From: FutureT@aol.com
Date: Fri, 17 May 2002 10:56:04 EDT
Subject: Re: [TSSP] Topload breakout potentials
--part1_82.1bbbec3f.2a167404_boundary Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit In a message dated 5/17/02 9:21:39 AM Eastern Daylight Time, paul@abelian.demon.co.uk writes: > Anyway, with this extra bit of primary voltage the model peaks at > 24.5kV/cm at your -12% tuning. The surface field at MCTV is 38kV/cm > so I'm wondering if the initial 24.5kV is enough to start a small amount > of breakout, which adds a little to the Ctop, which brings the secondary > down a little into better tune, therefore more Vtop, more breakout, etc, > and so on, eventually multiplying to give you copious breakout with the > coil effectively in tune. I guess this process might take place over a > few consecutive bangs, or perhaps within one beat? Paul, It may take place both within a beat, and also over consecutive bangs perhaps. How critical would the tune point be, after breakout, considering that the Q will drop due to the streamer loading? > > Do you have difficulty quenching? At 12% detune the primary waveform > doesn't form a very good notch - it's still at 50% of the max amplitude, > thus retaining 1/4 of the bang energy in the primary. If some > corresponding detuning of the secondary/topload as described above > wasn't also occuring, then perhaps you would have trouble putting out > the gap arc. I don't think the quenching depends much on a perfect tune or perfect notches. In work I did, I saw little difference in quench-times as those parameters were varied some. Basically, it seems the gap will quench when the energy is gone... sucked out by the streamers, or lost in losses. More energy transfers will occur in a given amount of time with tight coupling or mistuning, but the total time to quench will be about the same provided the streamers are long, to provide loading and assist the quench. Coils such as Barts usually quench at the second beat notch perhaps. It seems that in a trade-off, tight coupling is preferable to first notch quenching. The overall losses may be lower that way, or as Bert H. suggested, the multiple energy transfers may help extend the sparks. The scope traces were somewhat blurry in my tests, so I may have gotten a mistaken result. I can't be sure. I think where quenching becomes very important is in high bps systems, where the gaps may power-arc more easily, or where a delayed quench may steal time from the shorter cap recharge cycle. Is it the case that some of the art of tuning the TC is > > finding just the right amount of primary detune so that the breakout > drops the sec Fres enough for the majority of energy to transfer? That > certainly makes tuning a much more interesting operation! I suppose in > that case, tuning for deepest primary current notch would be the thing > to do, since that implies max energy transfer is achieved during the > beat. It surely can't be coincidence that four of the five coils looked > at so far use a fair amount of detuning. The need for detuning seems real, but I'm not sure the nature of the notches matters much. The peak energy may be very important though. I may have missed something in my old tests though, so I'm certainly not too sure about what I'm saying. It may be mostly the actual tuning relationship that is most important, although I can understand how the beat notches affect all this. In any case, I've noticed in some coils, that the gap arc dims as the streamers grow, showing that the energy is being drained out faster from the system, and shortening the quench-times (when they dim) as would be expected. > If we take the tuning setups shown on the web page to be the those that > give optimum performance for each coil, then we can use the detune > percentage to calculate the effective capacitance of the streamer load > (assuming the streamers load the secondary down to match the primary), > and from that we can estimate the total streamer charge, which in turn > we can compare that with the 'surplus charge' value calculated on the > basis of how much larger the MCTV is above the breakout voltage. I wonder if the best tune point is a sort of compromise, rather than being tuned for max streamer length? Yet, I tend to think the coil does have to be tuned for max streamer length or so, whatever the reason. What do you think about Malcolm's suggestion about the "offset" tuning, which emphasizes the lower of the split spectrum frequencies? I would be interesting if there was more than one reason for the lower tuning. I guess if detuning helps to form just one streamer in some coils, that's another spark-length benefit from detuning. It would be interesting if Bart would run his coil at a lower power, so that he gets just one streamer, then tune to make that streamer as long as possible, then see if it "scales" with my TT-42; same freq offset, same ratio of top volts to toroid breakout voltage stress, etc. This may be especially meaningful if he is also using 120 bps, as I am. I wonder if I'd see much difference if I tune at 1/2 turn intervals, instead of only at 1 turn intervals as I'm presently doing? Cheers, John > > Let's just try some ballpark figures for ba0: > > The secondary+topload has a total shunt capacitance of 40.7pF and to > detune by 12% requires 51.1pF, so we suppose the streamers are providing > the extra 10.4pF. > > Now if the streamers are clamping the surface field at 26kV/cm, that > takes a topvolts of 26/0.043 = 605kV. Thus our 10.4pF of streamers > holds up a charge 10.4pF * 605kV = 6.3uC. > > Now the MCTV would allow the topvolts to reach 880kV, which exceeds the > breakout potential of 605kV by 275kV, and thus the coil is capable of > supplying a surplus charge over and above that required to reach break- > out of 275kV * Ctop = 275kV * (30.5pF+10.4pF) = 11.2uC, (The 30.5pF is > the topload capacitance). > > Thus if the tuning is 'ideal' for the coil, the streamers are holding up > around 6.3uC, but the coil is theoretically capable of delivering > 11.2uC. It's nice to see these two numbers landing in the same > ballpark, > and that the first is less than the second. > > For now, the latest coil models are uploaded to > > http://www.abelian.demon.co.uk/tssp/cmod/ > > and I'll see about adding some extra rows showing the streamer cap and > charge estimated from detuning. > -- > Paul Nicholson, > -- > --part1_82.1bbbec3f.2a167404_boundary Content-Type: text/html; charset="US-ASCII" Content-Transfer-Encoding: 7bit In a message dated 5/17/02 9:21:39 AM Eastern Daylight Time, paul@abelian.demon.co.uk writes:
Anyway, with this extra bit of primary voltage the model peaks at
24.5kV/cm at your -12% tuning. The surface field at MCTV is 38kV/cm
so I'm wondering if the initial 24.5kV is enough to start a small amount
of breakout, which adds a little to the Ctop, which brings the secondary
down a little into better tune, therefore more Vtop, more breakout, etc,
and so on, eventually multiplying to give you copious breakout with the
coil effectively in tune. I guess this process might take place over a
few consecutive bangs, or perhaps within one beat?
Paul,
It may take place both within a beat, and also over consecutive
bangs perhaps. How critical would the tune point be, after breakout,
considering that the Q will drop due to the streamer loading?
Do you have difficulty quenching? At 12% detune the primary waveform
doesn't form a very good notch - it's still at 50% of the max amplitude,
thus retaining 1/4 of the bang energy in the primary. If some
corresponding detuning of the secondary/topload as described above
wasn't also occuring, then perhaps you would have trouble putting out
the gap arc.
I don't think the quenching depends much on a perfect tune or
perfect notches. In work I did, I saw little difference in quench-times
as those parameters were varied some. Basically, it seems the gap
will quench when the energy is gone... sucked out by the streamers,
or lost in losses. More energy transfers will occur in a given
amount of time with tight coupling or mistuning, but the total
time to quench will be about the same provided the streamers
are long, to provide loading and assist the quench. Coils such
as Barts usually quench at the second beat notch perhaps.
It seems that in a trade-off, tight coupling is preferable to first
notch quenching. The overall losses may be lower that way,
or as Bert H. suggested, the multiple energy transfers may
help extend the sparks. The scope traces were somewhat
blurry in my tests, so I may have gotten a mistaken result.
I can't be sure. I think where quenching becomes very important
is in high bps systems, where the gaps may power-arc more
easily, or where a delayed quench may steal time from the
shorter cap recharge cycle.
Is it the case that some of the art of tuning the TC is
finding just the right amount of primary detune so that the breakout
drops the sec Fres enough for the majority of energy to transfer? That
certainly makes tuning a much more interesting operation! I suppose in
that case, tuning for deepest primary current notch would be the thing
to do, since that implies max energy transfer is achieved during the
beat. It surely can't be coincidence that four of the five coils looked
at so far use a fair amount of detuning.
The need for detuning seems real, but I'm not sure the nature of the notches
matters much. The peak energy may be very important though. I may
have missed something in my old tests though, so I'm certainly not
too sure about what I'm saying. It may be mostly the actual tuning
relationship that is most important, although I can understand how
the beat notches affect all this. In any case, I've noticed in some
coils, that the gap arc dims as the streamers grow, showing that
the energy is being drained out faster from the system, and shortening
the quench-times (when they dim) as would be expected.
If we take the tuning setups shown on the web page to be the those that
give optimum performance for each coil, then we can use the detune
percentage to calculate the effective capacitance of the streamer load
(assuming the streamers load the secondary down to match the primary),
and from that we can estimate the total streamer charge, which in turn
we can compare that with the 'surplus charge' value calculated on the
basis of how much larger the MCTV is above the breakout voltage.
I wonder if the best tune point is a sort of compromise, rather than
being tuned for max streamer length? Yet, I tend to think the coil
does have to be tuned for max streamer length or so, whatever
the reason. What do you think about Malcolm's suggestion
about the "offset" tuning, which emphasizes the lower of the
split spectrum frequencies? I would be interesting if there was
more than one reason for the lower tuning. I guess if detuning
helps to form just one streamer in some coils, that's another
spark-length benefit from detuning.
It would be interesting if Bart would run his coil at a lower power,
so that he gets just one streamer, then tune to make that
streamer as long as possible, then see if it "scales" with my
TT-42; same freq offset, same ratio of top volts to toroid breakout voltage
stress, etc. This may be especially meaningful if he is also
using 120 bps, as I am.
I wonder if I'd see much difference if I tune at 1/2 turn intervals,
instead of only at 1 turn intervals as I'm presently doing?
Cheers,
John
Let's just try some ballpark figures for ba0:
The secondary+topload has a total shunt capacitance of 40.7pF and to
detune by 12% requires 51.1pF, so we suppose the streamers are providing
the extra 10.4pF.
Now if the streamers are clamping the surface field at 26kV/cm, that
takes a topvolts of 26/0.043 = 605kV. Thus our 10.4pF of streamers
holds up a charge 10.4pF * 605kV = 6.3uC.
Now the MCTV would allow the topvolts to reach 880kV, which exceeds the
breakout potential of 605kV by 275kV, and thus the coil is capable of
supplying a surplus charge over and above that required to reach break-
out of 275kV * Ctop = 275kV * (30.5pF+10.4pF) = 11.2uC, (The 30.5pF is
the topload capacitance).
Thus if the tuning is 'ideal' for the coil, the streamers are holding up
around 6.3uC, but the coil is theoretically capable of delivering
11.2uC. It's nice to see these two numbers landing in the same
ballpark,
and that the first is less than the second.
For now, the latest coil models are uploaded to
http://www.abelian.demon.co.uk/tssp/cmod/
and I'll see about adding some extra rows showing the streamer cap and
charge estimated from detuning.
--
Paul Nicholson,
--
--part1_82.1bbbec3f.2a167404_boundary--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.