth flow
of power.  

Static gaps fire when the voltage reaches a certain level.  Right or wrong,
when the voltage gets to the value of X, it fires.  If the system is
carefully designed and the gap is set right, this may be pretty optimal.
However, there is a lot that can go wrong to make this a poor performer.
Proper primary system design is really important and it is not easy to do
without some pretty hefty design work.  Coupling also plays a big role here
as does quenching.

Un-synced rotary gaps seem to work well with very high power systems that
can charge the primary cap very quickly and thus get around the 120 Hz
charging effects of neon type systems.  Systems, such as Marco's, that have
nice big DC supplies will really tell what break rate is best.  However, is
seems to be a bit above 120Hz.  Perhaps with three phase power and three
transformers and perhaps some diodes one could play some tricks here to get
to 360Hz.  The Electrum is similar to this, but I think it was rectified to
DC...

Sync rotaries have some surprising advantages.  They fire just when they
are supposed to.  No lost effort or firing at too low of voltage.  They
also can fire latter than the highest voltage point and get resonant
advantages as seen when using large primary cap values.  No doubt about it.
 Sync rotaries give the best performance from what I have seen.  Even on
high power machines they seen to run much better than the others.  It is
interesting to note that there are a number of "sweet spots" as far as gap
timing goes that only sync rotaries can take advantage of.  For large value
primary cap systems, this is critical since they need to force firing after
the normal high voltage point to charge the big cap but still get to full
voltage.

I haven't done this but it would be neat.  If you have a pig system that
you know can charge the primary cap several times during the AC cycle.  One
could use a sync gap that had a series of gaps that were spaced at the
charging rate.  In other words instead of having 4 electrodes on a 1800 RPM
rotor, you could have say 4 groups of 5 or 20 total electrodes.  Each of
the five electrodes in a group would be spaced so that the cap could
recharge before it would fire next.  In fact, they could be spaced
proportionally to account for faster or slower charging of the primary cap
depending on exactly where on the AC cycle they are programmed to fire.
During the low voltage parts of the AC cycle there would be no electrodes
and thus no wasted energy.  This would truly be an optimal sync gap.  It
would also be a great exercise for the computer models to do something
really useful :-) 

Comments and suggestions are welcome on one of the last and most complex
frontiers...

	Terry