Sciencemadness Discussion Board - Primaryless Detonators

Primaryless Detonators

jpsmith123 - 3-1-2006 at 08:36

Attached (hopefully) is a patent wherein the inventors use a "high energy pyrotechnic" mixture, i.e., basically flash powder, instead of a primary.

If this kind of thing can work, then I can imagine taking it a step further: What about an electrical impulse fed into a CuO/Al mixture? IOW, suppose your "bridgewire" is, say, a small paper tube with a few hundred mg of pressed in CuO/Al mixture? Now, being that CuO is a semiconductor with a negative temperature coefficient of resistance, maybe it would be possible to get the joule heaing effect to spread out over a large enough volume of material that you would get a reasonable shock wave out of it. (I would imagine the power supply requirements would be much less problematic than those of a "conventional" EBW).

Attachment: 5945627.pdf (964kB)
This file has been downloaded 963 times

12AX7 - 3-1-2006 at 10:12

FYI, Cu2O (red) is a semiconductor, in fact it used to be used in "copper oxide rectifiers", which were then supplanted by selenium (cough!) and finally germanium and silicon. (Silicon carbide is now making inroads for high speed applications, though it's still pretty expensive.) AFAIK, CuO (black) is nonconductive, but I might have to take an ohmmeter to my jar of it just to make sure. The Al (or Mg if you prefer

) in the comp may be conductive, as conductive lampblack or graphite in bridgeless compositions. If the comp isn't conductive, a bridgewire might be used, or a high voltage pulse could be used instead.

Tim

jpsmith123 - 3-1-2006 at 15:20

If I remember correctly, both CuO and Cu2O are p type semiconductors; CuO has a bandgap of 1.2V while the bandgap for Cu2O is 2V or something like that.

Regards,
Joe

Boomer - 4-1-2006 at 03:16

Without giving that patent, this might have looked like those kewl "can i detonate xyz with flash?" posts. So the idiots from totse were right, you *can* det ANFO with flash - they just overlooked you need the flash in a re-enforced cap on top of some PETN, and inside a pentolite booster...

Joking aside, thats interestring. Reminds me of the good old times when I had all the nitric I wanted, therefore all the secondaries, but TATP was unknown, I had no mercury, and getting silver fulminate out of fun snaps was a PITA. I came up with the same idea, but lacking perc used armstrongs on top of the MHN. NOT safer than a primary, but they worked. Needed a thick walled tube though.

Even the other idea mentioned in the patent is not new (the flier thingie). I gave it up back then since they got cumbersome. Imagine a two-foot-long detonator, propellant charge in one end, shooting a ball bearing into the compressed MHN at the other end.

jpsmith123 - 4-1-2006 at 06:07

Those were my thoughts as well. In fact I hesitated a bit before posting it, as I don't want to feed the "k3wls", but I don't like to withhold any interesting info from others who share my interests in energetic materials, either.

Regards,
Joe

quicksilver - 4-1-2006 at 06:47

Just for discussion's sake what are the power requirments for conventional EBW dets? Is there a standard???? I have seen various designs and, frankly, they seem to require vastly differing juice. (what gives?)

But thinking along the lines of "flash powder" dets I believe it was Konske writing for the American Pyrotechnics Assoc. in a series of papers on flash itself that had done some research on the amount of electrical stimulation the standard "70/30" flash needed to fire. IIRC it was somewhere in a tenth of a Joule (the discussion was a safety issue re: mixing flash). Thus a "flash based" cap would be a dangerous animal when compared to most anything with a bridgewire. However I would imagine that the finer the bridgewire the greater the possibility of inducing stray electrical energy via radio waves, etc. - That issue has always been brought up and signs in the area of a large mine where I live warn of not transmitting during blasting (a light flashes on a dirt road, etc) but it seems that the transmitter would have to be a big one (not some 5 watt mini) to get any juice in the air enough to pose a threat. Static electricity seems much more of a significant issue.

Do any of the higher quality multi-meters have the capacity to measure static, or is that just too fine a curve for standard electrical measurment devices?

[Edited on 4-1-2006 by quicksilver]

12AX7 - 4-1-2006 at 10:35

The standard DMM has a 10Mohm (or so) impedance. You'll find it somewhere in the manual. Most are also only rated to 1kV, so you aren't going to be measuring much static without a voltage divider (needing a 100M for 10x, or 1Gohm for 100x reading). For example, a 100M impedance reading an object of 100pF charged to 10kV (reasonable values for a human walking over carpet, with wool socks, in winter) will discharge it to 3.6kV in only 10 miliseconds!

Special devices have been made to measure electric field directly, with typical input impedances on the order of teraohms (or more?). 1V corresponds to a mere 160,000 electrons per second flowing!

Of course, conduction through *air*, let alone over dirty surfaces, is going to kill that. Gigaohm resistors aren't easy to maintain...

Note: 100pF at 10kV is 5mJ. E = 1/2 * C * V^2, with C in farads, V in volts and E in joules.

Tim

jpsmith123 - 4-1-2006 at 17:27

Quote:

But thinking along the lines of "flash powder" dets I believe it was Konske writing for the American Pyrotechnics Assoc. in a series of papers on flash itself that had done some research on the amount of electrical stimulation the standard "70/30" flash needed to fire. IIRC it was somewhere in a tenth of a Joule (the discussion was a safety issue re: mixing flash). Thus a "flash based" cap would be a dangerous animal when compared to most anything with a bridgewire.

That's partly why I suggested using CuO/Al rather than a flash composition. The thought of a blob of flash powder with antennae scares me.

Quote:

Do any of the higher quality multi-meters have the capacity to measure static, or is that just too fine a curve for standard electrical measurment devices?

You'd have to buy or build some kind of high impedance electrometer for that. Depending on exactly what you want to do, maybe you could rig something up to deliver a certain amount of energy using a capacitor and a spark gap or a pulse transformer or something?

Boomer - 5-1-2006 at 03:09

"Just for discussion's sake what are the power requirements for conventional EBW dets? "

More or less standard are 1.5 mil gold bridge wires 40 mil long. These need at least 200 amps to deliver the 'burst action' (I^2*t integral) that explodes the bridge and dets
(not ignites) the PETN. 400-500 amps are recommended though, and should be delivered in preferably under a microsecond.

Remember that the capacitor bank only sees the cable impedance at first. Say you have a 2kV cap and a 50 ohm cable. So a 40A pulse travels down the cable, gets reflected at the detonator (low impedance)
and adds to the cap voltage. A 80A pulse travels towards the EBW, gets reflected and so on. The resulting current through the bridge ramps up in 40A steps while it heats up. It must reach 200 (better 400) amps
before the bridge evaporates. This is not too bad since that 'travelling' is done at 2/3 light speed.

For home brew EBWs using TR5 fuses for bridge+housing, I succeeded at 15µF, 1.5kV, 1.5kA within 600ns. That is a power rise speed of 3 TW/s btw. Succeeded means 100% full dets, it *starts* working
at 60% of those values. And this was with MHN, not PETN!
P.S: forget fotoflash capacitors, electrolytics are too slow! Or get thinner gold wires...

12AX7 - 5-1-2006 at 04:20

I'd hate to see the inductance figures there... if you have say 10uF and a mere 100nH, that's like what 160kHz resonance, or 1.6µs quarter wave risetime, no?

I should learn about RF and transmission line theory some time...

Tim

jpsmith123 - 5-1-2006 at 04:23

Hello Boomer,

I'm wondering, what was the current rating of the fuses you used? Also, what kind of switch did you use?

Regards,
Joe

Boomer - 6-1-2006 at 08:47

It worked with between 1A and 2.5A fast action TR5. The lower values were easier to get working, but below 1A the bridge tends to rupture
when you press the secondary in. 1A were still far above the 1.5 mil used comercially (hence my higher current requirement).

And the cap bank was massively parallel. Maybe that's why current rose so fast (would indicate 30 nH). OTOH I am the anti-RF guru but I feel that
feeding current from a cap into a wire is different from shorting a cap with a coil, even if coil + cable have the same inductance.

Marvin - 7-1-2006 at 04:41

Boomer, low impedence terminations reflect with inversion, so the reflected voltage will be subtracted from, not added to, the supplied voltage. I'm not convinced transmission line theory can or should be really be applied to this unless the pulse lengths are shorter than the cable.

It would be fair to say that the cable has an inductance that contributes to the LCR time constant that dictates the rate of rise of current.

quicksilver - 7-1-2006 at 07:54

Is there a "clean" method to disasemble a TR5? Do they come apart with a solvent or is one forced to bore a hole to access the wire material?

The use of a single amp (or there about) is much more attractive than building a multi hundred amp box that requires unique switching for long term employment. In the original tread in E&W forum much was said about the need for unique electronics for an EBW but at 1.5+ amps the testing equipment would be commonly available.

Boomer - 9-1-2006 at 01:55

Marvin, maybe I put that wrong. It is reflected in a way that once back at the source it doubles the current. I'm no HF guru, quite the opposite, but I have seen the oscillograms of the current ramping up. Plus this description was taken directly from Cooper, Paul W. - Explosives Engineering.

My digital scope is too slow to show the steps for short cables, all I saw was a current ramping from zero to 1500A within 600ns.

And Quicksilver, you got that wrong. I *do* use a one-amp-fuse, but you still have to *explode* the fuse wire within a microsecond. That means *not* the rated fuse current, but said 1500 Amps! Switching was done by simply touching two massive wires. If you need timing you could use a triggered spark gap, but looking at the scope the current rise time varies surprisingly little with hand-switching!

P.S. The top of the plastic fuse body was simply grinded off. This leaves a cylindrical housing with the wire in the centre. Fill, seal and enjoy

jpsmith123 - 9-1-2006 at 02:31

Boomer, have you tried anything other than MHN next to the wire?

Twospoons - 9-1-2006 at 20:04

Its because the voltage reflection is inverted that the current goes up! Its like connecting two batteries together - match + to + and you get no current, but match + to - and you get lots of current.

The applicability of transmission line theory is dependent on the speed to the switch, relative to the length of the line - if the risetime of the pulse is shorter than the pulse return time then you need to treat the system as a transmission line.

Rock Fragmentation System Using Goldschmidt Method

jpsmith123 - 10-1-2006 at 00:38

Here's an interesting patent related to the substance of this thread. It would have been more interesting if they'd gone into some more detail though.

Attachment: 5773750.pdf (265kB)
This file has been downloaded 661 times

Boomer - 10-1-2006 at 00:59

FPMAGEL: Absolutely no comparing this to an EBW:

- You cannot det gunpowder with an exploding wire. It would take gigawatts and a 3-inch diameter column. All you do is ignite it, *then* it explodes!

- The welder cannot deliver the required current. At some ohms for the bridge (especially at melting/evaporating), you hardly get 20 amps!

- Even if it could, the rise time is magnitudes too slow. Your wire will still lazily warm up while an EBW has already detonated the main charge!

Wonder why they use 2-5 kV capacitors?

And yes I tried RDX. No way to get it working with my system. NG works sometimes (better if absorbed) but is hard to seal. I need PE ...

quicksilver - 10-1-2006 at 06:57

Of the material I have seen there exists little information detailing the level of current for various materials (Urbanski, etc). Do you know of any study wherein such a gathering of material exists?
[What level of electrical energy for what material stimulates det?]

Of both GMAW & SMAW; MiG & stick, I believe that there are many welders available that operate @ 200 amps+ . The Miller (& Lincoln) stick welders particularly operate at 200 amps at the lowest entry level stick welders. Whille the more professional models operate at an astoundingly high level of current. That is actually the only "OTC" source of high current I could think of. However it had been mentioned that a level of 1500 amps would be needed. Searching my memory that would be available in some of the deep cut engine driven model welders available for ship-yard work. Capable of fusing together one inch plate in a single pass. Is that not "over-kill" for a mere wire?
When welding with a home unit of 200+ amps the stick could easly be 1/8". Would this not vaperize a fine wire?

jpsmith123 - 10-1-2006 at 07:57

Quicksilver I have to say that I don't understand exactly what you are getting at here.

With a conventional *exploding* bridgewire arrangement, as others have endeavored to point out, the idea is to deposit energy in the wire at such a high rate that it doesn't just melt or vaporize...it explodes.

Whereas a welder can produce a high average power of perhaps several kilowatts, that can easily melt the wire, the EBW power supply must at a minimum provide literally megawatts of power, but only for a short time.

In this thread I suggested that it might be possible to make some kind of a - for lack of a better term - call it "hybrid" detonator, whereby, instead of having to add enough energy from an external source to instantaneously vaporize a wire, the idea is to bring some finite volume of an electrically conductive energetic mixture up to ignition temperature very rapidly...maybe something analogous to running a piece of primacord lengthwise through some ANFO to make it more effective; i.e., the ANFO is like a power amplifier.

Axt - 10-1-2006 at 08:53

What if the wire carried its own energy. The following was actually done with a welder @ 115Amp the top is straight Aluminium foil for reference, the bottom is 2x 0.2mm platinium wires sheathed in Al foil.

<center><img src="http://www.sciencemadness.org/scipics/axt/bridgewires.jpg"></center>

You cant really tell much from the frames, the movie host I was using is down at the moment so you'll have to wait for that.

I did pull the energy figures for this to compare to conventional explosives once, cant find the exact figures right now but the base reaction Pt + Al -> PtAl ended up holding more energy per volume then HMX (or was that for Pd/Al vs HMX? forget, probably both).

Marvin - 10-1-2006 at 13:21

Boomer, Twospoons.

I think what must happen is from the point of view of the conductor,

1. When the switch closes it starts feeding current into the transmission line.
2. The voltage pulse travels along the line at V, while the current going into the line is determined by the impedence of the line.
3. When the voltage pulse reaches the end it hits the short termination, inverts and cancels out the line voltage.
4. The ground voltage edge travels back along the line canceling the voltage but the current is still flowing.
5. When this pulse hits the capacitor end current is still flowing but the voltage over the conductors is now 0v at full current, so the voltage over the same impedence causes an additional current to flow in the line, ie the current doubles but the voltage pulse is now V, and travels back down toward the EBD and when it hits it, the current through it doubles, the 0v voltage pulse after inversion goes back up the line for another step upwards to 3x initial current.

Axt, interesting idea, but assuming you could make a bridgewire fast enough, would'nt that essentially be a primary explosive in its own right with all the inherent disadvantages EBW's are used to avoid?

jpsmith123 - 10-1-2006 at 16:25

We're on the same wavelength. That's exactly what I was getting at with the CuO/Al.

Quote:

What if the wire carried its own energy.

Axt - 10-1-2006 at 19:37

Quote:

Originally posted by Axt
Pt + Al -> PtAl ended up holding more energy per volume then HMX (or was that for Pd/Al vs HMX? forget, probably both).

It was for Pd, Pt was slightly less.

Pd + Al (2890cal/cm3)
Pt + Al (2510 cal/cm3)
3CuO + 2Al (4976 cal/cm3)
HMX (2589 cal/cm3)

CuO/Al thermite releases considerably more energy per volume, but getting it to release it instantly would be much harder then a thin wire. There probably some issue with the energy release & when it actually comes into effect with the intermetallic wire, considering its no detonation rather lattice energy of alloying PdAl.

Quote:

Originally posted by Marvin
would'nt that essentially be a primary explosive in its own right with all the inherent disadvantages EBW's are used to avoid?

I dont think so, less sensitive to heat/impact ect. then the secondary explosive your trying to initiate. Its disadvantages are more likely to lie on the EBW side, cant just connect it to a battery, and I dont think a welder would work for actual initiating a secondary explosive, not that its practical anyway.

The <a href="http://ww1.webtop100.net/~62552/xmovies.webtop100.net/banners/intermetallic-bridgewire.mpg">movie</a> of the frames above.

Twospoons - 10-1-2006 at 22:03

Perhaps your bimetallic bridge wire could be made by plating one metal on top of the other? Maybe by plating Pt on Al wire. The intimate contact between the metals should allow the reaction to start much faster, which would be important in this case.

Cu0/Al by plating Cu over Al then oxidising the Cu?

nitro-genes - 11-1-2006 at 01:35

Quote:

quote]Originally posted by Twospoons
Cu0/Al by plating Cu over Al then oxidising the Cu?

Selectively oxidising the copper while in contact with aluminium is inpossible...But maybe thin aluminiumwire can be dipped in some sort of CuO/resin mixture to coat the aluminium wire. I have no idea if aluminium is usable as a bridgewire however...

Boomer - 11-1-2006 at 06:28

To put the welder idea to the grave: As jpsmith pointed out you need MEGAwatts, i.e. a thousand welders in parallel. Plus, you need it FAST. Think 3 TW/s or four *billion* horsepower / second rise speed. Cap banks have NANOhenries, a welder has henries inductance. That's a *billion* times more. The cap discharge produces a shock wave of tens of thousands of PSI, and a deafening bang that makes the ears ring - without the explosive! It powders adjacent materials by the shock before they have time to even melt.

Of course a welder could heat said Pd/Al wire to its reaction temp, as could a NiCd pack. But for EBWs said metals make no sense. You evaporate and then explode the wire, and the energy Axt quoted is lattice energy from alloying them. No alloy in a plasma!
And using such a wire in a normal cap would also defeat the purpose. It will not detonate a secondary (other than by cook-off), and putting lead azide in means you can as well use normal wire. This will go off before the Pd/Al reaction starts (I assume it needs over 300C?).

But we miss the idea of the thread in discussing EBWs. The question whether Cu/Al would work if heated electrically can be rephrased as "Can this thermite produce a shock to detonate a secondary?" Putting in additional electrical energy would be a waste, since due to the higher mass it would need even more power than exploding a bridge wire. The mix has to have the potential to det say PETN from its own reaction rate. Anyone tried that?

quicksilver - 11-1-2006 at 06:44

Quote:

Originally posted by Boomer
To put the welder idea to the grave: As jpsmith pointed out you need MEGAwatts, i.e. a thousand welders in parallel. Plus, you need it FAST. Think 3 TW/s or four *billion* horsepower / second rise speed. Cap banks have NANOhenries, a welder has henries inductance. That's a *billion* times more. The cap discharge produces a shock wave of tens of thousands of PSI, and a deafening bang that makes the ears ring - without the explosive! It powders adjacent materials by the shock before they have time to even melt.

Below is the best I could come up with and - in fact - they use just about the same concept as originally mentioned!

These guys are quoting about 180amps! (for some models)
http://www.risi-usa.com/0products/1ebw/new/new.html
Therefore I thought that the thrust of "the welder" concept was what is in line with commercial models.
It appears that commercial stuff stays within what would be considered "low power" in so far as the discussion thus far.

[Edited on 11-1-2006 by quicksilver]

jpsmith123 - 11-1-2006 at 17:51

Actually, my idea is that the "reaction rate" is artificially "accelerated" by the electrical energy input.
IOW, the electrical energy doesn't just "ignite" the thermite, rather, it ignites a certain weight or volume of the mixture essentially instantaneously.

I'll try to put some numbers in to clarify the idea. I'm just "thinking out loud" and these numbers are really ball park and probably incorrect, so please correct me where I'm wrong.

Let's say you want the equivalent energy output (heat of detonation) of 0.25 grams of mercury fulminate, i.e., about 0.45 KJ.

The number I have for a stoichiometric CuO/Al mixture (heat of combustion) is about 4.4 KJ/gram, so you need about 0.1 grams.

I don't know what the ignition temperature of CuO/Al is, but for the sake of argument, say it's 600 degrees C. (I suppose you could add a small amount of another oxidizer to adjust it downward if need be).

So the idea is to try to bring 0.1 g of CuO/Al up to it's ignition temperature, essentially all at once.

Assuming a heat capacity of about 0.6 J/(gram)(degree K)
(a composite number based on 20% Al and 80% CuO), it would take about 36 J of energy to reach 600 degrees C.

As I see it, the rate at which you add the energy isn't that critical, vis-a-vis a conventional EBW, since the thermite mixture is packed into, say, a narrow paper tube, so electrically, it's all in "series". If the mixture in the tube is uniform, and the negative temperature coefficient of resistance of the CuO is working in your favor, then maybe the bulk of the material in the tube will heat up approximately uniformly.

Maybe a photoflash cap switched with a mosfet would work in this case. Without some experimenting, I don't know what the electrical load would look like exactly, but 450uF at 400 V would give you 36 J.

Quote:

Originally posted by Boomer
"Can this thermite produce a shock to detonate a secondary?" Putting in additional electrical energy would be a waste, since due to the higher mass it would need even more power than exploding a bridge wire. The mix has to have the potential to det say PETN from its own reaction rate. Anyone tried that?

12AX7 - 11-1-2006 at 18:44

I wouldn't use a photoflash, and I wouldn't use a FET in series with any sort of discharge, ever. What you want, if anything, is a robust switch (ignitron for example, Google "quarter shrinker") and, if higher voltage is needed, a transformer.

Tim

jpsmith123 - 11-1-2006 at 19:38

Well, the idea is that hopefully you don't need exotic, expensive parts, otherwise you're back to an EBW configuration and much of the proposed advantage is lost. If you don't like Mosfets for some reason you can probably use a BJT, SCR, relay, etc.

Off hand, I have no idea what the electrical load would look like, and I'm basically just "thinking out loud" here, but something like the attached circuit may be a reasonable starting point.

Attachment: Circuit.pdf (7kB)
This file has been downloaded 637 times

Twospoons - 11-1-2006 at 22:30

A chunky thyristor might do it. I designed a 36J electric fence energiser some years back that used an 85 amp (RMS) thyristor to deliver a 2000A pulse, 100 microseconds long. The capacitor bank ran at 900V.
A triggered spark gap is another fast, high current switch thats fairly easy to build.

You need to be very careful with that sort of hardware - the caps can retain their charge for a long time. As I found out when I picked up a circuit board I'd turned off an hour earlier. The experience was .... unpleasant. It could have been fatal.

[Edited on 12-1-2006 by Twospoons]

FPMAGEL - 12-1-2006 at 00:04

Most people hear will proalbly know this but for the rest.
Exploding wire has a VOD of 4000m/s, for .5mm wire 40v 150amp will cause 20cm to vapourize, it the current not just the amps.

Theroy: the wire gets current flowing throught it, which creats a Mag feild, as the wire heats up to 1040C(temp of copper) it loses its abilty to conducted electricty and the Mag feild dies, releaseing the densi plasma.

12AX7 - 12-1-2006 at 05:39

SCRs would work, if you get fast units. The problem is getting microsecond edges, not something slow SCRs are apt to do. SCR inverters don't go much over say, 20kHz.

JP: your schematic is basically the right idea (ignoring the FET as mentioned), but for the transmission line to operate as described, the load has to be zero ohms. In practice, the wire has resistance, especially due to skin effect at this rate, it could be considerable. Thus, the transmission line ought to have a reasonable impedance mismatched to the load, if I'm understanding TL theory properly.

An IGBT may also work, certainly fast enough, but keeping it from exploding might be difficult.

An ignitron isn't an exotic piece of equipment, obviously you didn't Google "quarter shrinker".

It's two bolts with air inbetween and an ionizing electrode. Put voltage between the bolts near the breakdown voltage, spark the ignitor and it discharges with minimal impedance. Such devices are used to switch 10kJ or so in microseconds, so one would certainly be suitable.

Tim

12AX7 - 12-1-2006 at 05:46

Quote:

Originally posted by FPMAGEL
Most people hear will proalbly know this but for the rest.
Exploding wire has a VOD of 4000m/s, for .5mm wire 40v 150amp will cause 20cm to vapourize, it the current not just the amps.

I find this suspect. Do you have a reference?

Quote:

Theroy: the wire gets current flowing throught it, which creats a Mag feild, as the wire heats up to 1040C(temp of copper) it loses its abilty to conducted electricty and the Mag feild dies, releaseing the densi plasma.

Well, you'd be suprised to know I can pass large amounts of electricity through molten copper.

At this rate of change, the wire does not act as a fuse. As I recall, the function is such that, in the first microsecond, current only flows in the outer ten-thousandth of an inch of the conductor (skin effect). So much current is sent down the wire that the layer which is conducting is instantly joule heated to vapor. This transformation occurs in microseconds, so nothing has moved physically, and it tends to create a shockwave. I suppose you would estimate VOD based on energy, risetime and dimensions.

I don't know if these EBW use this effect, the wires may be too thin for layer burning and instead the whole wire conducts at that risetime.

Tim

Boomer - 12-1-2006 at 06:45

Right, it conducts completely. Skin depth for 1 MHz is 66µm, for 10 MHz 21µm. Wire dia is 38µ (commercial units).

I know that website (RISI), thats where I got the ballpark number of 200 Amps. They do state the delay time for 400-500 Amps though, and that is the recommended minimal current for reliable operation. I only need 1 kA min and 1.5 kA for reliability because my bridge is too thick.

The 4000 m/s are BS. Velocity of *detonation* for a gold plasma? The current keeps flowing even after the wire *evaporates* (not only melts). "it the current not just the amps...and the Mag feild dies, releaseing (sic) the densi (sic) plasma" says it all. FPMAGEL are you gonna delete your post *again* after we proved it makes no sense?

I get the idea of speeding up the thermite by bringing it up to reacton temp electrically. The question remains whether the subsequent reaction will detonate a secondary. Even the most sensitive need kilobars, most need tens of kilobars, in a sharp-rising pulse.

Twoospoons I remember that fence. Are you possibly on E+W under another name?

jpsmith123 - 12-1-2006 at 07:40

Quote:

SCRs would work, if you get fast units. The problem is getting microsecond edges, not something slow SCRs are apt to do. SCR inverters don't go much over say, 20kHz.

AFAIK, the relevant parameter would be turn-on di/dt. A fast inverter SCR might be capable of 1000 A/usec, which is probably plenty fast enough. As I recall, the limiting factor for inverter duty is the relatively slow turn off, which is obviously irrelevant here.

Quote:

JP: your schematic is basically the right idea (ignoring the FET as mentioned),

Ok, at the risk of going off topic, I'll bite; explain to me the fundamental problem with MOSFETs. (BTW, the MOSFET in this circuit is strictly representational at this point, primarily since I have no clue what the resistance of the CuO/Al mixture (when used as described) would be. Nor do I know exactly how fast the energy would need to be added.

Quote:

but for the transmission line to operate as described, the load has to be zero ohms.

Huh? If the load were zero ohms, basic physics dictates that I could not deposit energy into it. A transmission line in this case is merely a convenient way of transmitting power from a source to remote load.

Quote:

In practice, the wire has resistance, especially due to skin effect at this rate, it could be considerable. Thus, the transmission line ought to have a reasonable impedance mismatched to the load, if I'm understanding TL theory properly.

Apparently you have it backwards. Generally, the idea is to match the transmission line to the load reasonably well.

Quote:

An IGBT may also work, certainly fast enough, but keeping it from exploding might be difficult.

LOL!

Quote:

An ignitron isn't an exotic piece of equipment, obviously you didn't Google "quarter shrinker".

I don't need to "Google 'quarter shrinker'" to know what an ignitron is. Sorry but compared to a MOSFET or SCR I can get from stock at Allied or Digikey for probably less than fifty bucks, an ignitron is exotic and expensive.

Quote:

You've just described a triggered spark gap, not an ignitron.

12AX7 - 12-1-2006 at 09:10

Quote:

Originally posted by jpsmith123
AFAIK, the relevant parameter would be turn-on di/dt. A fast inverter SCR might be capable of 1000 A/usec, which is probably plenty fast enough. As I recall, the limiting factor for inverter duty is the relatively slow turn off, which is obviously irrelevant here.

Ah yes, that's right. I remember reading quarter shrinkers get slower results with SCRs than with a spark gap, so there's bound to be some cost of speed at any rate. 1kA/us is certainly fast enough here.

Quote:

JP: your schematic is basically the right idea (ignoring the FET as mentioned),

Right. Well the thing is, the MOSFET looks roughly like a resistor when on. A FET of the required voltage is going to be around half an ohm Rds(on), which means peak current can only possibly be, say, 400 / 0.5 = 1kA, if any part of the chip can even handle that (I wonder if the silicon even has that many charge carriers available or something?). The silicon chip and bonding wires will have I^2t ratings comparable to the EBW wire itself, making an ugly situation!

You can of course use a bunch in parallel, but then...you need a whole bunch...

IGBTs look like a few-volt diode drop when on (I have a quad of IRG4PC50UD that run up to 200A peak for all of 3Vce!). I don't know what kind of peak capacity you can get out of one, and I'm not about to test either ;-)

A good stocky, fast SCR is probably the way to go, especially since this is one-shot.

Quote:

but for the transmission line to operate as described, the load has to be zero ohms.

Well yeah, but I mean for developing current through the load- obviously it won't dissipate power, but looking at the I^2R loss, you want to maximize I.

A power match would be best, so you meet V^2/R and I^2R (that is, current and voltage are both proportioned to the capabilities of the driver), but if the idea is to have the transmission line bounce current around a bit, you at least aren't going to be able to match line to load.

Quote:

Apparently you have it backwards. Generally, the idea is to match the transmission line to the load reasonably well.

Yeah, that's kind of what I'm wondering. Not like it's a very well-behaved load, though...

Tim

Goldstein - 12-1-2006 at 15:48

LANL describes atleast two primaryless hot wire detonators based on both DDT under some confinement and by using a flyer plate propelled by deflagration.

There are a number of documents with experimental results on the LANL DVDs, under titles such as:

"The deflagration-to-detonation transition in PETN and HMX"
"All secondary explosive hot-wire devices"
"All-secondary explosive flying-plate detonators"
"Nonprimary-explosive, hot-wire detonator"

Also patented under US3,978,791 and US4,317,412.

Another patent worth looking at is US3,726,217 describing another clever principle for initiation of a PETN or HMX column.

And for those interested in reactive bridges. I'm sure there is more on the topic both in patent and academic databases.

jpsmith123 - 12-1-2006 at 16:41

Quote:

Well, there's going to be some kind of on-state loss with any switching device you choose. MOSFETs aren't unique in that regard. The essence of the design process is making various tradeoffs based on the applicable parameters and constraints. Obviously, for the sake of reasonable efficiency and reliability, you would generally need to make sure that R(load) >> RDS(on). Anyway, IIRC, you can get a 500V MOSFET with RDS(on) less than 0.1 ohms for under $20.00.

In the case at hand, once you have some idea of what the load looks like and how fast you need to put the energy in, you can nail down the design.

Quote:

Well yeah, but I mean for developing current through the load- obviously it won't dissipate power, but looking at the I^2R loss, you want to maximize I.
A power match would be best, so you meet V^2/R and I^2R (that is, current and voltage are both proportioned to the capabilities of the driver), but if the idea is to have the transmission line bounce current around a bit, you at least aren't going to be able to match line to load.

Other than to accomodate the current I need to handle to meet my ostensible design goals of maximum power into the load at minimal cost and inconvenience, I don't care about current...I certainly may not want to "maximize" it. Generally, I want to maximize the power and minimize the cost; ipso facto - I don't want it to "bounce around", either. Don't forget, this transmission line might be rather long and will have losses.

Twospoons - 12-1-2006 at 16:59

Quote:

Originally posted by Boomer

Twoospoons I remember that fence. Are you possibly on E+W under another name?

Yes, that was me. I changed names to dissociate myself from E&W - I don't go there anymore.

MOSFET conduction loss is I^2 R. Since I is very big, I^2 R will be huge.

IGBT conduction loss is I x Vsat. Trouble is if you try to poke too much current through one they come out of saturation and the power loss skyrockets - followed shortly by the IGBT!

Thyristors are nice because they handle massive peak currents - generally 10 to 20 times their rated contiuous current. The datasheet will usually tell you the I^2 T rating of the device - which is rather handy. The key to getting the buggers to turn on fast is delivering a really big gate current pulse - this can be done with a smaller thyristor.

Nothing is as robust as a spark gap, however.

[Edited on 13-1-2006 by Twospoons]

jpsmith123 - 12-1-2006 at 17:21

Thanks for that great link. I knew we were on the right track. Particularly interesting to me is the idea of the integrated shaped charge.

A similar idea I had was to put an EBW on "rails". IOW, in order to try to take advantage of energy that might otherwise be lost, especially with a slower energy input; i.e., making 0.5LI^2 energy more effective by electromagnetically driving the plasma.

I

FPMAGEL - 12-1-2006 at 17:24

"says it all. FPMAGEL are you gonna delete your post *again* after we proved it makes no sense?"
PROVED it makes no sense, well first off ,how about you prove it. And the replies from you two, a pinch of salt will have more sway over me.

"I find this suspect. Do you have a reference?"
Its from a book, a friend of mine in laser research told me.

"Well, you'd be suprised to know I can pass large amounts of electricity through molten copper."
Stop nit picking.You should try to get a containg Mag feild(liner field) around molten copper wire.

"instantly joule heated to vapor."
And the Mfeild dies, realseing that dense plasma.

[Edited on 13-1-2006 by FPMAGEL]

[Edited on 13-1-2006 by FPMAGEL]

jpsmith123 - 12-1-2006 at 17:47

In the course of looking to see what mixtures might work best, I came across some interesting parers:

http://www.me.umn.edu/~ap/aerosolgel.pdf

http://www.menet.umn.edu/~ap/KMnO4.pdf

Quote:

MOSFET conduction loss is I^2 R. Since I is very big, I^2 R will be huge.

Hopefully for the thermite idea as I've stated it, the load would be measured in ohms and the minimum power required would be no more than a few tens of KW, in which case a MOSFET would work fine, IMHO.

[Edited on 13-1-2006 by jpsmith123]

Twospoons - 12-1-2006 at 20:44

A MOSFET and a thyristor of similar rating will be about the same price, but I'd put money on the thyristor being more rugged. Its also much easier to get high voltage thyristors, should the need arise. You'll hunt for a long time for a 1200V MOSFET with the current rating you will need.

Just looked in the RS catalogue:
1200V 8A fet = $86,
1200V 55A thyristor = $17
And that thyristor will handle a 500A pulse, while the fet is only rated at 32A pulsed.

[Edited on 13-1-2006 by Twospoons]

jpsmith123 - 12-1-2006 at 22:24

Perhaps we should start another thread entitled: "MOSFET vs SCR for as yet completely undefined loads".

As I've endeavored to point out, ad nauseum, my use of the MOSFET in that circuit was representational (and reflected my optimism that you don't need a massive amount of power); in practice, determine what you need to do and use the appropriate device of your choice.

I responded to 12AX7 because I disagreed with his unqualified assertion that you never "ever" use a MOSFET for such purposes. That's really the extent of my interest in this "side issue" at this point.

(BTW, if you look in Digikey you can find a 500V, 50 amp (and probably at least 3 or 4 times that value for a non-repetitive pulse) MOSFET for $8.75. I would hope this would suffice, but as I said, I have no idea).

[Edited on 13-1-2006 by jpsmith123]

Another Interesting Patent

jpsmith123 - 13-1-2006 at 07:39

In US Patent 6227116, it is claimed that fuel rich thermite mixtures can cause a charge of PETN to detonate. I haven't read the whole patent yet, but here's the abstract and an example from the text:

Abstract
A detonator comprising a shell with a secondary explosive base charge, igniting means and an intermediate pyrotechnical train, said train comprising a novel ignition composition with a specific redox-pair of a metal fuel and a metal oxide oxidant, said fuel being present in excess to the amount of stoichiometrically being required to reduce the metal oxide, the ignition composition being able to ignite said secondary explosive into a convective deflagrating state to reliably detonate the same. Use of said novel ignition composition for the ignition of secondary explosives in general.

"Into 20 initiating elements in the form of aluminium tubes, each having a length of 20 mm and an internal diameter of 3 mm and an outside diameter of 6 mm, an ignition charge consisting of 20% by weight of Ti+80% by weight of Bi.sub.2 O.sub.3 was pressed to a column height of 5 mm. Adjacent thereto a column of PETN with a density of 1.3 g/cm.sup.3 was pressed.

In the same way 20 initiating elements were manufactured with the exception that the ignition charge (i.e. 20% of Ti+80% of Bi.sub.2 O.sub.3) also contained 8% by weight of Fe.sub.2 O.sub.3 as an additive.

This experiment showed that all 40 detonators containing said initiating elements worked excellently with a qualitative detonation of the base charge".

franklyn - 4-7-2006 at 22:23

Exploding bridgewires are a rather mature technology at this point. It is
afterall how the implosion assembly of A-bonbs has always been triggered.
The firing of automotive sparkplugs is less high tech and except for the smaller
energy required essentially the same.

Attached below are two pages excerpted from Explosive Trains
of the Army Material Command

Exploding Wires is the definitive reference on the technology ( just click submit at this page ) ->
http://catalog.loc.gov/cgi-bin/Pwebrecon.cgi?Search_Arg=expl...

Coaxial or instead low inductance stranded cable keep impedance low. See ->
http://www.litz-wire.com/applications.html

Low inductance capacitors are also required ->
http://www.sbelectronics.com/pdf/PPC2005%20Presentation.pdf

Of further interest see my post here about a conductive explosive _
http://www.sciencemadness.org/talk/viewthread.php?tid=851#pi...

At this thread ->
http://www.sciencemadness.org/talk/viewthread.php?tid=851#pi...

.

[Edited on 5-7-2006 by franklyn]

Attachment: Explosive Trains AMCP.pdf (888kB)
This file has been downloaded 691 times