Rail/Coil Gun Timing

Philldapill

About 2 years ago, I built a high power coil gun setup. I was using what might be considered overkill - qty: 25, 3900uF 400V Capacitors in parallel/series to fire a gun with 10gauge wire for the coils. That's 7.8K Joules.·I got the projectile to fire at a pretty high speed. I was using a series of optical triggers to trigger the next stage, but never used any uProcessors or anything. I've gotten the gun out of storage and want to play with it again. Now, I'm thinking about using the prop to control the firing!

Here's what I'm thinking. Now, I know this will probably never happen, but I'd·love to get this bad boy to fire a projectile at the speed of sound ~1100ft/sec. Now, that's about 356m/sec. Each coil on my gun is about 5cm long, so I figure at top speed, the projectile will travel through each 5cm coil in about 140 microSeconds. I'm thinking that I can throw away my whole optical triggering setup(which takes up about 1/4" between stages). I think I can use the prop to simply trigger each stage in sequence at a particular, predetermined, interval.

The question to all this is do you think the propeller can handle triggering it's outputs in time to reach this 140 uSec. interval? I plan on finding the best trigger times by simple trial and error with little to no calculations involved. In addition, I'd like to use the prop along with a simple optical sensor at the end of the barrel to find the speed of the projectile.

Phil Pilgrim (PhiPi)

You must've just gotten your copy of Nuts & Volts today!

Mike Green

In assembly language with an 80MHz clock, the instruction time is 50ns for most instructions.
That's 2800 instructions during the transit time through a coil. I don't think the Propeller's
speed will be a limiting factor. It'll be other parts of the system.

stevenmess2004

Philldapill said...
I'm thinking that I can throw away my whole optical triggering setup(which takes up about 1/4" between stages). I think I can use the prop to simply trigger each stage in sequence at a particular, predetermined, interval.

This will mean that you will only be able to use one size/weight for your projectile. Hook the optical sensors up to the prop and then you can do other fun things.

william chan

I had a portable pistol type coil gun using one 2,500uF 400v capacitor charged to 320v but it's performance is dismal.

How many stages ( coils ) do you have?
Are you using SCRs for the switches?


I think one problem could be the response time of the optical switches.
Another potential problem could be that the high EMF produced during a coil discharge could make the Propeller go haywire.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my

Jonathan

What are you using for a projectile and barrel?

I have built a few (much more humble than yours) rail guns. I tried using a·uC (an SX chip) for triggering on one. I got more erratic results than using opto triggering.·My conclusion after much messing with it was that variances in strating postion, projectiles and barrel friction were making the timing variable.

With a 5cm projectile, what spacing are you using on the coils? Are you sure you aren't getting any "suck back"? Are the coils the same length as the projectile?

What kind of cap are you using? Is it a really low ESR pulse type, or a power supply type cap? Looking at my last rail gun on a scope, I can see the discharge of the caps is way too slow, and the projectile is long gone before the discharge is finished. This means mine has the dreaded "suck back".

I doubt that at this point the timing is what is holding you back. What speed are you getting, and how are you measuring it?

Jonathan

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.madlabs.info - Home of the Hydrogen Fuel Cell Robot

Jonathan

Err, substitue "coil" for "rail" above. I've built both, had a lil brain fart there.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.madlabs.info - Home of the Hydrogen Fuel Cell Robot

Rayman

You can get 1-us precision output pulses with my delay generator object here:

http://www.rayslogic.com/propeller/Programming/RaysStuff/RaysStuff.htm

Aleks

As is being pointed out already, using a timing sequence to fire the gun is impractical in that the slightest difference in any component of the gun will directly affect the speed of the projectile. Every projectile will fire a little bit differently, and definitely not at the same speed. Thus, by always using the same time intervals for your triggers you aren't getting the maximum effect needed for what you desire. I have to agree with stevenmess2004 and say implementing your original optical triggers in conjunction with the ucontroller will give you the best results.

Philldapill

I failed to mention the beauty of my design. I'm not using "stages" in the normal meaning. I'm actually using 5 connected stages. Each stage is connected to the next, but the connection point goes to an HIGH POWER SCR. Basically, you trigger the first "stage" which is actually the full length of the stages in series. The next trigger, is an SCR that "shorts" the coil closest to the starting position. The next SCR shorts the previous two coils, and so on. What I believe I'm ending up with is a solenoid who's "center" is moving closer to the end of the barrel. The final stage simply shorts out the remaining two coils. I know, there is probably a good bit of energy wasted there, but hey, this is a tabletop gun, not a hand held pistol.

In regards to the EM interference that may make the uC go nuts, I'm worried about it. When I fire this thing, nuts that are 4" away jump 1/2" over to the coils. That's just the primary field that it generates, not to mention all the induced current in other things(circuits). I'm thinking about using a high speed opto-isoloator to try and cut down on any emf produced from the pulse.

BTW, this is a project behind my CNC machine which I don't expect to finish in the next couple months so the gun will have to wait. Thanks for the suggestion though guys!

Philldapill

Phil, no I don't get nuts and Volts. Is there something in it about this?

Jonathan

So, you only have one cap? And it fires through all of the coils?

And what speeds are you getting? What are you using to measure it? And what kind of barrel and projectile are you using?

Inquiring minds want to know :P

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.madlabs.info - Home of the Hydrogen Fuel Cell Robot

Philldapill

Ahh, by the way, my giant SCR is rated for something like 45KA, 1200V. My cap bank is configured as 5 sets of 5 caps in parallel(3900uf, 400V each cap), and those 5 sets are in parallel. Do you think I should maybe only use 3 banks of 5 each in SERIES, charged slightly less than the 400V limit, for a total of ~1200V? This way, I think I can get an even faster rise time on the pulse?

Yes, I have gotten suckback before. On the start-side of my gun where the projectile goes, I have a 1/2" plexiglass shield configured in a way that it blocks the projectile in case of suckback. The shield got severely cracked one time when I was toying with some adjustments, and the proj. sucked back so hard, it slammed into it, cracking it.

In regards to the traditional coil gun efficiency, I was wondering something... Do you suppose that it may be a better idea to use the huge solenoid to suck a metal rod IN, and as the metal rod goes in, stretch/compress a high-spring constant, spring? This way, you could probably get much more kinetic energy conversion as the spring launches a projectile afterwards?

Philldapill

Oops! Sorry Jonathan, I forgot that. I'm using a copper tubing barrel(eww, I know), electrolytic "computer grade" caps(25 total, 3900uF, 400V, 312J each). I WAS using the optical sensors to start and stop a 555 timer, which would then toggle flip flops and drive a 7 seg display... And then calculate the speed from the display readings. Pretty accurate. The Projectile is just 3/8" steel rod from home depot. Oh, and the barrel is 1/2" I believe.

Jonathan

What you are calling a suckback I always called a backfire. What I meant by suckback is when the projectile has passed the center point of the coil, but the cap is still dumping power thru the coil. This results in slowing the projectile, as it is being attracted back to the center of the coil. The ideal is to have all of the juice dumped just as the projectile hits the center point.

I also got a performance bonus when I started making projectiles from brads or small finish nails. I would cut off the heads, grind them smooth, then coat them with a light coat of varnish. Then I would fill a 1/4" ID wax impregnated cardbaord tube wityh epoxy, then stuffed the nails in. After the epoxy dries, remove and dress. I always wanted to get some hex stock to try that with.

I would have thought a copper pipe would act like a shorted turn. Hmm.

Anyway, we are gettig way OT here, so if you would like to talk coil guns, send me an email to jpeakall AT madlabs DOT info

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.madlabs.info - Home of the Hydrogen Fuel Cell Robot

william chan

You can't use a copper tube as a coil form (barrel) b'cos too much energy would be wasted in eddy currents.
Try a plastic tube instead.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my

Philldapill

Jonathan, yes, what I'm refering to is suckback. What is happening, is as the proj moves through the center point, it is slowed by the still active magnetic field and pulled BACK in. The direction of travel changes completely 180 degrees. By the time it re-enters the center, the field begins to collapse and it continues moving in that direction(backwards).

In regards to the copper pipe, I mentioned "eww" earlier. There is a fair amount of eddy current in the tube, but remember... 7.8KJ of energy is being dumped into the coils. The field is pretty intense even on the outside where most of the field is being canceled out.

I tried a little test one time. I put the projectile in at one end of the coils, stuffed a wooden dowl rod in in front of it and fired it with only the first triggering. That dowel rod shot across my shop, hit the other wall, and splinterred in half. I'm just glad I wasn't in front!

BTW, I don't think we're too off topic. Well, maybe sorta... This is a propeller forum after all - not a coil gun hobbiest forum!

So, with that said, next question. What kind of precautions should I take to prevent the massive EM from messing with the prop? I wanted to see how harmful the gun is to electronics, so I put a cheap old calculator right under the coils a few times. The 4th firing, the calculator wouldn't turn on, so I guess I killed it.

Chuck McManis

The last time I tried this I was using a loop of tube. The idea I had was to work the speed of a BB (actually a .250" ball bearing but the same concept) by sequencing coils outside a formed polycarbonate tube and using optical sensors to shut down the coils (and re-power them when the BB got to where it was back in the accelleration path) The "exit" was an electro magnet that supplied the 'pull' force over the gap where the BB could exit (this was based on a classic synchrotron design I read about in my quantum physics textbook, except they were steering protons and I was trying to steer a steel ball.

To be completely realistic, it did not work nearly as well as I had hoped it would. If the BB had enough energy to go around the tube the solenoid rarely had enough energy to keep it from exiting, and if I pumped up the current in the solenoid to keep it in the tube the BB would be slowed by the attraction to it and wouldn't keep accelerating. All in all quite dissatisfying.

I recovered some satisfaction when I turned the whole contraption into a sort of clock. Because I had uC control of the coils I could, by modulating the current into them, position the BB at various places on the circle and make it stop. I never had the time or energy to do the math to have two BB's in the loop where one was acting as the minute hand and the other the hour hand. Frankly I don't think the physics allows for them to be in the same tube (even if you changed which was which when the minute hand passed the hour hand) and the interference between the coils would make controlling an inner loop and an outer loop very very complicated.

--Chuck

Chuck McManis

From the Nuts-n-Volts web site cover page. Their cover article is a coil gun.

--Chuck

Beau Schwabe

Philldapill,

There are a couple of sentences that you said that I will focus on....

"In regards to the copper pipe, I mentioned "eww" earlier. There is a fair amount of eddy current in the tube, but remember... 7.8KJ of energy is being dumped into the coils. The field is pretty intense even on the outside where most of the field is being canceled out."

Even using a copper pipe, and producing results capable of launching a wooden dowel rod via a metallic object within the copper tube across your shop with enough force to splinter in half I'm surprised that your copper pipe doesn't experience severe overheating problems or even "hourglass pinching" at the main coil location.


"....next question. What kind of precautions should I take to prevent the massive EM from messing with the prop?...."
Consider the effects of your "EMP blaster" <- It's really not a coil gun at this point is it? .... and also consider that the general design inside the IC to deliver Power and Ground supplies to the IC is made of continuous metal loops around the perimeter of the IC core. It's not just the Propeller IC, this is a common technique for many other IC's.... eventually that Power and Ground "ring" will super heat and go POP!!

Any hope of shielding is going to be in the form of a thick metallic enclosure with several thick laminate layers forming a box inside of a box scenario.



Don't get me wrong, I like high voltage power experiments, and have done my fair share of them, but what you are messing with can be extremely lethal not only to yourself, but others. Use EXTREME caution at ALL times.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

stevenmess2004

If your worried about em than just put the prop in a faraday cage like Beau suggested.

Philldapill said...
In regards to the traditional coil gun efficiency, I was wondering something... Do you suppose that it may be a better idea to use the huge solenoid to suck a metal rod IN, and as the metal rod goes in, stretch/compress a high-spring constant, spring? This way, you could probably get much more kinetic energy conversion as the spring launches a projectile afterwards?

For a spring the stored energy is E=(kx^2)/2 with E-energy in joules, k-spring constant in N/m and x-deflection in m.

If we solve this for x and assume your energy of 7800J and a spring constant of 50kN/m (seems to be about a car spring) than you will need a deflection of about 560mm (almost 2 ft). You will also have a force of around 28kN (about 2800 kg) so it would have to be pretty strong. I don't think that you would get any efficiency gain as you are introducing 2 more energy transfers into the system. However, if you used something else to compress the spring (some hydraulics or pneumatics or something. but be careful. Hydraulics can be deadly!) than you would probably get much better efficiency.

