John,
There's two ways to get a faster rate of transfer. You can use a higherfrequency (in a way) or a tighter coupling. Using a tighter coupling willreduce gap losses during the transfer, but the gap will be harder to quench, so it probably won't quench on first notch unless a ground strike occurs.If coupling is too tight, you'll get racing sparks and pri to sec arc-overs.Using a higher frequency will mean the gap is firing for a shorter time. The rate of energy transfer will be faster (from a clock point of view, not from an RF cycle point of view), so the losses will also be speeded up. In the end the gap losses will be the same. The key is how you measure *rate of transfer*. If you measure it with the clock (in microseconds), then yes a high frequency will give a faster energy transfer. But if you measure by the RF cycle, then for a given coupling value, the same number of RF cycles will be needed to reach the notch whether the frequency is low or high. Think of it this way: Suppose you are shoveling coal. You shovel at the rate of 1 shovelfull per second. But each time you handle the shovel, some coal (10%) falls off the shovel and never makes it into the dumptruck you are trying to fill. This 10% of the coal represents the losses. Now lets say you start shoveling twice as fast. You shovel at the rate of two shovelfulls per second. Now 10% of each shovelfull will still fall off the shovel each time you handle the shovel. In the end of the job, the dumptruck will be full of coal, but there wlll be another 10% as much forming a pile of coal on the ground that fell off the shovel. This pile will be the same size whether you shovel fast or slow. The coal losses are the same for fast or slow (high or low frequency) shoveling (or coiling).
I think the "rate" is the confusion.... let me see if I can work it out in single point questions...
Now 100 kHz to 500khz, the amount of time spent over the gap will be less... (clock point of view?)
BUT it will still take for example 10 cycles to complete.. (RF cycles?)The losses per cycle will still be the same (10% ?), as there is still the same number of cycles..
Instead of losses say in 500uS the higher frequency will have all the losses in say 100uS.
So as the losses are "compressed" into a shorter time frame, then the gap will be "hotter" and be harder to quench.
I see it as 100uS 10 cycles, or 500uS 10 cycles.I do not class myself the cycle count as a factor, since the cycles will really remain the same (all other factors been the same that is). It is just a matter of clock time I was referring too.
On this basis the spark gap will probable run hotter and be harder to quench, but regardless of that for the moment, energy still went from 500uS down to 100uS. Losses are the same per cycle, and will still need 10 cycles.
Another point I was trying to get at, is a higher frequency should also ionise the spark gap a lot faster, which will conduct a lot better, hopefully reducing losses. Though on the negative side of things, transferring the cycles in a shorter time frame could make the gap harder to quench. If this is the case then using a higher frequency is probably very bad from a quench point of view...
I think it is the same problem as passing higher currents over the spark gap (hence large primary inductance to help counter act this to a point) I am looking at it from the point that faster and harder is better (which it is) BUT, the spark gap is the let down which in itself counteracts these "improvements" and probably will make quench time a lot longer. So if the spark gap remains a constant (just for example) of 100uS to quench, then the higher frequency could cycle 50 cycles across the spark gap in 100us, in which case a lot more energy is lost. It will still take 10 cycles to transfer energy, but would not quench until the 5th notch..
A lower frequency taking 10 cycles in 100uS would mean 1st notch quench and 4 over all less trips from primary to secondary. Over all, the higher frequency could have 50 cycles, (5 notches) so if you take each round trip (per notch) as a waste in energy of 10% then the higher frequency would loose 40% more energy. Really it all comes down to keeping the spark gap transfers in all possible ways to a minimum...
this is now at the point I was referring too. That if you use a higher frequency, then you have to come up with some method to speed the RSG break time up, as hard as it may be.... *if* it could be done and quench time was not a factor, then more current and higher frequency should not be a problem, over all the system should work a lot better... I think...
I suppose the idea is a bit of a dream in a way, though when you progress to SCR based coils the frequency and notch times are not factors at all. Really there is no reason not to use MHz range as the SCR (spark gap) will always "Quench" at the first notch....
I think yourself , myself and Bart maybe looking at the "transfer speed" in slightly different ways. I am looking from a frequency vs. uS time, though I suspect Bart and yourself are maybe looking at a RF cycle point of view....
So really, lets try and round up a little in another way...If you use a higher frequency, higher current, less uS across the gap, 10 RF cycles and all spark breaks out and the tank cap has zero energy left...
If you do the same as above but with low frequency then the only difference between the 2 systems is the total time taken from the tank cap to spark output.. A higher frequency will make this time a lot less.. a lower frequency will just take a longer time.. over all if it sparks to ground and drains the tank cap, then quench times are meaningless, and RF cycles assumed to still be 10 in both high and low frequency. . So if you rule out the quench problems, higher frequency should work just the same just faster.
Chris _______________________________________________ Tesla mailing list Tesla@xxxxxxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla