Dr. Resonance,If lower inductance charges faster, then this also means a higher frequency. There should be more chance of a good spark gap quench as the time across the spark gap is greatly reduced. If the secondary is charged faster then the spark should spark out faster draining the tank cap even faster... I think...
Higher frequency also suggests less primary turns, as a direct results should pull more amps from the tank cap, so an even better tank cap should be used..( higher peek current)
The problem now is that more current will be placed across the spark gap which will probably make it harder to quench. I am not sure if this would be an issue anyway is "all" the tank energy is pumped into the spark in just a few cycles.. Though due to the lower tank cap value anyway it may not even be a problem...
There is another problem with running higher frequency, and that is a lower tank cap value is needed, so it favours a higher input voltage in the order of 20-40KV with a tank cap of 10nF or there abouts. Obtaining current will need many parallel caps to cope.. Not much of a problem really.... Though now we have a much higher voltage across the gap, so the question is would the gap quench faster due to a higher voltage and a smaller tank cap running at higher frequency...And the ultimate question would the design work overall more efficient ?
The spark gap causes no end of problems, I think a higher frequency coil should be used with a solid state switch instead, though if running at lower voltages it causes its own trade off problems... Though using resonance you do not necessarily need to run the IGBT's at 50khz or more. This is the idea I am working on with my "high Q design".. I run an SCR which triggers at 100hz and pumps power into a resonant circuit of 2mhz.. In simulations it works great... real world........ Well not sure until I finish building it!
It is no different really than the 50hz NST pumping into a 100khz primary circuit... the spark gap is still make/breaking at 50hz but the frequency "through" the device is a lot faster... Actually my SCR design only runs at 100hz and discharges the tank cap into the primary in less than a single cycle so as soon as the current drops in the tank low enough, the SCR turns off. Energy is trapped into the primary LC which has a resonance of 2mhz... The secondary is also 2mhz so the primary should drain over the next set of cycles and charge the secondary as normal.
In anycase the secondary inductance is around 200uH. I hope the secondary will charge at a fast rate, and also hope that the frequency in itself will ionise the air easier around the toroid promoting spark breakout.
I ran some figures last night on the coil... CAPACITANCE Gain = sqrt (Cp / Cs)1,000,000nF (1000uF) primary, 0.016nF sec = gain of 7905. So 25V input on 7905=197KV
Q FACTOR(all approx) = 1mhz, 500uS cycle time ( 1uS per cycle) 500 cycles 25V tank voltage ( 12VDC input?) sqrt(25)=5 (rms cycle voltage) Secondary Q factor 100?
5V x 500cycles = 2500 2500 x Q (100) = 250,000 (250KV) JOULES = 1,000uF 25V = 0.312J 0.312 / 12.5 = 0.02496 sqrt *1000 = 158KVI also ran my normal Tesla coil figures much the same way, and they all come out the same voltage output, of around 200KV (10nF 10KV) So If I can actually physically build it then with 25VDC input it *should* output around a 20" spark... (10Kv 10nF=0.315J, or, 1,000uF 25V = 0.315J)
The system is limited by the tank cap's voltage of 250VDC so the max throughput will be 31Joules.... 10KV 200nF is only about 10 Joules...
I also think the very low secondary resistance will help make the sparks brighter along with the higher frequency. DC resistance is less than a single ohm, AC resistance according to JavaTC is 3ohms... Though I am using coax cable for my secodnary, so DC resistance could be 2 ohms and AC resistance maybe 2ohms.. in anycase far lower than normal secodnary coils which normally in the 20-100ohm area maybe.
A point which I do not think anyone picked up on was the coupling vs frequency... After doing some low 12V testing with coupling higher frequency is many times more efficient than a lower frequency. This could be more related to radio terms, maybe an old argument vs LW and FM transmissions of quality vs range.... Not by beat area of knowledge, though I Think a high frequency will also couple more energy a lot more efficiently than a lower frequency.
This brings onto a new idea of using a low coupling of below 0.1K to prvent the secodnary from effecting the primary inductance and lowing the current pulse within. I ran the figures in JavaTC also, primary was 1uH and secodnat 220uH or there abouts, mutual inductance I think was less than 1uH..
Needless to say my design is pretty much totally backwards to any Tesla Coil built ( to my knowledge).. I think the idea of a high frequency solid state coil as a high Q system as it uses safe voltages and only needs 100turns on the secodnary coil of a wide diameter! Hopefully will provide a higher current to the sparks...
I have built the coil, It was untuned and I used a IGBT which I will replace with an SCR soon as my simulation after I pondered over some design mistakes did not even work at all with a IGBT! Even so I did maged to obtain 1 or 2KV from a 10V input and it did run at about 2mhz. This suggests to me a Q factor of 100, though I used 100 as my Q factor calculation, 10V input time Q = 1,000volts. This was untuned and was feeding the secondary at 100hz.....
I do not know if it is correct to assume the gain is 100, I do not know if the cycle count is accurate though discharging a 1,000uF tank cap is not easy. It was simulation figures which I based the cycle count on and Q factor based on inital testing... probably will turn out wrong though...
Lots to ponder over! Chris----- Original Message ----- From: "resonance" <resonance@xxxxxxxxxxxx>
To: "Tesla Coil Mailing List" <tesla@xxxxxxxxxx> Sent: Thursday, November 22, 2007 9:48 PM Subject: [TCML] Sec design trade-offs and considerations
Chris raises an interesting point here. This is much like a catch-22 analogy.A lower inductance would mean faster charging of the magnetic field, however, max output potential favors a higher inductance and large dia sec coilform.Also, the IGBTs used in solid state coils like to switch at or below 50 kHz which usually again requires larger inductance and either large dia coilform or lots of sec turns. Too many turns and you begin to run into the tradeoff points between getting large inductance vs. too much resistance.Some experimenters use as many as 2,600 turns on the sec to keep the switching freq low while other experimenters prefer to use large dia sec coilforms. With solid state coil designs, I prefer around 1,800 sec turns and the largest dia. sec coilform possible. The balance here, of course, it the size of the sec toroid which increases with large dia sec coilforms --- and the cost rises with the larger toroid sizes.I use k value of 0.17 to 0.18 for most designs. You can do a lot of "what-ifs" in JAVATC to check the coupling values for different inner and outer pri coil radii.Certainly a lot of trade-offs to consider for optimum performance at reasonable cost.Bart Anderson's JAVATC program enables experimenters to try many different configurations to achieve desired performance, and then a cost analysis is conducted for each different configuration.My personal preference is to "best guesstimate" the cap size and then use the "auto-tune" feature in JAVATC to determine best pri number of turns to match. In most solid state coils values between 0.6 uF and 1.5 uF are common with pri sizes of around 4 turns. You can use less but then tuning points on the primary become very critical. Use 4-5 turns with a 0.6 to 1.5 uF pri cap and see what JAVATC tells you to use for the primary. I like around 1.0 to 1.5 uF with 3.5 to 4.5 turns on the pri tap. Then, I check the calculated JAVATC coeff. of coupling (k value) to see if it is in the 0.17 to 0.18 range. I make inner and outer radii adjustments as necessary to get the k value in this critical range. The idea is to couple as tight as possible but definitely to not "overcouple" and get into the "racing sparks" range --- usually a value of 0.2 or above.16 inch dia. sec coilform with 34 x 8.5 toroid atop a 24 x 6 inch toroid. These can be spun toroids or, if cost is a major consideration, the good old aluminum tape dryer duct type toroids work fine. Sonotubes coilforms or fiberglass work equally well.BTW, if anyone is working on a solid state design I have developed some "datasheets" that can be helpful to setup of over design and also a separate datasheet for pri-sec geometry. Filling in these sheets first with dimensions helps transfer the data into JAVATC for the test runs. Contact me off-list if you want them emailed to you. Also, if anyone needs help with JAVATC initial run contact me off-list for assistance.Many happy sparks, Dr. Resonance Resonance Research Corp. www.resonanceresearch.com----- Original Message ----- From: <FutureT@xxxxxxx>To: <tesla@xxxxxxxxxx> Sent: Thursday, November 22, 2007 11:39 AM Subject: Re: [TCML] quench times againIn a message dated 11/22/2007 11:47:26 A.M. US Eastern Standard Time, list@xxxxxxxxxxxxxxxxxxxxxxxxx writes:I am also trying to work out, that other than coupling and frequency which effect the tank "transfer speed" to secondary... can the secondary itselfbecome "easier to drive" to make the transfer quicker ? this is why I thought that a lower inductance would take less time to "charge" and the energy transfer would be quicker than a lower inductance.... though this could just be down to a higher frequency...cheers, ChrisChris, I'll ignore the arc-to-ground case which is a special case. Mostfolks like to see mostly air streamers I think. The transfer speed to the secondary is not the problem preventing fast quenching. The real problemis streamer (actually leader) impedance. If the streamers were somehowof a lower impedance, this would drain the energy faster from the secondary.Low impedance results in a heavier loading effect by the streamers. If the streamers were of low enough impedance, then there would be noenergy left to go back into the primary and prevent quenching. The effect of streamer loading reflects back through the system to affect the quenchtime. In the case of the arc-to-ground, the streamer impedance becomes very low, and drains the energy quickly. If the energy transfers to the secondary quickly, but can't get out quickly via the streamers, then there's a bottle-neck, a traffic jam. It's as if many cars are streaming onto a highway from various feeder roads, but up ahead a couple of lanes are shut down for repair.Now the traffic backs up. If the cars speed quickly to that bottle-neck, it won't do them a lick of good. They'll still have to slow down or stopuntil the traffic ahead makes its way through the constriction. Souping up the engines of the cars, or reducing the friction of the car's powertrain, etc. won't help. The only thing that will help is to remove the constriction, to open the lanes of the highway which are closed for repair. This opening of the lanes, would be analogous to reducing the impedance of the streamers of the TC. John**************************************Check out AOL's list of 2007's hottestproducts. (http://money.aol.com/special/hot-products-2007?NCID=aoltop00030000000001) _______________________________________________ Tesla mailing list Tesla@xxxxxxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla_______________________________________________ Tesla mailing list Tesla@xxxxxxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla -- No virus found in this incoming message.Checked by AVG Free Edition. Version: 7.5.503 / Virus Database: 269.16.4/1146 - Release Date: 22/11/2007 18:55
_______________________________________________ Tesla mailing list Tesla@xxxxxxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla