magnetics design -- 60mJ energy impedance matching

I

gives

Everyone seems to be thinking voltage transformers when current transformers would be more appropriate. Bigger wire and much lower turns counts. And yes, a 50 to 75 cycle 1 kHz pulse would solve many problems.

Well, I've got it working already with a CR2032 __without__ the transformer. The problem I wanted to wrestle with was the 100' away part.

What I'd like to do is allow most anything to work. Some people _like_ the idea of having their wire small enough to put into one of those fishing line disks for easy packing around in the pocket. I'd like to support that, if possible. Obviously, if it isn't possible at all, then ... well ... I'll still try to do the impossible. ;)

Distance gives you a lot of things. One obvious one is ease of tracking and giving a decent baseline for angle measurements for height, if someone wants to try that. Crowd safety could be another (the ability to _see_ someone approaching and to then have the _time_ to get to them before they reach the rocket site can be a help and clearances aid this.) To be honest, it may be even further.

I've done a LOT of that in my life. Had quite a number of explosions, too, due to things like hairline cracks in the fuel. I can't even remember how many bags of sand I've hid behind! :) And good thing.

This would be mostly for casual users (and kids) and this is a hobby project, not a commercial one. Wire is cheap, safety is worth a lot. And I've got to put a lot more thought into details, yet. That will take some practice with this to learn what is good and what is bad about what I start out with. Eventually, I'll get enough of the details right enough that I'll feel comfortable passing it on.

Jon

P.S. For me? I'd be using picric acid and potassium chlorate and much worse, if I can lay hands on it, and all of it in a steel cylinder with a turned steel engine nozzle I design and make. :P So for me, lots and lots of distance is a _very_ good thing. But I'm using the flash lamp method for anything I do, not this device. I get precision timing of multiple energy pulses, that way. No current, just a nice 6kV trigger on a wire.

Yes, it is balancing between the various losses versus cost, life, and efficiency.

One of the FIRST things I did was to look up current transformers. Not because I know what I'm doing -- it's plain enough I don't -- but because of the word "pulse" that often is included with the phrase "current transformer." I was looking for pulse stuff and sure enough that came up right away. However, I quickly realized from reading:

That they were designed for something else and, due to my own lack of knowledge (ignorance), I also knew I wasn't competent to decide if they would fit the application.

So I dropped the thought and went "back to the books," again.

Yes. I'm really beginning to see those volt-second Webery things in every waking moment of my life, now. Damnable things. I am learning to like __micro__-volt-second stuff. Tiny bits of volt-seconds are cool. Mega-volt-seconds are uncool.

Jon

Yeah. I'm getting a small clue, gradually. Since I figured on using a micro anyway, chopping is a doable approach. More parts, more expense, etc. But one cannot argue with results, either. I'm still going to give a shot to the massive P36/22 pot core (54 grams!) and see what I can dish out with that bastard. But then I expect I'm going to go back to the H-bridge choppy way of doing it and use a tiny little bump of a transformer on the other end.

The universe can be cruel. I want a free lunch, once in a while.

Jon

Okay, thanks. I liked your idea of a nylon bolt and nut. I can go there. $2 for 54 grams of mass is pretty good, I figure, so I won't look a gift horse in the mouth. And thanks for the time.

Jon

...

I have the impression you are thinking of the reflected impedance as being in series with the primary, rather than in parallel, as model at (eg) (near end of page) shows.

...

I have the impression you want to transfer most of the energy in a large initial pulse. Couldn't you use a small xfmr, ringing at a few kHz, to transfer the energy?

--
jiw

Yes, I was thinking that and it doesn't makes sense what I was thinking. Your point _does_ make sense to me. I thought that way because a diagram I saw __replaced__ the primary with the reflected impedance. I had considered the thought of it being in parallel, but because of the diagram I saw I set my own thoughts aside. I had since come to the conclusion that had to be in parallel, reflecting on the quandary posed by my own hypotheticals and the very real fact of others' experience, resolving that was the only choice that answered both things simultaneously and figured on showing it to myself, later on.

Yes. And I don't mean to say I fully apprehend all the details, but the problem I see here is that if the primary is "ringing" around a lot, I'm going to lose a lot of energy to the wiring out to the primary. I liked the idea of ringing because it can be used to cut down on the volt-seconds (so I imagine, anyway) requirement... but I expect to lose out to wiring losses, then.

(Regarding the volt seconds: the first positive part wouldn't be entirely compensated by the next negative going part, but that would be more than the following positive going part... etc., but it's better overall [I think] than going for the critically damped case where the hog's share is all on the first part.)

Ringing may be an approach with higher primary inductance (lower currents?) Now that you are making me think, the key is to minimize this:

Integral from 0 to 50ms of: (Vc(t)-V_primaryinductance(t))^2/R*dt

(Vr(t) of the wiring is the difference: Vc(t)-V_primaryinductance(t).)

I'll think a little on that.

Jon

I assume that you are going to dump the charge from a capacitor into a

100 ft transmission line with essentially a 0.6 ohm load at the far end. A voltage step change. What is the waveform going to look like at the load ?

Same question re-worded.

--
Best Regards:
                     Baron.

Well, a long line with a load on the other side of a transformer is my hope. It doesn't work without something like that, otherwise. All the energy winds up in the long wiring if just directly connected.

