During testing I'll always aim the waveguide or horn antenna at the sky...

Are you kidding me? Did it ever occur to you that any plane flying in that area will basically get EMP'd and will crash into your backyard? Please consider pointing it downward into ground, preferably surrounded by a wall of dirt on the sides, instead of upward.

 

Aircraft are mostly surrounded by aluminium shielding and the electronics themselves are generally inside a metal case of some sort. The actual 3 cm waves would be (mostly) deflected by the bottom of the fuselage. I don't think an aircraft flying 6 miles up will be in any danger. I don't think it's quite that easy to make a practical EMP weapon. If it were, fighter aircraft and bombers would be routinely falling out of the skies during their missions. Magnetron technology has been around since the 1930s. I've never heard of a single instance of an aircraft being knocked out of the sky by one either intentionally or unintentionally. Of course I'm sure quite a few have been shot down after being detected and tracked by magnetron powered radar.

And as far as the people inside the aircraft they are basically surrounded by a big faraday cage. I guess someone flying an ultralight or powered parachute or a glider made entirely of composite materials might be in some danger but the likelihood of someone flying one right across the path of one of my pulses is exceedingly unlikely. Especially in the location where this device would be operating.

 

Yes the risk is low, but still, would it be that hard to point it the other way where you can be sure no one can get hurt?
Anyway, what is the overall goal of this project? Just to be able to fire the magnetron, or to do something useful with it?

 

If you mean for testing, I don't think pointing the horn antenna at the ground would be a particularly great idea. I guess some of the pulse would be absorbed, but the rest would be reflected all over the place in an unpredictable manner. And for testing I'm going to be right there. I would see and probably hear any sort of low flying aircraft and probably commercial aircraft as well. I guess I could avoid testing on a cloudy day.

As for the goal of the project itself, I think discussing that in this thread might be a bit too off topic. It has already drifted pretty far off as it is. I'm sure at some point I'll start a thread about it in the projects forum. I'm planning to publish a whole blog on this ambitious DIY project with photos and updates on my progress. Based on what I've said so far I think it should be pretty easy to guess what the aim of the project probably is. A typical application for something like this would be weather radar, but I've already said that I am planning to modulate the pulses for communication purposes. I've said that the device will never be aimed at people or buildings and since 3.25 cm microwaves are strictly line of sight that rules out a lot of potential applications. No more than 2 occur to me offhand. Especially if you rule out commercial, profit motivated, applications.

 

So 1 watt = 1 joule per second. Power = energy/time, energy = power x time and 0.000003 seconds x 1000000 watts = 3 joules. But it occurred to me that 1 MW is the output of the magnetron. The magnetron is not 100% efficient. More like 55%. I actually need to supply the magnetron with 1.8 MW of power (60 A and 30 kV) for 3 μs. My stored energy requirements are actually 1800000w * 0.000003s = 5.4 joules.

Since it's questionable whether the magnetron can make use of 100% of the energy in the pulse and since I may even end up wanting to use a spark gap in series with the thyratron so that the PFN cannot discharge down to 0 volts, 5.4 joules seems like the barest of bare minimums. The 7.65 joules provided by 17 nF may not be quite sufficient. Going to 20 nF would give me 9 joules. That may work, but having a 2x safety factor would be awfully nice. That would require 24 nF.

Using Q = CV and Q = I*t: I*t = CV and t = CV/I; So for 17 nF: t = .000000017 F * 30000 V / 60A = 8.5 μs of 60 A current. 24 nF would result in 12 μs of 60 A current. Or 4 times what I actually need. That seems like a pretty decent safety factor.

 

As for the goal of the project itself, I think discussing that in this thread might be a bit too off topic. It has already drifted pretty far off as it is.

I think it odd that you are reticent about the purpose of this project. Could it be that you are trying to make an over-unity device?

 

No. What would that have to do with a communication transmitter? It doesn't involve free energy or perpetual motion or try to cheat the current laws of physics in any way. It doesn't involve any scientific innovation whatsoever. It's just an X band microwave transmitter intended solely for communication.

Mostly my reticence results from wanting to continue to receive help from you nice folks. What if you didn't agree with the aims of the project? You might be less inclined to help. The nature of the project is a bit, uh, controversial. At least in the professional community. And while I might not be considered a scientific crackpot for attempting it since I'm not violating any laws of physics or attempting to discover new ones, I might still be considered a bit of a crackpot for attempting it.

I have no intention of keeping the project secret. It just didn't seem necessary to bring it up in this thread, but maybe it is. I'm quite tempted to lie and say that it is a moon-bounce communication system, but you guys would eventually discover that that is not true and be super pissed.

I'm also curious whether anyone could guess. You guys do seem very intelligent to me, and intelligent people tend to like puzzles. Even minor ones. The clues I've given should be more than enough and I've just ruled out the moon-bounce possibility. Why else would someone want to build such a device and aim it at the sky? If I mention that the parabolic antenna is going to be at least 15 meters in diameter and hopefully 20 maybe that would help. I'm making this too easy.

 

This thread has given me the solution to my problem: use capacitor storage for the pulses and that will keep the mains transformer relatively isolated from them. And also you couldn't generate such pulses from a standard mains supply anyway and maybe not even from your own generator. And I also now know how much capacitance I need to store the pulses.

But my original question regarding how to size a conductor to carry pulse current still leaves me puzzled. It's not just a theoretical curiosity. Everything after the output side of the PFN will be facing 60 Amp 30 kV pulse current. Thyratrons and spark gaps won't have a problem with such currents, but what about the windings of the pulse transformer or the conductors connecting all the components after the PFN? It still seems like being able to calculate conductor size under pulsed current conditions is valuable.

I guess the problem would be heat generation. The instantaneous power is either going to be zero or 1.8 MW, but it seems like conductor size wouldn't be based on either of those figures. The average power is very low, but it seems like I can't base conductor size on that either.

How about this thought experiment. A PFN generates 30 kV 3000 amp 1 ns pulses with an instantaneous power dissipation of 90 MW. The pulses are only generated once per hour. So the average power is very low. The output goes to ground through an AWG 40 magnet wire. What happens? Does the 40 gauge wire melt? Glow red for 1 nanosecond? Or would it be able to carry the huge pulse because it only lasts 1 nanosecond?