Linux-powered iPad compatible Smart Rifle

[Starglider]

I recall seeing a prototype of this technology a few years back but AFAIK this is the first time it's been available off the shelf to consumers. Warms my heart to see the world getting a little bit more cyberpunk.

Steve has just delivered a .338 Lapua Magnum round directly onto a target about the size of a big dinner plate at a range of 1,008 yards—that's ten football fields, or a tick over 0.91 kilometers. It's his very first try. He has never fired a rifle before today.



Of course, Steve isn't some kind of super mutant marksman—he had a bit of help. We were plinking targets with $17,000+ Linux-powered hunting rifles, made by a small Austin company called TrackingPoint. Earlier this year, Ars reported on TrackingPoint's "Precision Guided Firearm" at CES, where the 59-employee company was giving the press a sneak-peak at their product before its official introduction at the Shot Show convention a week later.

The Precision Guided Firearm is a "whole widget" type of thing—it's not just a fancy scope on top of a fancy gun, but rather a tightly integrated system coupling a rifle, an ARM-powered scope running a modified version of Angström Linux (with some custom BitBake recipes and kernel modules to support the rifle's proprietary hardware), and a linked trigger mechanism whose weighting is controlled by the scope. TrackingPoint actually makes three different Precision Guided Firearms, two of which fire .300 Winchester Magnum rounds and one of which fires a larger .338 Lapua Magnum cartridge. The weapons themselves are crafted by Surgeon Rifles, and TrackingPoint adds the scope and trigger mechanism to the rifle and then sells it as a package. Since launch, TrackingPoint's sales have been very brisk: the company has nearly sold out of their entire allotment of PGFs for all of 2013.

I've fired a small number of bolt-action rifles before, but nothing as big as the .338 LM-chambered XS1, on which we spent most of our range time. The oddest thing about firing TrackingPoint's rifles, the thing wildly different from a standard hunting rifle, is the trigger mechanism. The rifle's "Tag-Track-Xact" technology means that there is a delay between when you pull the trigger and when the rifle fires—sometimes several seconds of delay, depending on how steady your aim is. This disconnect between trigger pull and firing can be weirdly disconcerting.



All of the interaction with the rifle's computer is done through the computerized tracking scope, which displays an image of the world in front of the rifle, overlaid with data. It's reminiscent of a fighter jet heads-up display. Through the tracking scope, you locate the thing you want to shoot, whether it's a metal plate like we were hitting on the range, a game animal, a tree, or whatever. You place the center reticle over the target and depress the red tagging button just forward of the trigger, which causes the tracking scope's range finder to briefly illuminate the target. The range finder is a powerful laser, rated at 75W and pulsed for an average power of 1mW, which measures the distance to the target in yards (1 yard is 0.9 meters, for metric readers) and displays that number in the upper left side of the scope's field of view. Once "tagged," the target gains a red pip. The scope image also reorients itself in a move that can be confusing the first few times it happens.

The reorientation is to take into account the "drop" of the shot. As demonstrated by MythBusters, a bullet is affected by gravity as it flies, losing height at the same rate as it would if dropped directly out of your hand. At 1000 yards, the bullet loses about 20 feet (6 meters) of height, and so when you tag a target with the tracking scope, the field of view immediately shifts down, by an amount proportional to the target's distance. This is so that the center of the scope accurately shows you where the round will go, rather than some point far above it. Someone firing a long-range shot with a traditional scope must make the same adjustment manually, but here the ballistic computer takes care of it.

Once the target is tagged, the laser ceases illuminating and the embedded system's image recognition routines take over. If the target is a game animal, the scope will track the animal's movement in the field of view, and the red pip will remain stuck to the point on the target where it was originally tagged. This level of precision obviously helps out at the range, but it also enables a hunter to more ethically harvest animals by precisely targeting specific areas—the tagging mechanism greatly simplifies the process of killing the game animal with a single shot, rather than wounding it and inflicting unnecessary suffering. After the target has been tagged, the scope's reticle changes to a large blue "X," and the weapon can be fired. To actually send a round downrange to the target, you depress the weapon's trigger. This doesn't cause the weapon to immediately fire, though—the reticle turns red, and while keeping the trigger held down, you must align the reticle with the tagged pip. Once the pip and the reticle coincide, the weapon fires.

Just about everyone who fires a pistol or rifle goes through an adjustment period where they anticipate and overcompensate for the recoil the weapon produces when fired. The overcompensation is an almost unconscious adaptation that the body makes. Overcoming it is part of learning to shoot. Firearms are loud, and when they are fired they jump back in your grip. So, less-experienced shooters cope with the recoil by bringing the weapon's barrel downward against the anticipated force—often before they've even fired the weapon. The only way to learn to cope effectively with a weapon's recoil is to practice, and the adjustment period lasts a different length of time for every person. There are things you can do to help (I've had instructors in the past recommended balancing a penny on your weapon's front sight and dry-firing until the penny doesn't move as a way to teach a steady trigger pull), but one way or another it's something everyone has to deal with. TrackingPoint's rifles, though, remove overcompensation from the equation by totally segregating the act of pulling the trigger from the weapon's firing. Also no longer necessary is holding your breath to avoid disturbing your sight picture or timing your shot to your heartbeat or other long-range tricks—the rifle's computer picks the most optimal time to fire. You pull and hold your trigger, and then carefully align the reticle with the floating red pip.

A lot of the delay is because what the user sees in the sight picture isn't always exactly what the rifle's computer is looking at. The rifle gets its view of the world through the computerized tracking scope, which looks out through a 35x fixed-magnification optic focused onto a digital imaging sensor with a vertical resolution of about 3600 pixels. Behind the optic assembly is an entire ARM-powered embedded system that knows everything there is to know about the characteristics of the rifle and its ammunition. The tracking scope also contains instrumentation to make it aware of ambient conditions, including temperature and humidity and atmospheric pressure, the incline and cant of the weapon, and its compass heading. The only thing you need to manually enter is the wind speed and direction, using a control on the top of the scope.

...

The elephant in the room, though, is that these rifles could have applications beyond sport hunting—what could be used to lock onto and take down a running deer at 1000 yards could conceivably be used to lock onto and shoot a person, too. Venturebeat editor-in-chief Matt Marshall went so far as to call TrackingPoint's products "disgusting" and "truly repulsive," saying that the rifles "are paramilitary weapons." He invoked the image of a madman like John Allen Muhammad using one to pick off innocent people from "a mile away."

That's the subtext behind a lot of the negative reaction to TrackingPoint's PGF technology: its potential use against people, either in the hands of a soldier, a policeman, or the proverbial crazy person. I'd argue that it's far more likely a mass shooting event would take place with conventional firearms than with a PGF; a bolt-action hunting rifle isn't exactly the quickest method of getting a lot of lead on a lot of different targets. The very nature of the PGF's "Tag-Track-Xact" scope encourages methodical target selection at range, and hauling even a hypothetical smaller PGF into a crowded place and letting loose would be enormously difficult. It's not a close-quarters weapon by any stretch of the imagination.

Fear-mongering aside, two things are quite obvious about the three PGF rifles that TrackingPoint is launching with. First, they're not particularly good "paramilitary weapons," as paramilitary weapons go. Second, though, is that they're only the first round of PGFs. The weapons have a heavy and relatively fragile computer on top of them, filled with electronics and glass. They're not rugged enough to survive anything like a battlefield, urban or otherwise. You don't want to drop them out of a tree or slam them into a wall or get them wet.

There's also a question of cost versus benefit. In repeated firing on the range, we didn't even bother with any targets closer than 250 yards—that was the "warm-up" distance. It doesn't particularly make sense for a government or a law enforcement agency to spend more than $17,000 on a PGF when they simply don't need its capabilities. "The average [law enforcement] sniper engagement is 75 yards, like from the house across the street....You look at cops and they all have short-barreled Remington 700s that weigh eight pounds."

That's not to say that TrackingPoint hasn't been demoing their products to law enforcement and some government agencies. "We've shot with the Austin SWAT guys," Boyd noted. "They all think it's great. We've flown out to DC and I've done demos with the Secret Service, their counter-sniper guys, and the FBI, their Hostage Rescue guys...and those are organizations we expect to be customers at some point. The thing that they like the most about it is not really the marksmanship and ballistics—they're already very good at that—but it's more the recording and the networking," elaborated Boyd. The PGF's integrated recording capabilities would function like the dashboard cameras in a police cruiser, documenting exactly what the shooter sees, exactly where his or her round was aimed, and the exact conditions under which a round was fired.



Hidden inside the tracking scope's guts is a Wi-Fi antenna, and the weapon can form an ad-hoc Wi-Fi network and link up with a specially crafted iOS app running on an iPad. Everything that the weapon's user sees in the optic is shown on the iPad's screen—all the ambient condition data, the reticle, the tagged point, everything. It's a powerful tool to assist a new hunter with how and where to line up a shot. On the target range, our spotter flipped back and forth between the iPad and a traditional spotting scope as we fired, using the iPad to confirm where our tags were placed on the targets and using the spotting scope to check for hits on the targets too far away to hear the impact. The scope does more than just stream video—it records video, too, along with audio, and the two videos embedded in this article were taken from our time on the range. The rifle's computer also takes stills when a target is illuminated with the rangefinder laser and when the rifle is fired. The iPad app even gives users the ability to upload videos of their shots to Facebook or YouTube.

The idea of duplicating the sight picture on a separate screen is only the start. The streaming video lets a spotter see exactly what the rifle's user sees, but it's passive—the iPad displays a live picture of the PGF's scope, but nothing more. TrackingPoint is kicking around the idea of eventually releasing an actual spotting scope equipped with the "Tag-Track-Xact" technology. It would be capable of linking up via Wi-Fi to a PGF and could be used to call shots, letting a spotter actually designate targets that would show up on the PGF's display. Something that's not particularly likely, at least for now, is integration with an augmented reality system like Google Glass. The idea seems like it could be a good fit—getting a sight picture beamed into your eye without having to peer into a scope—but TrackingPoint doesn't believe that Google would be interested in integrating with them and they haven't pursued any formal relationship. "Google actually has some reasonably prohibitive policies around firearms," explained Boyd. "I don't really think they're going to be a good partner for us."

...

"In 10 years, precision guided firearms is just going to be another class of weapons that a ton of people are going to make," he predicted as we were packing up. We'd spent several hours shooting in the cool, dry Austin air, racking up hit after hit on targets so far away that they couldn't be seen unaided, watching each other's scopes through iPads. It all felt a little unreal.

Anyone with more firearms experience care to comment on the near-future direction and adoption rate of this technology? Aliens-style smartguns seem to be just around the corner.

[Eternal_Freedom]

I can't comment on the possibilities, but that is a very pretty gun and a very cool idea. Now if only I had $17K spare and lived in America I would get me one of these.

[The Xeelee]

They can take my money!

I don't even like guns.

[Stark]

Oh man. Bullets drop in flight 'as demonstrated by Mythbusters'. But ballistic apps for phones are pretty old hat and this doesn't automatically adjust for wind so... Yay linux? :v

[Ted C]

Turns an average Joe into a sniper? It could certainly be abused, although, as they say, it's not the tool of choice for a suicide shooting spree.

[Starglider]

and this doesn't automatically adjust for wind so...

Interestingly LIDAR-based wind sensors now exist that can scan a volumetric flow field out to several hundred meters; they're used in high-end sailing competitions and for surverying wind power sites. The laser class is similar so in principle it should be possible to build a sight that will near-instantaneously sample and adjust for wind along the entire trajectory. That would be a significant improvement on any current adjustment method (based on a single speed/direction measurement); cost prohibitive at present of course but the necessary electronics will eventually be dirt cheap.

[Sea Skimmer]

The company claims really high effectiveness out to 700m, and then it drops off abruptly to almost nothing, which is nice, but they are using cartridges designed for 1000-1,500m ranges and assuming little or no wind error. Not a big deal against animals that need high power shots to be brought down and you can put out flags for wind, but not so impressive in a military context. Now meanwhile, for shots inside 400m a good marksman is not going to miss either with such high power weapons. So your looking at that 400-700m band in which this thing is a dramatic improvement; but for 10,000 extra dollars you can buy an awful big pile of training ammunition all the same. The company has also been touting its bolt action accuracy compared to in some cases semi automatic weapons, which is not a very honest approach all the more so given that the semi auto weapons also assume a 5mph wind error in US Army data. So neat toy, a lot of future ahead for the concept, but not so sure its a a revolution yet. Also not ruggedized. It did however work from a hovering helicopter in tests which is pretty impressive in its own right.

Overall I suspect DARPA guided bullet will probably be adapted first in view of its long range, or someone will finally make a satisfactory air burst shoulder arm that makes direct hits irrelevant. All of these systems suffer from the problem of the laser making your position blatant to a high tech enemy. So far I've not heard of anyone actually buying one, or being able to borrow one to test completely independently.

[madd0ct0r]

So if laser's are like setting up a disco ball above the sniper, I guess they'd be more useful for guy's who position is already compromised - ie squaddies moving up.
I guess they might work as a one per squad 'add-on' to a normal assault rifle, for those occasions when you really need to make that shot but spend most of the time waving the gun around in covering fire like normal.

[cosmicalstorm]

Didn't Bruce Willis do this in the Jackal? I suspect that within a not so distant future it will be possible to construct cheap drones with cheap weapons and automated aiming-systems attached. Maybe in significant numbers? Imagine a terrorist-attack involving hundreds of cheap drones with grenades and handguns swarming a building or a town. I wonder what the counter would be.

[Arthur_Tuxedo]

Overall I suspect DARPA guided bullet will probably be adapted first in view of its long range, or someone will finally make a satisfactory air burst shoulder arm that makes direct hits irrelevant. All of these systems suffer from the problem of the laser making your position blatant to a high tech enemy. So far I've not heard of anyone actually buying one, or being able to borrow one to test completely independently.

I thought the 25mm OCSW was supposed to be the satisfactory air burst shoulder arm you describe. What went wrong with it?

[born in shadow]

I thought the 25mm OCSW was supposed to be the satisfactory air burst shoulder arm you describe. What went wrong with it?

If I'm not mistaken, it had a tendency to have the shells explode about a foot out of the barrel. It had to do with the range finder not working correctly, but I could be mis-remembering it all.

[Jub]

I suspect that within a not so distant future it will be possible to construct cheap drones with cheap weapons and automated aiming-systems attached. Maybe in significant numbers? Imagine a terrorist-attack involving hundreds of cheap drones with grenades and handguns swarming a building or a town. I wonder what the counter would be.

Couldn't they already do that if they wanted to? Get an 18 wheeler to house the operators in and then buy large model aircraft and fit them with simple cameras and weapon systems. You don't need much sophistication if you don't need exceptional radio range or accuracy.

[Adam Reynolds]

I thought the 25mm OCSW was supposed to be the satisfactory air burst shoulder arm you describe. What went wrong with it?

If I'm not mistaken, it had a tendency to have the shells explode about a foot out of the barrel. It had to do with the range finder not working correctly, but I could be mis-remembering it all.

If you are referring to the XM25 CDTE that has seen test use in Afganistan, it was recently removed from service after an incident on Feb 2, 2013 in which a weapon exploded during live fire training. This explosion was a result of a double feed, fortunately only the primer and propellant detonated which prevented serious injuries to the user. It appears to be the reason for the removal was the fact that this had also happened previously. The company does claim that they have redesigned the feed system to prevent this in the future.
see here for more

[born in shadow]

If you are referring to the XM25 CDTE that has seen test use in Afganistan, it was recently removed from service after an incident on Feb 2, 2013 in which a weapon exploded during live fire training. This explosion was a result of a double feed, fortunately only the primer and propellant detonated which prevented serious injuries to the user. It appears to be the reason for the removal was the fact that this had also happened previously. The company does claim that they have redesigned the feed system to prevent this in the future.
see here for more

Hm, you may be right. I can't find anything on what I was thinking of. I could've sworn I remembered hearing about this more recently than February though.

[Sea Skimmer]

Didn't Bruce Willis do this in the Jackal? I suspect that within a not so distant future it will be possible to construct cheap drones with cheap weapons and automated aiming-systems attached. Maybe in significant numbers? Imagine a terrorist-attack involving hundreds of cheap drones with grenades and handguns swarming a building or a town. I wonder what the counter would be.

No such thing as a 'cheap' automatic aiming system, not on a terrorist budget. Best your going to do is have a time of flight based release which will be enough to hit a town, but not a specific building. Anything remote controlled meanwhile can be jammed very easily without using very expensive communications equipment. Also people are going to notice you making hundreds of armed RC planes and then trying to figure out a way to launch them all. I'd worry a lot more about other things the terrorist might do with 200 hand grenades. Ideas like this are actually much more useful to a nation state that can simply order 100,000 small armed drones from a factory and bring them to action by the truckload and find a way to easily launch and control them all.

The counter will be and already is lasers, it'd also be really easy to start making .50cal class anti aircraft weapons again. One twin .50cal mount and an optical tracking system will saw through many dozens of drones very quickly.

Couldn't they already do that if they wanted to? Get an 18 wheeler to house the operators in and then buy large model aircraft and fit them with simple cameras and weapon systems. You don't need much sophistication if you don't need exceptional radio range or accuracy.

The US threw someone in jail not long ago for plotting such an attack; but the amount of damage you are going to do with a large RC plane is still very limited, on par with a package bomb or sniper attack on a building. Depending on what explosives you have you might not even breach a good wall. Main advantage is it can fly over fences, the person in question wanted to hit the Pentagon IIRC. I'd be a million times more concerned about someone filling that 18 wheeler with 100,000 pounds of explosives and incendiary material.

[Sea Skimmer]

I thought the 25mm OCSW was supposed to be the satisfactory air burst shoulder arm you describe. What went wrong with it?

Na, ignoring the recent teething problem, its still not want everything people really want. Still too heavy among other thing. It will be well used, but I doubt its going to be around for very long before a complete replacement is designed.

[Starglider]

One twin .50cal mount and an optical tracking system will saw through many dozens of drones very quickly.

Fine in a warzone or open country but the collateral damage of all those falling rounds over say Washington D.C. must be significant.
Even a low-power (kilowatt) laser would destroy civillian-R/C-aircraft type drones quite easily though, if it could keep the spot steady on the airframe.

I'd be a million times more concerned about someone filling that 18 wheeler with 100,000 pounds of explosives and incendiary material.

Ditto but these things will continue to get cheaper and smarter, even without looking at lethal applications the nuissance value for protesters must be significant. Certainly the Occupy crowd would have been more annoying if they had a few hundred R/C quadcopters to buzz around harassing people and dropping smoke bombs into air conditioning inlets.

[Zaune]

Wasn't that guy's plan to use the drone to make a small bang and maybe start a fire, then be waiting near the designated assembly point with an automatic weapon?

[Sea Skimmer]

Even a low-power (kilowatt) laser would destroy civillian-R/C-aircraft type drones quite easily though, if it could keep the spot steady on the airframe.

Target drones have already been downed by such lasers years ago. Higher power lasers took out Harpoon missiles in US navy tests in the 1980s. Pointing isn't a big deal.

Ditto but these things will continue to get cheaper and smarter, even without looking at lethal applications the nuissance value for protesters must be significant. Certainly the Occupy crowd would have been more annoying if they had a few hundred R/C quadcopters to buzz around harassing people and dropping smoke bombs into air conditioning inlets.

Annoying? Sure. They'd also all be arrested on terrorism charges if they did such a thing, most likely ending the occupy movement months earlier then was the case and thus making them far less annoying in total, and provoking far more overt hostility against them. Also since most RC toys are now 2.4 GHz all the equipment police and governments have to intercept and jam wireless devices would already work to shut them down in an area of one city easily. Military aircraft it would more like we can jam this at ~200nm or even much further as the equipment is operating on fractional watt power levels vs jammers intended to barrage megawatt class radars.

[Starglider]

Certainly the Occupy crowd would have been more annoying if they had a few hundred R/C quadcopters to buzz around harassing people and dropping smoke bombs into air conditioning inlets.

Annoying? Sure. They'd also all be arrested on terrorism charges if they did such a thing, most likely ending the occupy movement months earlier then was the case and thus making them far less annoying in total, and provoking far more overt hostility against them.

How would that work, legally speaking, if they use lightweight unarmed equipment that can't really hurt anyone? Perhaps anything that would cause evacuation of buildings could be considered a bomb threat and prosecuted under relevant legislation, but harassment such as hovering next to offices and recording meetings or playing music? Also you can trivially jam them, but tracing who is actually controlling one would be difficult, either for the directly controlled (saturated time-sliced airspace, inprecise direction finding particularly in built-up areas) or cellphone network based (same difficultly as tracing any Internet hacker) ones. Long-term area jamming of digital networking frequencies would cause more disruption (in a city center) than the UAVs themselves.

[Terralthra]

Uh, yeah. Sea Skimmer, your "jamming" plan effectively shuts down 802.11b, 802.11g, and 802.11n (one band out of two). That's something like 99% of all WLAN connections in the US.

[cosmicalstorm]

Jamming would not be an issue for systems capable of autonomous action. I am under the impression that the necessary technology to create a camera that recognizes human shapes already exists. It seems it would be possible to build a "Terminator" drone that would be based on the very cheapest materials. It would be a flying machine with a cheap weapon and a few rounds of ammo. It would be released outside a city and proceed to fly a certain distance, once activated, a killing program would execute and at that point it would locate the nearest six humans and shoot each one of them.
Now this is not a civilization ending issue like biotech, machine intelligence and such. But it seems like a real possibility. I do not doubt that it would be easy to destroy. Maybe government and military will have to build their own immune-defense-ish systems to deal with problems like these as they occur.

[Sea Skimmer]

How would that work, legally speaking, if they use lightweight unarmed equipment that can't really hurt anyone? Perhaps anything that would cause evacuation of buildings could be considered a bomb threat and prosecuted under relevant legislation, but harassment such as hovering next to offices and recording meetings or playing music?

Using an incendiary device to pump toxic fumes aka smoke into a building is going to be treated a lot more seriously then just making a bomb threat. That's easily a 25 year jail sentence before we start thinking up all the other possible charges that could be piled on or the likelyhood of states rapidly passing laws to require license for UAVs even if the fed gov wont act. Hovering next to offices or playing music isn't that annoying and still faces legal issues. How big a speaker do you think a RC helicopter is really going to carry?

Also you can trivially jam them, but tracing who is actually controlling one would be difficult, either for the directly controlled (saturated time-sliced airspace, inprecise direction finding particularly in built-up areas) or cellphone network based (same difficultly as tracing any Internet hacker) ones.

You know just with software mods we made the cell networks in the US capable of locating phones with under 10m precision using time of arrival, and most peoples phones are now uploading GPS data constantly anyway which is being stored. Locating people would be a lot easier then you think if anyone felt like it or was using dedicated equipment.

Long-term area jamming of digital networking frequencies would cause more disruption (in a city center) than the UAVs themselves.

Who said long term? Turn it on for 1 minute at a time and the UAV crashes. Most people wouldn't even notice, and indoor wireless devices might not be affected at all.

[Sea Skimmer]

Jamming would not be an issue for systems capable of autonomous action. I am under the impression that the necessary technology to create a camera that recognizes human shapes already exists. It seems it would be possible to build a "Terminator" drone that would be based on the very cheapest materials. It would be a flying machine with a cheap weapon and a few rounds of ammo. It would be released outside a city and proceed to fly a certain distance, once activated, a killing program would execute and at that point it would locate the nearest six humans and shoot each one of them.
Now this is not a civilization ending issue like biotech, machine intelligence and such. But it seems like a real possibility. I do not doubt that it would be easy to destroy. Maybe government and military will have to build their own immune-defense-ish systems to deal with problems like these as they occur.

You might be able to build such a thing now, but its not going to work very well nor be cheaply created by a jolly band of terrorists, let alone in any serious numbers. Making autonomous weapons is incredibly difficult even against large fixed targets. Identifying one person, not so hard. Identifying one person in a crowd while the camera is moving an shaking like crazy because its mounted on a tiny UAV, much harder.

[Sky Captain]

Jamming would not be an issue for systems capable of autonomous action. I am under the impression that the necessary technology to create a camera that recognizes human shapes already exists. It seems it would be possible to build a "Terminator" drone that would be based on the very cheapest materials. It would be a flying machine with a cheap weapon and a few rounds of ammo. It would be released outside a city and proceed to fly a certain distance, once activated, a killing program would execute and at that point it would locate the nearest six humans and shoot each one of them.
Now this is not a civilization ending issue like biotech, machine intelligence and such. But it seems like a real possibility. I do not doubt that it would be easy to destroy. Maybe government and military will have to build their own immune-defense-ish systems to deal with problems like these as they occur.

It would be too much trouble for too little gain. Want to kill some people - build a homemade bomb and leave it in a crowded buss. Much easier and more reliable than trying to make armed drone from RC plane parts.
Small bomb that can be carried by RC plane only could be useful to attack something volatile that doesn't take much to set off. Maybe such attack could be effective against oil refinery after all it don't take much to cause major fire in industrial facility full of flammable liquids.