close-circle
Close
0%
0%

Gauss rifle v2.0

An improved version of my first gauss rifle.

Similar projects worth following
This project is going to be an improvement over my first multi-stage coil gun design, which only archived a muzzle energy of 3J. This time i want the muzzle energy to be in the tens of joules.

This new design will be able to shoot steel BBs in full-auto mode and spin stabilized bullet-shaped projectiles for greater range and accuracy.

In this project i will document the design and build process of my second Gauss rifle.

In every log i will walk you through the design and construction of this second Gauss rifle or "projectile accelerator" as i like to call it, i will explain the changes i make; why and how they improve upon the design of the first version and suggest further improvements in case anyone chooses to replicate my project.

I hope you enjoy reading it and take inspiration from my work.

stencil.png

Stencil template for the v2.0 PCBs

Portable Network Graphics (PNG) - 645.00 bytes - 02/15/2017 at 10:57

eye
Preview
download-circle
Download

Gauss rifle v2.0 gerbers.zip

V2.0 PCb gerbers

Zip Archive - 23.95 kB - 02/15/2017 at 10:56

download-circle
Download

Gauss rifle.sch

V2.0 schematic

sch - 346.76 kB - 01/01/2017 at 00:48

blank
See BOM
download-circle
Download

Gauss rifle.brd

V2.0 PCB

brd - 85.21 kB - 01/01/2017 at 00:48

download-circle
Download

coilgun.sch

V1.0 schematic, in case someone needs it

sch - 322.62 kB - 01/01/2017 at 00:08

blank
See BOM
download-circle
Download

View all 6 files

  • The busbar connection

    Leandro02/28/2017 at 11:33 0 comments

    To isolate the copper from the aluminium profiles, i used a long piece of biberglass mosquito net.

    The net keeps the copper from rubbing against the aluminium, keeping them electricaly isolated.

    Here you can see how the copper bars sit nicely in the net lined assembly, the PCB fits on the copper bars, and the coil wires and photodiodes have just enough room to get through.

    Once all the PCBs were placed, the photodiodes were soldered and the drill positions marked onto the copper bars.

    Holes were drilled and taped for M3 screws.

    The electrical connection between the PCBs and the copper bars is made with 5mm brass screws. Steel screws would do ok (i used regular ones in v1.0), but i managed to find them in brass. They will lower the power losses by electrical resistance.

    I planned to glue the bars to the aluminium profiles through the net, but i think that when the coil wires get soldered to the boards, the busbars will be held in place just fine.

    Now i need to add connectors for the batteries. The batteries i bought use 5.5mm gold plated connectors, which should be adequate for the high currents the rifle will draw.

    I cut two pieces of round copper bar and turn down the ends to press fit the connectors.

    The connectors were soldered to improve the electrical connection and the bars bent to about 90º.

    They will be connected to the busbars facing opposite directions and slightly angled down.

    For this, i carved a recess in the busbars and a mating flat surface in the connector bar. A hole was carved in the aluminium profiles too, leaving a bit of room for insulation on the copper bar.

    Then, i soldered the connector bar to the busbar and filed it flat to match the busbar surface. I made sure not to solder the connector for the battery + cable onto the - busbar and viceversa.

    I repeated the process with the other connector and this is what they look like:

    I covered the connectors with heat shrink tubing so there would be no exposed metal to short the batteries together.

    For now, the busbars are done. I leave the back end exposed because i will need to place a board there to control whatever projectile injection mechanism i come up with in the future. As you can see, all the boards fit nicely on the connection part, which is centered on the coils so as to reduce the resistance to the furthest ones.

    The batteries wil be placed under the coil section of the rifle, leaving a gap between them where i will place a grip for the rifle and the connection between the batteries, which will likely include a protection circuit.

  • The aluminium structure

    Leandro02/15/2017 at 10:48 0 comments

    The barrel and coils are somewhat flexible by themselves, so they need something to keep them straight and give some rigidity to the structure.

    In v1.0 (picture above) i used anodised aluminium T profiles as both busbars and structural elements. They proved to work well as the structure for the barrel, but they presented a few problems as conductors.

    Even with the thick oxide coating, shorting of the busbars happened a few times, and given the power that the batteries can deliver it made for quite a firework show. Also, such a thin aluminium profile doesn't hold threads very well; tightening a screw a bit more than necessary would tear them right off.

    As i liked how well they fit over the coils and kept the barrel straight, i decided to use them again for v2.0, but with a few changes in design.

    Instead of screwing the profiles to the spacers as in v1.0, i decided to glue them in place to give the gun a cleaner look.

    The T profiles are 15x15mm. This size leaves the sides of the coils visible.

    This time i decided to use copper busbars to carry the current instead of the aluminium profiles themselves. I looked for 4x10mm or 4x12mm copper profile, but i couldn't find any so i had to use trimmings from 4mm copper sheet.

    In the top and bottom, the profiles leave a gap that allows the coil wires and photodiodes to come through, and a nice pocket to place the busbars and PCBs.

    I also like this kind of profiles because they're a good place to fasten other components to the rifle.

    Like the solenoid that feeded the projectiles from the magazine in v1.0.

    In v1.0 the gun handle, the battery holders and the strap rings were also bolted to the T profiles.

    In the v2.0 rifle i plant to fasten things in a similar way, given how easy it is to tread this aluminium profiles. , and how convenient the spacing between them is.

    I just have to be careful to spread loads along many bolts because as i said, any excessive force will destroy the aluminium threads.

  • The PCBs

    Leandro02/15/2017 at 10:37 0 comments

    Once the PCBs arrived it was time for some soldering.

    This are the components needed for each board, not including the 3mm IR photodiode, which would be added later in the build.

    The missing resistor *R1* was later found out to be 100k, so add a fourth one to the list.

    As i was going to populate 10 identical boards just for this project, and would probably do it again in the future, i decided to etch myself a solderpaste stencil out of an aluminium can.

    It makes soldering the components much quicker than solder wire and iron. This way, the components can be placed on the solderpaste blobs without having to solder them individually.

    A minute under the hot air gun, or in a reflow oven makes short work of soldering the whole board.

    For this project i needed ten boards, but it can probably be scaled up by using a few more.

    Using the stencil and hot air method i was able to populate all ten boards in under an hour of work.

    The boards are designed to make electrical contact with busbars through the screw mounting holes. But the bare pads under the MOSFETs are there just for thermal purposes and will be isolated from the busbars with silicone pads.

    In order to leave a bit of space for that pad, and to avoid rubbing of the board with the busbars on the soldermask areas (to prevent wear and shorts), i soldered small pieces of tinned copper sheet arround the mounting holes, to maintain electrical contact but lift the board a little bit from the suface it sits on.


  • The coils

    Leandro01/19/2017 at 13:39 0 comments

    Once all the spacers were finished, the next step was winding the coils.

    In the picture above you can see the dimensions of the barrel with spacers: the tube is 10mm in diameter, centered in a 30x30mm square, and the distance between spacers is 35mm. This meant that the coil would have an inner and outer diameters of 10 and 30mm respectively, and it would be 35mm in lenght.

    The other requirement for the coils was a working current of 100A.

    The coils will be powered by a 22v pack of batteries, so the resistance should be close to 220mΩ.

    (If you're not interested in the math, skip to the paragraph above the next picture)

    The resistance of a metal wire is given by the following equation:


    Where ρ is the metal conductivity (1.678×10−8 Ω·m for copper), l is the wire lenght in m, and A is the Cross-sectional area of the wire im m2.

    Writing A as a function of the wire diameter :

    To find the length of the cable l we need know the number of layers that the coil will have, which we get by dividing the coil thickness (the diference between the outer and inner radiuses) by the wire thickness:

    Where D and d are the outer and inner diameters, (which we know).

    Then we get the number of turns in each layer by dividing the coil length L by the wire diameter:

    And their average circumference:

    The total wire length is the the number of layers times the turns in each layer times the circumference of each turn:

    Feeding this equation into the resistance one we get:

    Finally, solving for ∅ we get the final equation:

    This equation can be used to calculate the wire diameter for a coil made out of any metal given its dimensions (inner and outer diameters, and length) and resistance.


    Using ρ = 1.678×10−8Ω·m (copper), R = 0.22Ω, D = 0.03m (30mm), d = 0.01m (10mm) and L = 0.035m (35mm), we get a wire diamteter ∅ = 1.2×10−3m (1.2mm)

    A coil wound with 1.2mm copper wire would need a length of 15.3m, and it would have 8 layers with 29 turns each. The resistance would be around 0.226Ω. Close enough for me.

    Once the math was done, i bought 160m of 1.2mm enameled copper wire and started winding the coils.

    Each coil was coated with clear epoxy for protection.

    i ended up winding the coils by turning the barrel with a drill. This could be done with a lathe too, but i would advise to drill the holes in the spacers for the IR bridges AFTER you wind the coils and not before; the holes through the tube compromise its torsion resistance and it can break at those weak points.

    Once all the coils were done, i gave them a second coat of epoxy alf left it to cure.


  • The barrel

    Leandro01/07/2017 at 15:46 0 comments

    When i had designed the PCBs i started making the barrel for the rifle. The distance between coils was dictated by the lenght of the PCBs, which are 40mm long. The coil spacers were going to be 5mm thick, just enough to allow 3mm holes to be drilled accross for the IR LED and photodiode pairs. This meant that coils would be 35mm long.

    The barrel itself was going to be a plastic tube again. It could be aluminium, but i would either have to accept significant induction losses or cut a long slit lenghtwise, which didn't sound like an easy task.

    V1.0 used a PVC tube with an OD of 10mm and an ID of 8mm, I tried very hard to find a similar tube but made out of teflon, which would reduce friction with the projectiles and improve efficiency. After a long time searching to no avail, i settled for the PVC tube again.

    V1.0 barrel:

    For reference, this is how the v1.0 barrel looked like before winding the coils. The coil spacers were made out of pine wood, 10x30x30mm. Coils were 50mm long, making each stage 60mm long.

    Investigaing arround, i found out that making the spacers out of a ferromagnetic material, would improve the efficiency of the rifle.

    From Delta-v engineering: "Flux augmentation is the practice of encasing the coils within some magnetic material in order to reduce the reluctance of the magnetic circuit and improve the magnetic linkage between the projectile and coil. Research suggests that ferrous end caps can give a significant performance boost to low power coils"

    So i ditched wood in favour of iron filled epoxy.

    I made a couple of silicone molds for 5x30x30mm spacers and i spent the next month casting spacers, at a rate of one or two per day. (i made a 20 coil barrel too just in case)Keeping the spacers square, parallel to each other and consistently spaced was challenging to say the least. I would recommend building a rig to help you if you need any kind of precision.

    After casting the 11 spacers needed for a 10 coil rifle, i sanded the faces and drilled through 3mm holes for the IR bridges.

    I spaced the spacers slightly more than 40mm to have some clearance between the PCBs. The boards arrived just in time to test the fit, which was as good as i could expect.

  • V2.0 circuit design

    Leandro12/31/2016 at 18:42 0 comments

    I had to redesign the circuit i used in v1.0 to fix the problems it created, and with such a simple circuit to begin with, that meant starting from scratch again.

    (V1.0 circuit diagram for reference)

    V1.0 circuit diagram for reference

    The first thing that i noticed in v1.0 is how slow was the voltage ramp at the MOSFET gates. I traced the problem bach to the photodiode itself. As you can see in the diagram above, the signal at the the photodiode output is the same as the VGS applied to the mosfet (the BJT pair just increase the current availible to the gate). As i've mentioned before, this slow ramp at the MOSFET gate is the most likely cause of the many explosions they produced. In a high current application such as this, the MOSFETs should only work as switches, this meant i had to reduce drasticly the ammount of time they spent switching ON.

    As a side note i should point out that the MOSFETs switched OFF orders of magnitude faster than ON, and the schottky diode absorbed the induction spike just fine, leaving no trace of overshoot at all.

    This is the voltage at the MOSFET gate in the first stage (the one with longer ON time and the most likely to explode). It's also taken when i was using 24v to drive the gates, which was changed to 12v shortly after.

    You can see that the MOSFET was working in its linear region for almost 2ms. This is what i needed to change.

    Other aspect i wanted to change about the rifle as a whole is the sequence in which the coils activate. In v1.0 only one coil was ON at any time, which was the simple way of doing things.

    This is the voltage measured accross a 7mOhm resistor in series with the whole gun.

    We can see that its current draw peaked at ~110A with the first coil, but the other 7 coils didn't reach their steady-state; they never got to completely build up their magnetic field, which happens when they draw the full current as in a DC situation.

    This means that most coils never turn fully ON, and never get to apply the maximum force they could on the projectile.

    I decided to improve the v2.0 design by switching ON the coils several stages ahead of the projectile, so when the bullet is about to reach that coil, it has had time to fully build up a field and impart the most ammount of force it can. The stages still have to turn OFF as the projectile is in the middle of their coil, so as to not slow it down at all.

    This meant that i had to design a circuit that would switch ON the coil when receiving an external signal (from a previous stage) and still switch OFF by itself when it detects the projectile.

    The following diagram is what i came up with:

    The signal that switches ON the coil enters each stage at the supply syimbol with INT on it. A logic 1 (+12v) on this point will turn ON Q5, which will in turn pull down the base of Q2. Q2 pulls high the MOSFET gate and also the base of Q3, which closes this positive feedback loop that keeps the MOSFET ON even when the signal coming from a previous stage (the +12v at INT) drops low again.

    When the projectile reaches the end of the coil (it will be in the middle of the coil if they have the same lenght) the IR bridge will be broken and the voltage at the base of Q1 will drop.

    With this circuit it doesn't matter if it takes 2ms for the voltage accross the photodiode to reach ~10v, which was the point at which the MOSFET would be fully ON in v1.0.

    Here, as soon as the photodiode is a bit shaded, the voltage at the base of Q1 will drop below VCC-0.7v (switching Q1 ON) and it will start conducting current through R2. The supply symbol with IO on it is at the voltage accross R2, and is the signal sent to a stage ahead of this one in order to turn that one ON.

    This signal also turns OFF the coil of its own stage by pulling up the base of Q4, which in turn pulls down the gate of the MOSFET.

    The two diodes in parallel with the coil (PAD1 & PAD2) are two schottky diodes is a single package, the energy stored in the coil when the MOSFET switches OFF is dissipated in these diodes, preventing the formation of a HV spike...

    Read more »

  • v1.0 overview

    Leandro12/25/2016 at 20:27 0 comments

    This is the first coilgun / Gauss rifle i built, let's call it v1.0

    it has 8 stages, all independent from one another. Each stage switches on its coil as soon as the projectile breaks the IR bridge in front of the coil and shuts it off when the projectile exits the IR bridge, when it is in the middle of the coil.

    Steel projectiles are stored in a "rubber loaded" magazine wich attaches to the gun with a pair of neodymium magnets.

    A solenoid in the rear pushes the topmost projectile into the first coil, wich in turn pulls it out of the magazine and propelles it forward to the next stage.

    It is capable of semi-auto fire with a pretty high fire rate. Each projectile weights about 14g and carries almost 3J of kinetic energy, enough to punch through cans and shatter ceramic tiles and glass bottles but nothing too serious.

    The circuit design is fully discrete, and i plan to keep it that way for v2.0.

    In v1.0 each IR photodiode signal is fed into a totem pole transistor pair, which in turn drive the gate of the big MOSFET that switches on the coil. Pretty simple stuff.

    The problem is that the signal from the photodiode is quite a slow ramp, and the BJTs are feeding the exact same signal to the MOSFET, this causes the MOSFET to work in the linear region for a relatively long ammount of time (just a few ms, which sounds like nothing but is actually quite long). At the current levels that each coil operates (~100A) this means a lot of power dissipated in the MOSFETs, which caused them to often fail rather spectacularly.

    This is one of the aspects that i have planned to change. I've designed a whole new circuit that will drive the MOSFETs' gates up way quicker, and hopefully solve the unexpected fireworks problem.

    In between explosions i had enought time to use the rifle on a lot of targets. I definitely had a lot of fun building and using this coilgun, and so did my friends. I leave you a couple of clips that illustrate the v1.0 in action

View all 7 project logs

Enjoy this project?

Share

Discussions

ActualDragon wrote 04/15/2017 at 22:59 point

have you checked out some of these designs? 

http://www.deltaveng.com/gauss-machine-gun/

  Are you sure? yes | no

Leandro wrote 04/16/2017 at 09:15 point

In fact, i built the v1.0 rifle based on that.

  Are you sure? yes | no

ActualDragon wrote 04/18/2017 at 11:47 point

oh XD i was going to build one for a while, but just haven't gotten around to it. nice project!

  Are you sure? yes | no

Chadd Van Komen wrote 03/05/2017 at 16:14 point

Very impressed with your progress, can't wait to see a video.  Out of curiosity have you played around with coil separation to see if one coils resistance to the induced magnetic field from the previous coil is hampering efficiency?

  Are you sure? yes | no

Leandro wrote 03/05/2017 at 19:17 point

I haven't tested v2.0 yet, but i will tell you that energy efficiency is not what i'm after, so i'm cramming as many coils as i can in a short rifle. I want the most energy per shot i can get, but i haven't mut that much thought on the physics behind this.

i uploaded a video of the first version, it is in the first log. v2.0 can only get more powerful than that.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates