Big SGTC / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Big SGTC

1 2 next

Move Thread

LAN_403

monokel

Sun Sept 19 2010, 05:01PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

I'm planning a big SGTC. The secondary has a hight of about 1.6 meters and a diameter of about 0.4 meters.
I'll use a DC voltage supply (a three phase rectifier with 6 MOTs and voltage doubler yielding about 10 kV DC).
I have made an air coil with 5 H ; is has a brooks coil geometry.

I use a rotary with about 470 or 235 BPS.
I'm not sure whether the plastic disc of the rotary (consisting of polycarbonate) will withstand the heat beeing generated by the sparks. Has anyone made the experience that the plastic was melting due to the heat ?

Here are some pics:

The secondary

My MMC (220 nF, max. 30 kV, consisting of 576 WIMA FKP-1 caps each 220 nF, 1250 V):

Air coil with max. 5 H:

Rotary with about 2850 rpm:

The circuit:

This is a "classic" circuit for DC SGTC. The inductance 0.92 H is not ideal; but an air coil with the ideal inductance of 8.3 H would be too costly. With 0.92 H the cap is charged, discharged and charged again before firing.

Because of the assumed danger that the plastic disk was melting due to heat, I deliberated using another circuit: In the primary oscillating circuit, there's a static spark gap, but a rotary is used to control the charging of the cap. This rotary must eventually have a longer ON time than a rotary beeing in the primary oscillating circuit, because it is used to charge the cap.

An advantage of this configuration is that the current of the rotary is less and thus less heat is developped. Maybe you can replace the rotary by a semiconductor switch if the semiconductor could withstand the high voltages.

What do you think about this alternative circuit ?

ScotchTapeLord

Sun Sept 19 2010, 06:25PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

I do not know why you would want to limit the charging to the capacitor. With such a big system you will want the benefit of quenching in the tank circuit that the rotary gap provides along with a high current charging (that's a lot of capacitance!).

With a beast like this, you won't want to use resistive ballasting (which would include the RSG in the second circuit), as it severely diminishes efficiency... which is bad for a system that is going to demand many kW of power- like this will!

If you want to take some stress off the rotary gap, just use a stationary gap in series with it.

Regarding the inductor, if you have access to more MOTs, you could maybe rewind a couple with a good deal of space between windings and core and drop it in oil to get some higher inductance.

I myself am working on a similarly sized (a bit smaller) coil, but mine is still in the planning/purchasing phase.

monokel

Sun Sept 19 2010, 07:10PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

ScotchTapeLord wrote ...

I do not know why you would want to limit the charging to the capacitor. With such a big system you will want the benefit of quenching in the tank circuit that the rotary gap provides along with a high current charging (that's a lot of capacitance!).

With a beast like this, you won't want to use resistive ballasting (which would include the RSG in the second circuit), as it severely diminishes efficiency... which is bad for a system that is going to demand many kW of power- like this will!

I think you refer to the 100 Ohm resistor in series with the 2 H inductor. It could be the copper resistance if you take an air coil (2 H). But the RMS current of the resistor has an order of magnitude of 1 A, and so the power loss is about 100 Watt which isn't big compared to the total power throughput.
In some simulations with an ideal inductor and no series resistor the current through the inductor increased continuously to infinity, because the inductor could'nt transfer all its energy to the cap. But in other simulations (when changing some parameters) this phenomenon didn't occure even with no series resistor. Maybe I'll make some more simulations to clarify this.

In the alternative circuit the rotary is necessary. If you would short-circuit the rotary, the static gap wouldn't stop arcing because the voltage source would supply the current continuously.

Marko

Sun Sept 19 2010, 07:47PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Wow, that's one bigass secondary over there.

Truly - how much wire did it go into your air-cored 5H inductor? I still find it hard to believe it has as much as 5H, and resistance as low as 100ohms.

Couldn't you get away simply by seriesing/paralleling a bunch of mercury lamp or fluorescent tube ballasts?

Also, many people here advise using heat-resistant termosetting resin based plastics for the disks. PCB material (epoxy/fiberglass) is good up to like 300 degrees C... and since the disk is spinning in air this adds to cooling greatly. Small heatsinks(grooved aluminium studs) can be added between your electrodes and the disk to improve dissipation.

Marko

ScotchTapeLord

Sun Sept 19 2010, 10:30PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

I think I have a better understanding of what you're trying to do now. I like the idea of the semiconductor, better, though it would need to be made up of quite a few devices in series for the voltage. Also, if you want to be sneaky, you could even cut off the power supply when there is current flowing in the tank circuit using a current transformer and associated circuitry...

monokel

Mon Sept 20 2010, 03:05PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

Marko wrote ...

Truly - how much wire did it go into your air-cored 5H inductor? I still find it hard to believe it has as much as 5H, and resistance as low as 100ohms.

I used a bit more than 10 kg of copper wire of diameter 0.65 mm (0.025 inch) and length of about 3000 m. Number of windings is about 6100. If you want to construct an air cored inductor with specified inductivity and wire diameter and minimal wire length, you have to build it like a "brooks coil":
Brooks coil calculator
(In my coil, c=5.2 cm (Coil Height/Width))

Marko wrote ...

Couldn't you get away simply by seriesing/paralleling a bunch of mercury lamp or fluorescent tube ballasts?

