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

Re: Re[2]: Capacitor Size (long)



Original poster: FIFTYGUY@xxxxxxx

In a message dated 10/8/05 12:41:04 AM Eastern Daylight Time, tesla@xxxxxxxxxx writes:

Original poster: "Lau, Gary" <Gary.Lau@xxxxxx>

> I disagree on two points.

>1) The width of a static gap should NOT be set to
>achieve 120BPS.  While I agree that adjusting the
>static gap width does permit one to adjust the
>BPS to almost any desired value, it ignores the
>most important consideration in setting gap
>width.  The width must be such that the breakdown
>voltage does not exceed what's safe for the NST.


I think we would all agree that this would be the most prudent setting without a safety gap. However, I think it can't hurt to open up the main gap to where it works best, so long as the safety gap is adjusted properly.



> In my earlier post today where I cited >measured BPS measurements for my mini coil, the >BPS was between 157 and 257, depending on the >tank cap size. The gap width was fixed and had >been set by connecting only the NST to the gap >and adjusting it just wider than where it began >to arc. Had I adjusted the gap wider still to >where the BPS measured 120, it would have been >operating at a higher and unsafe bang voltage.


Well, Gary, you scoped the bangs - but did you measure the actual voltages at the main gap and the transformer terminals? I'm not being facetious - I'm the last guy you want trying to do any meaningful measurements. Better someone with good technique and equipment!



>It's not clear what voltage you consider the plus >and minus peaks to be. Depending upon how close >to mains-resonant the cap is and if you're >saturating the NST core, mains-resonance will >occur to varying degrees. If it didn't, nothing >would happen when the Variac was at 50%. The >point is, the NST & cap will achieve 60 Hz >resonant rise to the point of the gap breakdown >voltage, and if the gap is set to 30kV, a 15kV >NST can easily ring up to that. So it's not like >a 15kV NST can't charge a cap to greater than >SQRT(2)*15kV and that's what the gap must be set for.

Granted, but that takes more than AC half-cycle to resonate up to higher-than-normal voltage.


> 2) A static gap firing more than 120 times per > second is not uncommon, it's the norm. I base > this opinion on scope measurements of my coils.


So assuming folks "properly" set their main gap by using just the NST output alone (no prim cap), how come the break rate is *higher* than 120? How is resonant charging occurring to drive the cap charge voltage up, if the bps is higher than 120 and the cap value is properly LTR?


>Think of it another way.  For any given NST size,
>the recommended cap size for a 120BPS sync gap is
>about twice the value as for an LTR static
>gap.  The goal for both cases is to achieve the
>largest bang as is safe, so the bang voltage must be comparable.


I'm inclined to believe that larger bangs energies make longer sparks, all other things equal. And it seems we need a minimum of 100 bps, for some reason.



> Original poster: "D.C. Cox" <resonance@xxxxxxxxxx> > > It's done with the setting (distance) of the > spark gap itself. It is adjusted such that the > gap can only fire when the AC waveform hits its > plus and minus peaks, thus insuring it fires at 120 pps.


If this was true, then what's the point of running at 240 bps? Are the bangs at the zero-crossing smaller ones, where the cap isn't charged fully, yet any additional bang gives a contribution?



> If the gap is set too close together it could > re-fire more than 120 times per sec but this is uncommon. > > Dr. Resonance

    And so Gary Lau disagrees...


> > Thursday, October 6, 2005, 10:08:55 PM, you wrote: > > > > Original poster: "Lau, Gary" <<mailto:Gary.Lau@xxxxxx>Gary.Lau@xxxxxx> > > > In my early naïve days of coiling I thought that > > 120 BPS was the natural and normal mode of > > operation for a static gap. I know better now > > and don't wish to perpetuate that misunderstanding.

    I'm glad you brought it up, because it got me thinking....


> >> Original poster: "D.C. Cox" > <<mailto:resonance@xxxxxxxxxx>resonance@xxxxxxxxxx> > > > >> Erms resonant value times 1.57 for best effective C value. With 30 > >> mA its near .008 and doubles to 0.[0]16 at 60 mA. This works best with > >> a static copper tube sparkgap firing 120 pps.


Well, at least one thing we're certain of with proper *rotary* gaps: the bangs aren't occurring any *faster* than the design break rate.
So here's what I'm thinking, how Gary Lau and D.C. can *both* be right:


We know we need a minimum break rate to propagate nice long streamers. This means there's some time constant in the air that allows streamer channels to stay open/charged/whatever. I'm still waiting for somebody to give a detailed explanation of why that is - which ion species it is that we have to worry about, how many need to be excited to what energy, how the energy dissipates, and how to model the streamer up from the small to the full-size. It may seem academic, because we're stuck with sparking through air - but see below.
There seems to be a wide performance in spark gaps, depending on the design. Folks have tried all sorts of things, and I'm not going to try to reinvent the wheel. Especially when I don't have a meaningful way to accurately measure the many important parameters at once. But it seems like the things we try to quench the gap would be the *opposite* of what we want in order to better propagate the streamers off the secondary.
Likewise, if the secondary streamers appear to "like" 100+ bps, why wouldn't the main gap "like" that is well? After all, they're both high-voltage, high-frequency, pulsed discharges in air. And on a SGTC, the two rates are the same.
So what if - by picking a break rate over minimum, and by discharging repetitively through the same air space, we are doing the same thing on the primary gap that we're trying to do at the streamer: decreasing the voltage necessary to propagate the next spark.
I think this would explain a few things:
With a properly set gap (barely over breakdown for X-fmr), the break rate goes up past the nominal 120 bps. This could be because the cap does not have to charge to full voltage to break down the gap. Since it doesn't have to charge to full voltage, it can be charged quicker by the same X-fmr. That's why similar performance from a LTR cap - it's not being charged to full voltage anyway, so the increased size doesn't matter much.
A way to fix this would be to open up the main gap - as D.C. advocates (with an adequate safety gap that doesn't fire prematurely or frequently, of course!). This would move the main gap's necessary firing voltage back up, and increase performance.
Force-ventilated gaps, and gaps with serious cooling surfaces take the heat/charge/whatever out of the gap region, or replace energized/charged air with fresh. This would bring the gap firing voltage back up, and increase performance.
However, we have a double-edged sword. Whatever we do to increase the breakdown voltage for the gap, by replacing the air or increasing the gap distance, also hurts us by increasing the on-state resistance of the gap, and therefore the gap losses (more gap losses also make it harder to keep the gap voltage up!). And so it spirals, with marginal gains.
We need a way to keep the excited/energized air in the primary gap, so we don't waste the energy we've already put into it. Keeping the gap width/number low would keep the excited air in the gap where we could use its low resistance for the next bang. Keeping this gap as short as possible might keep our losses low as well. But we also need a way to make the gap fire at only the proper high voltage. Perhaps an interesting case with the magnetically-quenched and/or triggered gaps?
Since RSGTC's can run even bigger caps with better performance, there must be some way the humble NST is actually getting more energy into the cap than the resonant charging numbers say it can. And if we can improve the static gaps, then we can do the same to the rotary gaps.
And thus gleaning what goes on in the secondary streamer helps us at both ends.
Must be nice using solid-state switching, where everything in the primary resets for the next bang.


-Phil LaBudde