The waveform isn't crucial. The discussion I had with those who make and sell the squibs is that the current at the squib cannot exceed

40A, as there may be some problems elsewhere (bonding connections, clipping to the leads, etc.) before the tip gets a chance to heat. The pyro material has very poor thermal conductivity, so it heats up quickly and, so long as one isn't too slow about it, will ignite.

Jon

John, I thought I'd comment on this thread one more time.

Thanks for the input. I've had some more time to consider the details and your point that "it's a bit better than a rule of thumb" is nail-on, forgiving the pun. ;)

You explained where the AcAw product arrives, well. I've come at it from different angles, arriving at similar places. So it feels a lot better to me. Here's one angle I tried:

mmf = N I B = U0 Ur H H = mmf / l dH = (N / l) dI dB = U0 Ur dH

so,

dB = (N / l) U0 Ur dI

introducing time,

dB/dt = [(N / l ) U0 Ur] dI/dt

But dI/dt = V/L = (V l) / (Ac U0 Ur N^2)

So,

dB/dt = [(N / l ) U0 Ur] [(V l) / (Ac U0 Ur N^2)] = V / (Ac N)

Lots of stuff canceling out, which then rearranged gives me this,

Ac N = (V dt) / dB

Using finite differences to replace the infinitesimals gets me the same place. Units work out, as well. (I gather that Tesla is just Joules-seconds/Coulombs-m^2.)

Things seem to make more sense. mmf is equivalent to volts, but driving a hypothetical magnetic fluid through a hypothetical magnetic reluctance. Since electrons moving across a linear magnetic field have a force vector acting on them which is always perpendicular to their motion vector, they spiral around. But this also means that forcing electric field sources (electrons) to spiral around, generates a linear magnetic force field with what amounts to a + and - end. That acts on vacuum, air, or whatever, to induce a "current," so to speak. H is like a local field strength (similar to volts/meter); B is tied to H, but includes the effective permeability; etc.

I'm still working at the concepts. The main thing is that the window area thrust into the AcAw product makes a great deal more sense, now. I see how and why and at least some times when not to go there. There is another factor, current density in copper, that comes into play at times. And I can see the replacement of dt with 1/f in equations, as well. So it is slowly coming together.

I still need to see how to develop mmf=N*I as a consequence, I assume, of the spherical integrals in Maxwell. But that's for another day.

Jon

I never looked into it. See

John

Result = 1, so Jon's units are correct.

Yes, they give it (in SI units) as mass*time^-2*current^-1, which yields what I wrote above (using the derived-SI unit of Coulombs, rather than the primary unit of Amps.)

I almost always go back to SI units to do a dimensional analysis check on any expression I see to make sure the units work out. If they don't, usually it means there is a constant whose units I didn't apply correctly, I misunderstood the units of the variables involved, or else there are hidden constants with units the author didn't include but which points out my own need to go track it down. There is another possibility, of course, which is that the author didn't know what they were quoting well. Which means setting that aside and looking for better advice.

I'm more of a 'counter' type person. I prefer thinking in terms of objects I can count, like electrons into Coulomb units, than in terms of combined units like Amps, which SI prefers because of our ability to measure, right now. And I keep in mind a few things that also make sense to me from classical mechanics, like angular momementum which is easily derived as a necessary consequence of assumed Euclidean space and linear time, so Joule-seconds are meaningful to me for that reason and for keeping one idea about electron spin in mind.

So Volts become Joules/Coulomb to me, which is easy to understand from accelerating electons across a pair of charged plates in a vacuum. Complete sense there. Ohms are in Joule-seconds/Coulomb^2, Farads are in Coulomb^2/Joule, and Henries are in Joules-second^2/Coulomb^2. Clearly, then, the multiplication of Henries and Farads yields seconds^2, which must be square-rooted to get seconds out. Etc.

I am in the process, now, of going back to understanding Maxwell, conduction current, displacement current and dielectrics, E fields, H fields, Poynting vectors, which if I'm guessing right moves me towards a closer understanding of near-field and far-field, as well (out of phase nearby moving towards in-phase further out.) I need to factor in ideas on back-to-back electric charge motion in a loop which from a distance appears to be no motion of charge at all, quantum fluctuations (1/2*k*T in each of 3 dimensions), etc. I've never taken it on and I can see I need/want to. I'd like to get to the point where I can derive mmf=N*I from a more fundamental understanding.

Jon

Nifty!

Jon

Works in google, too:

*seconds)%2F(Coulombs*meters^2))

Jon

you want a transformer that doesn't store energy. there's no such thing.

I'd consider using a buck converter.

If you are referring to my comment about "no energy stored there" I meant I don't want an energy source at the rocket end. No batteries, etc.

If you are referring to the transformer idea itself and not to that comment, then transformers aren't designed with an eye to energy storage. Flybacks are. But not transformers. So far as I'm aware, anyway.

That would likely be unworkable in this application. At least, although I have built one before I still don't have a clear idea what to consider here that would work for the intended purpose.

Jon

No he doesn't.

A transformer is capable of storing energy, but until you put some energy into it, it isn't actually storing energy.

He's asking for something where the energy stays in his pocket until he is at a safe distance from the rocket, at which point the energy is sent down the wire to the rocket.