I've no experience with such lamps, but I think they would act similar like a static spark gap (?). If you would replace the inductor with such lamps or a SG, the DC voltage source could be short-circuited if all the 3 spark gaps (the rotary and the two static gaps) would fire simultaneously, couldn't it ?

monokel wrote ...

In the alternative circuit the rotary is necessary. If you would short-circuit the rotary, the static gap wouldn't stop arcing because the voltage source would supply the current continuously.

Is this correct ? It sounds logical, but I'm not 100 % sure whether it's correct. A similar question is: Is it possible that an AC powered SGTC with a simple static spark gap can fire more than twice per mains period if the power supply is strong enough ? In other words: Will the SG stop arcing before the next zero-crossing of the supply voltage ? Somewhere else I've read that someone has recorded the voltages of such a coil with an oscilloscope and there were more than 2 firings per mains period.

monokel

HV Enthusiast

Mon Sept 20 2010, 06:52PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Nice coil.

HOWEVER!

That second circuit isn't right. The spark gap (switch) needs to be between the primary capacitor (energy storage) and primary coil. Your DC supply will charge up the primary cap, and then the spark gap will discharge that energy into the primary coil. The power supply doesn't have the peak current capability to discharge like you have it.

Also, get rid of the charging choke. You can probably use it between the DC supply and primary capcaitor, but NOT between the primary capacitor, spark gap, and primary coil. Remember, you want good quencing in the spark gap. A huge honking inductor in series with the spark gap will not provide good quencing. When you try to disconnect the circuit (spark gap), current will want to continue flowing and voltage will start increasing, which will simply try to keep the arc going.

Other than that, looks good.

Stick with your first schematic - thats on the right track.

monokel

Wed Sept 22 2010, 06:59PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

EastVoltResearch wrote ...

Remember, you want good quencing in the spark gap. A huge honking inductor in series with the spark gap will not provide good quencing. When you try to disconnect the circuit (spark gap), current will want to continue flowing and voltage will start increasing, which will simply try to keep the arc going.

Thank you for that hint. It's an error I have not seen before.

Maybe a rotary is, in fact, best for a high power DC coil, if it's well designed. For example one could make the ON time so that quenching is at the "first notch". This cannot be achieved with a static gap. But I like thinking about possible alternatives even if there are some disadvantages (but also advantages).

Quenching of a static gap could be facilitated be short-circuiting the gap after most energy of the cap has been consumed. This could be made with a series connection of some MOSFETs. I assume that such HV switches which can be turned on and off arbitrarily are available.
I've simulated the following circuit. I assumed 235 BPS, thus a cycle duration of 1/235 s = 4.26 ms. I assume that a controller turns the switch 1 on at the time 0% of the cycle duration and off at 40%. The switch 2 is turned on between 70% and 90% of the cycle duration. As soon as the steady state is reached after switching the power on, the cap will reach the voltage of 16680 V at 48% of the cycle duration. The spark gap will probably quench at some time between 70% and 90% of the cycle duration (when the switch 2 is on). In the simulations I assumed that this is at 80 %.
In this setting the real power consumed in streamers and the secondary and primary oscillating circuits is about 6600 Watts. The power supply delivers a real power of about 6830 Watts. The time-average of the current through the inductor is 865 mA, and the RMS current is 1.2 A.
The simulations were made with some idealizations, e.g. ideal switches.

Of course, there're some problems is practice, e.g. building the HV switches. The switch 2 must have a low ON resistance, so that the spark gap will really quench. If switch 2 accidently turns on before the cap is mostly discharged, the semiconductors will break down due to high current. Probably the BPS (beats per second) is limited because a minimal time with low current is needed to quench the spark gap.

Another interesting alternative seem to be the "SIDAC / IGBT SPARK GAP" modules of Terry Fritz. But IGBTs beeing adequate for SGTCs with > 6 KWatts seem to be expensive. And, if I see this right, a parallel connection of these modules would be problematic (?).

Hazmatt_(The Underdog)

Thu Sept 23 2010, 03:51AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

If you want to go down this path, you're going to want to look at Terry Fritz's SIDAC work like you already mentioned.

Personally, I think your diode there is going to explode on turn-on. You have two energized branches with currents when switch 2 closes, and the transient response is going to be a lot bigger then you expect. If you're going to take this approach, start with something reasonable like 1KV and get bigger from there.

monokel

Thu Sept 23 2010, 04:41AM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

Hazmatt_(The Underdog) wrote ...

Personally, I think your diode there is going to explode on turn-on. You have two energized branches with currents when switch 2 closes, and the transient response is going to be a lot bigger then you expect.

I see this in another way. The inductor doesn't change currents abruptly. And further more, if switch 2 is closed, switch 1 is open and the voltage at the inductor is nearly zero, thus the inductor doesn't change its current while switch 2 is closed. I've looked at the simulations again. The max. current through the diode occuring after power-turn-on is 2.03 A.

However, there can be situations when the diode explodes. If the time average of the voltage at the inductor is other than zero, then the current through the inductor will increase towards infinity bit by bit (of course, the 100 Ohm resistor would prevent infinite currents, but with poor switching times the currents could get very large). Probably, a zero time average of the inductor voltage in the steady state can be achieved if the switch 1 doesn't close immediately after the quenching of the spark gap or the opening of the switch 2; thus the current through the inductor can decrease in the meanwhile.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob