The tests were very successful. I was quite pleased with the results even if they were less than I expected.  The new rig stayed together and made 6 successful shots before we ran out of the custom machined projectiles.  There was a failed shot somewhere in the middle and I was able to safely and successfully discharge the cap bank, and quite quickly I might add.  The super diode was fully successful in restricting the caps from reverse polarity and no caps were damaged during the tests.

Results were less than I expected because the targets weren't penetrated until point blank range. This is mostly because the projectiles twisted during flight to hit the targets in a broadside fashion.  Its not that they weren't firing with great speed, one ripped the phone book cover and several pages underneath over its entire side length (approx 1.25"). It was also difficult to determine how well the projectile was making contact, and how much it was disintegrating due to the current cause they exploded on impact. one was recovered and it doesn't paint a nice picture.  Here it is next to an unfired projectile.

It seems that the current flowing through the projectile is still too much. Again the projectile is being damaged (surface pitting and material exchange with the rail) by each shot.  The projectile is also very hot after firing. A small chunk (.1" square) that had been cooling for ~10 seconds still had enough heat to burn me when I rested my hand on it accidentally.

I think a modification is needed before testing can begin again. The rail guides will be re-machined to accommodate a wider projectile so that there is more surface area touching the rail, and a wider path for current to flow. This change should help with heating and pitting.  A larger projectile might also stand a better chance of staying together after impacting the target.  This should be a fairly simple change to make before the next tests.

Wednesday May 31, 2006

Construction on the new rail has been underway for quite a while now.  the new system is quite a bit more complex than the last.  Today I'm going to outline and describe the major components and how they have changed form the last version. 

The charging, monitoring and firing systems have changed considerably.  The last version used several independent pieces to do these jobs.  The original charging circuit was powerful but dangerous to say the least, and there was no way to discharge the cap bank once it was charged. There was a double blade switch mounted to the top of it that allowed disconnecting from the caps during firing, but the fully dangerous energies of the caps were still exposed and out in the open begging for trouble. The monitoring system was a couple of volt meters sitting near the rail, one measuring across a resistor to determine the current flow to or from the cap bank and one measured the voltage across the caps.  It worked well enough but I was always worried they would be damaged being directly connected to the rail system during firing.  The firing system involved an air tank with a nozzle next to the charger. The BB was placed in a long nylon tube and air pressure was used to send the BB down to the rail. This long tube put a lot of spin on the BB contributing to previous issues.

Overall, the new system is more complex and much safer. The same voltage multiplier charger was used but the blade switch was exchanged for a relay controlled by the firing circuit.  The charger is disconnected from the caps and AC source when not in use and a discharge resistor is connected to the output (removes any stored energy for safety).  When the charge switch is flipped, AC is applied to the input, the discharge resistor is removed and the cap bank is connected to the output.  Once the caps are charged the switch is turned off and the charger goes back to a disconnected and discharged state.

The new system also employs a discharger for the cap bank, in case the shot has to be scrubbed. It is usually disconnected from everything and powered off. When discharge is selected the caps are connected to the energy sink, a fan is turned on and a soft start circuit enables a variable load.  The load is designed to dissipate 100W of power continuously until the caps are discharged.  The circuit uses several methods of feedback to maintain a steady power consumption.  All the power dissipated is turned to heat, placed on a heat sink and blown away by the fan.

Here are some pictures of the PC board I created for the discharger circuit.

A projectile is still fired into the rail by pressurized air but now the air tank is part of rail gun assembly and the release is an electrically controlled solenoid no manual.  The control box has several pneumatic valves for charging and discharging the air tank and a gauge for measuring the pressure in the tank.  A run of the mill compressor is connected to the control box as a pressurized air supply. A button is pressed opening a valve and the tank fills. Another button opens a valve to a muffler that releases the air in the tank in a controlled fashion.

The firing circuit has two stages. The same switch that turns on the charger is flipped the other way and the fire control is armed. When armed an red LED is illuminated and the trigger is energized. The fire button does two things; it turns on a relay that disconnects the control box from the cap bank, and it energizes the air release solenoid at the rail.

The cap bank monitoring is done in the same fashion, with a couple volt meters, but they are connected to ports on the control box not the caps themselves. They are physically away from the rail, disconnected when fire is pressed, and reconnected when the fire button is released.

Mounted right to the rail assembly is the air tank used by the injector.  To get the best performance it was designed to be as strait as possible, minimizing turbulence caused by corners.  The charge bottle is made of 4" PVC with a Y at one end. after the Y, the parts are metal.  There is a T that connects the bottle, a connection for charging and discharging the tank, and the release solenoid. The tank and solenoid are aligned for best flow during shots. The last link to the rail is 3/8" nylon tubing, to add a little flexibility to the union.

Complexity, is what the new rail gun assembly has become. It is a cluster of custom machined plastic bits, rails, wires, screws and magnets.  To complicate things, the puzzle has to be assembled and threaded into a piece of Unistrut that is a snug fit and about 3 ft long.  Luckily the Unistrut is segmented allowing separate access to the rail section, but full disassembly is required to adjust the injector.  The segment around the rails has many screws, from opposite sides, for the fine adjustment of the axial alignment. It is difficult to describe the design of the rail section so I'll just let the pictures do most of the talking. 

The injector is made of four pieces of acrylic held together with screws. The projectile guide in the injector is shimmed to +/- .001" (very tedious work, but great for the attaining the most speed).  The injector is supported by the Unistrut and the power cable running to the rail (the rubber of the cable insulator acts at a spring centering the injector). By running the power cables inside the Unistrut, and on axis with the rail, they will contribute to the force applied to the projectile. The rails are separated by Teflon for its low friction. The Teflon guides are aligned by nylon with lots of holes for mounting the magnets. Those pieces are further supported by acrylic and the Unistrut it self.  The rails are pre-loaded together by the Teflon shims. The rails get .008" closer together by the end of the rails (the EM fields are forcing the rails apart, preloading attempts to counteract this).

The capacitor bank has also been improved. the old design had two groups of 12 caps each. The groups were housed in a giant Tupperware tub, as seen in the pictures and movies.  Only four of the caps were ever used by the last rail design.  The new rail design should handle much larger energies, so I might as well redesign the cap bank at the same time.  The new cap bank will have 36 caps in strings of six that can be configured in multiple ways (series and parallel). The power supply has been measured outputting above 1500V, perfect for three groups in series. The cap bank has been housed in a Plexiglas box for safety purposes.  The power cables will be connected to the caps inside the box to remove exposed electrical conductors.  The box will also contain any splatter if a cap decides to pop.

The new design will also incorporate a HUGE diode to stop the coil from drawing the caps negative.  When an inductor is operating it resists changes in current because it develops a magnetic field around it. Current continues to flow as long as this magnetic field is present. In this case, once the caps are discharged the coil continues to carry current forcing the caps to go negative. Electrolytic caps are easily damaged by reversing their polarity. By shunting the caps with a massive diode all the current that would have tried to damage them, harmlessly flows through the diode until the coils magnetic field collapses.  I will be building the diode before testing begins.

Sunday, May 14, 2006

Forgive me rail gun blog for I have sinned.... Its been nearly a year since my last post.

I'm back! the blog may have been neglected but the work continues. Since the last post the rail gun project has begun a rather large revision.  A brand new rail design is under construction along with a new control/firing box.

Why the big revision with so little use of the last design?  BTW it has been fired several times without mention here and the results were less than consistent.  So, why redesign?  It is getting a makeover for a lot of reasons.  The projectile (on many levels), the rail design, and other factors. BB were a cheap and easy way to begin these experiments, but they are a horrible projectile to be using.

First reason the BB's suck; the BB's are made of chrome plated steel, which means they are affected by magnetic fields. when the rail gun fires, it creates an electro-magnetic field to launch the projectile. when that field travels through the BB it polarizes the steel and, most likely, magnetically saturates the material. When saturated it takes a long time for the fields to change orientation. If the BB is spinning in any direction the Electro-Magnetic (EM) field will be constantly changing and not aligned properly to make the best use of the field created by the rails, lowering efficiency on an already inefficient design. 

By design, the BB's are spinning when they enter the rail.  A long piece of nylon tubing was used to inject the BB into the rail gun.  No part of the tube was strait because it is coiled for storage, so the BB is constantly rubbing a side causing it to roll down the length and into the rails.

Another reason not to use BB's is, the cheap steel they are made of has a low melting point. low enough for the current applied during firing to melt the BB. Any energy used to destroy the projectile is being wasted. The steel in the BB has a resistance to the flow of electricity which, all be it small for any other purpose, is too large for this project. The electrical power consumed by any device is described by P (power) = V * I = R * I^2 = (V^2) / R.  In this case the current (I) is VERY large. on the order of .5-3kA (guessing?), so if the resistance of the BB is anything the power being dissipated in the BB is huge for its physical size. The pictures of mangled BB's is proof of this.

The flaws in the rail are due to the design its self.  the rails are aligned so that the large faces of the rails are on the same plane, and the small faces towards each other. Between the rails, is where the EM field is the strongest. By using the small sides, this rail design is susceptible to large variations in the EM field due to eddy currents and variations in the material. By placing the rails in such a way that the broad faces are towards each other the field lines become more linear between the rails.  With  linear EM-fields the force on the projectile should be more consistent. If the fields vary or differ between the rails the projectile will want to twist and deflect into the rails or guides. The new design should help with this.

The old design had bolts running through the rails, clamping it together. This is bad for several reasons. the bolts are energized when the rail is.  Not that anyone really hangs around close to the rail gun while the caps are charged, but never the less, its not the safest having lots of exposed "hot" bolts.

The bigger fault in having bolts running through the rails is the non-linear path of current flow.  Electrons take the path of least resistance as a rule.  An extension of that is; electrons want to travel alone. Explanation: looking at a cross-section of a conductor flowing current you will see a, roughly, even distribution of current through the entire area (they don't all take the same path).  By placing holes through the conductor, the electrons split and bottleneck around the holes and spread back out afterward. This causes huge variations in the EM fields.

EM fields wrap around wire carrying current. When the "single wire" of the rail is split by a bolt hole the current splits, creating two fields that don't add up to the original (where it counts, between the rails).  Some of the flux escapes through the hole instead of flowing all the way around between the rails.  So, where there is a hole, the field strength between the rails decreases, still using energy but not adding as much force to the projectile.  The bolts holding the rail gun together are made of steel which, being ferrous, distorts the fields even more. 

All of these factor add up to my reasoning for the redesign. 

The new rail design is very different from the last. The rails are solid (even current flow), with broad faces towards each other (linear flux). The projectile will be rectangular (no changing orientation mid shot) and made of carbon (higher resistance for a longer energy pulse, and much higher melting point).  The guide rails will be made of Teflon (slicker than nylon) and the injector will be permanently mounted to the rail (accuracy of axial alignment).  Magnets will line the sides of the rail amplifying the fields created by current flow (permanent magnets fields amplify electromagnetic fields).  And the entire rail assembly (rails and injector) will be enclosed in an aluminum extrusion for rigidity (axial alignment of injector), strength (repulsive effects of firing) and shrapnel containment (possible flying magnets).

Tuesday, June 7, 2005

We shot the rail gun today and got great results.  New movies are posted here.  The new setup works great.  kind of heavy when its full of everything, but so simple to move around compared to before.  It is ready for its public demonstration

We only used 4 of the 24 caps mounted in the box.  From previous experience we know that BB's begin to break down when we fire using energies above 1kJ.  So using all the caps would be futile. Below is a picture of a BB after a 1kJ shot.  It shows signs of severe damage.

Monday, June 6, 2005

Today I got all crazy with my buddy Drew.  We went to Home Depot and got supplies to put the party gun together.  We will be building this fixture around the new capacitors.  This will be our first use of them.

We got 2/0 gauge wire used for industrial electrical purposes, figuring a low stranded high gauge wire would help choking the initial current surge.  The first current that flows is on the surface of the wire (the steep rise in current produces skin effects) when the projectile completes the circuit. With only 7 strands it should have much less surface area than a high stranded wire. effectively acting as a choke.

The core of the choke was chosen to be a piece of 3" PVC pipe instead of the design proposed in the last entry.  Construction using the pipe core should be much easier, and the pipe will be more sturdy.  We started with a hole at one end that is off center, to limit the tight radius corner. zip ties were used in a staggered pattern between each winding, with a mount to the tube after 10 windings.  With the amount of material available a coil of 25 turns was made.

The coil is surrounded by a piece of aluminum vent pipe and mounted to the side of a large plastic tub that holds the capacitor banks.  the shielding is mostly for the capacitors.  In a previous version the coil was right next to the caps and one that was closest died during a high powered shot.  We don't want to repeat the failure so a new mounting location and the addition of earth grounded (only during test shots) shield should help.  6 gauge wire, that was the other coil, is used as the ground strap for the shielding. below is a look inside the box.

To complete the fixture some generic mounts were attached to the top, that will make using other rails very easy.  the wire come out of the tub curving up to the rail completing this leg of the project.

Using the tub worked out great.  It makes a handy carrying case for all the stuff needed to fire.  Extension cords, multi-meters, nylon tubing, the rail, and tools all fit with the lid closed.  tomorrow were going to go play with it.

Saturday, June 4, 2005

The Rail gun will be making its first public appearance in a few weeks.  I met a guy (John) at a bar and became friends with him.  We start chatting about hobbies and I blurt out "rail gun."  And he replies, "jigga what?"  So he goes on to mention a party he's throwing and figures he and his friends would enjoy a demonstration.  I just hope his neighbors don't get scared easily.

So now the project is to make the Rail gun and all the support equipment portable, so it can travel to the party.  It also needs to be easy to setup once we arrive.  The current idea is to put everything in a huge plastic tub.  The Rail gun will have to be mounted to the tub in such a way that its sturdy but collapsible.

A redesign of the inductor is also due.  Current ideas involve copper tubing commonly used to plumb auxiliary water fixtures in homes.  The tubing would be have to be coated with an insulating layer in order to wrap like a regular coil. The insulator might be heat-shrink tubing or an enamel lacquer. 

An inductor could be made without adding an insulating layer simply by using plastic to support the coil like a backbone.  Two pieces interlinked with notches along the edge would be simple and effective. A proposed design is shown below.

Friday, June 3, 2005

Some work was done on the 18" pieces of nylon for Rail B.  It was decided that the holes in this version would all be spaced regularly.  1/4" holes are drilled every 3/4" along both sides.  It would allow us to change the configuration more easily.  Rail A has some holes, used for electrical connections, that are in irregular places making a design change more difficult. Multiple rail/injector assemblies could be produced ant tested.  Changing from one setup to the other would be quick and easy.

Wednesday, June 1, 2005

Beginning of the Inspired Rail BLOG!  This BLOG is beginning after this project has been active for around 6 months.  The project began as a collaborative effort between myself and my fellow "Mad Scientist" Brad. 

A not-so-brief rundown of the project so far:

It started with scrap pieces of angle iron screwed to pieces of plywood, simple transformer based chargers and a smallish cap bank (40 x 2700uf, 200V electrolytic) capable of storing approximately 54J each or 2160J put together. Q=.5*C*V^2  

The original  projectile chosen was the common American penny. Cheap, easy to find, fairly standard size, all made it a logical choice.  That's right, we defaced money, we're going to hell. 

The Proposed target is a stack of phone books.  I remember seeing on TV, in project run by the government, a pile of phone books was used to catch a projectile after it had punctured 2" of plate steel.  If its good enough for government work its good enough for us.

The first experiments weren't all that successful. It was difficult to ensure the rails were parallel and evenly spaced along the entire length (especially important for good conduction).  Plywood is not the best substrate when accuracy matters. Several of the first shots were "successful" though not very impressive. successful = The muzzle velocity was higher than the injection velocity. 

The scrap steel angle iron was replaced with aluminum from the local Ace hardware. Using new angle iron, instead of used, made aligning the rails easier, but aluminum melts at very low temperatures and began disintegrating rapidly. after 2-3 shots the rail had to be disassembled to clean the surface of the rails so the penny would roll through.

The angle iron Rail gun has several inherent design flaws. It was found that a considerable amount of current was flowing through the plywood while we were staging the shot since the rail is always connected to the caps.  Wood is an insulator but its not that great. Also, without a cover over the rail, the penny generally just jumps out of the top of the rail before the end.  This may be caused by the fields developed by the odd shape of the rails, or the un-smooth surface of the plywood. Either way the penny would enter, accelerate for several inches then lift out of the rail, loose connection and fall back into the rail and spark some more with no acceleration. 

The solution to the imbalance of the vertical forces is to use rails that are symmetrical about the projectile. Brad and I brainstormed many hours trying to find a new rail solution.  The new solution needed to fill many requirements.

• Simple. few parts that stack together, without interlocking pieces, that can be swiftly assembled and disassembled to perform maintenance on the rails every few shots.
• Strong. If you have read the "Theory" page you know that the force put on the projectile is 1/10 (or less) the force applied to splitting the rails apart. The new rail design must be over engineered to prevent failure for monetary and safety reasons.
• Cost effective. We needed to find materials that had close to ideal properties while staying in budget.  There are exotic hybrid plastics, like Garolite, that are incredibly strong, high temp, resistive, and don't shatter, but are prohibitively expensive.

Our new design was going to all theses things and more.  The requirements say the two rails need to be electrically isolated, mechanically ridged, and resistant to flashes of outrageous heat. The simplest design we could think of is two strips of copper sandwiched between thick nylon with a ridiculous number of bolts holding it all together. 

Nylon is a good insulator, strong, but not so temperature resistant.  Fortunately the heat from the plasma arcs comes and goes quickly, so the immediate surface of the plastic may melt but the block will not weaken structurally.  Copper was an obvious choice for the rails; great conductor, great heat dissipation, relatively easy to find, and it won't break the bank.

The new design readily lends itself to experimenting with multiple configurations. The backbone of plastic can be used with many different rail designs (tall and narrow, wide and short, square, etc) enabling many different experiments.

As this BLOG is beginning, The first of the newly designed Rail gun's is complete.  We'll refer to it as Rail A.  It has 3/16"x1"x6" copper rails and 3/4"x3"x8" pieces of nylon and 3/16"x2"x2" pieces of injector mounting.  Rail A is built to fire BB's. A brass tube is mounted in the injector portion of the Rail gun, and a nylon tube is connected to that.  The nylon tube we used is generally used to supply water to a fridge for the ice maker.

Rail B will be made with 2 pieces of 3/4"x3"x16" nylon. and Rail C is slated to use 3/4"x3"x24" pieces of nylon.

When Rail A is fired, a BB is loaded into the nylon tube at the open end, away from the rail.  compressed air is used to send the BB down the tube at high speed.  The BB enters the brass tube and its trajectory is aligned with the rails, and then it enters the rails and sparks fly! 

It was found that without the brass tube and a near perfect entry into the rails, the BB will ricochet all the way to the end of the rail. The ricocheting BB wastes considerable amounts of kinetic energy to friction. In one case the BB left the muzzle at higher speed under air pressure alone than the powered ricochet shot.  The Rail gun was dismantled to investigate and a series of dents were discovered that created a pattern alternating from rail to rail spacing out as they neared the end of the rail.  Obviously not very efficient.

The Best shot from Rail A so far was on the order of 1000V, 650J. This shot, by a BB, ripped 80 pages into a phone book. All the way to Brown.  The BB showed no signs of its Rail gun experience. The surface was smooth and shinny just like it had started.

The test shots afterwards were on the order of 800-1000V, 2000-3000J.  In the higher powered shots the BB never penetrated the phone book.  It left dents but bounced off and rolled away.  In most of those shots we couldn't find the BB's cause the shop was scattered with them. 

We did finally find one.  It was mangled, and completely dull (all of the shiny surface was gone).  the BB was now 75-80% of its original mass and oval.  looking at it in one direction it was still circular, but turn it 90º and its an oval.

It appears BB's, although cheap and easy to find, begin disintegrating under higher power/current.  I had heard somewhere that ferrous materials don't work well in rail guns. maybe it has to do with the fact that they can be truly magnetized, whereas other metals can't.  The reaction in rail guns is caused by electromagnetic forces which are transient by nature.  Another armature must be found for higher powered shots.

The project got a boost in Jan-Feb of 05' when we picked up some caps from a guy in Florida.  there are 48, 450V, 2400uf Mallory electrolytic's.  The caps probably came from UPS units used in industrial machines. they cost $80 and $130 to ship from Florida to Oregon. The box weighed 131lbs.  I feel bad for the UPS guy that had to get them in and out of his truck.  If you run the numbers, they are capable of storing about 240J each or 11.6kJ as a bank. That's approximately 6 times larger than the original bank.