For some numbers. If your gun is 100% efficient than the velocity of the projectile is equal to sqrt(2*E/m) where E is energy in Joules, m is mass in kg and the velocity in m/s.

Now if your projectile was 100 g than you would get a speed of 395m/s which is faster than sound. So you clearly have more than enough energy. The problem is how efficient is it? If it was 50% efficient than you would get 279m/s.

Post Edited (stevenmess2004) : 2/29/2008 9:00:37 PM GMT

Philldapill

Haha, no I hadn't seen the article, Chuck. I'll have to read up on it.

Beau, thanks for the insight. The copper tube(and coils for that matter) get VERY hot after only 2 consecutive, full power firings. As you may suspect from that, it takes a LONG time to caliberate things in between firings because of the time it takes to cool down. I'm sure I could come up with some sort of cooling method, but I like working on something, taking a 30 minute break to eat or do whatever. haha As for the pinching, I haven't seen any. The coils are made up of 10gauge wire, 4 layers, and the whole length of coil(s) is about 10" long. Remember, ALL the length of coil is energized at once.

Steven, if I remember from my physics class, PE = 1/2(kx^2), but I think you got the deflection right. Doing some calculations, I find that all this coil gun business is nonsense. I'd love to see a design that achieves 25% eff. with even a decent amount of energy, let alone 50% conversion. I said my goal was to get the proj going faster than the speed of sound and now I'm lauging at the thought. Ha-Ha.

Thanks guys!

stevenmess2004

Philldapill said...
PE = 1/2(kx^2)

If you mean potential energy than it is PE=mgh where m=mass in kg, g=9.8m/s and h=height (this is not strictly right but it is a good approximation close to earths surface
Spring energy is SE=1/2(kx^2) where k=spring constant in N/m and x=deflection in m.
Kinetic energy if KE=1/2(mv^2) where m=mass in kg and v=velocity in m/s

yeah, I missed the ^2 out below. I used it in the calcs I did. Will go and fix it

Another interesting power=amps*volts so you have for the SCR 45kA*1200V=54MW. Where on earth did you get something that handles thst much power?

Philldapill

My first design for the gun was just one big stage, and one big switch - the SCR. I got it for a whopping $100 on ebay. I've attached a spec sheet with this post.

I really have no need for this thing anymore. Anyone want it for $50?

Philldapill

I gotta say, this must be a fascinating topic. 2 days later and this post has over 400 views, and 23 replies.

In regards to me selling the SCR, looks like I got a buyer in under 8 hours. Thanks Bill!

Sleazy - G

· THE GUINNESS RECORD HOLDING·method of making hypersonic railguns is not with coils, but straight rails.·The system includes an insulative projectile·(usually ceramic disks) , and a conductive slug (copper) immediately behind this disk.· When current is sent through the parallel rails , the conductive·slug vaporizes, resulting in·ionic gas/plasma.· This charged copper gas/plasma phase is "forced" down the barrel by the natural magnetic field generated between the rails.·



The accel approaches 0 to 1000km/sec in 1ms· .· This is no typo, actually a physics textbook statistic i had committed to memory

dont use coils in railguns if you want to be for real

Philldapill

Yeah, I think I've heard about that, too. I saw a picture one time of the "gun" which is on wheels and enormous.

Here's a site that's pretty interesting. All sorts of things related to this. Oh, and he has a pretty hefty RAIL gun that seems to work pretty well.

http://www.powerlabs.org/

deSilva

stevenmess2004 said...
Another interesting power=amps*volts so you have for the SCR 45kA*1200V=54MW. Where on earth did you get something that handles thst much power?

Power is a myth What do you think happens inside a crystal (or a PC chip)?
54 MW times 100 ps is a meek 5 mJ

Philldapill

deSilva, where did you get 100ps from? In regards to the 45KA surge, the time on the datasheet is 10ms. 10ms is 1/100s, so correction - 540KJ. I see your point though. It's not like the SCR is dissipating that much power, it's just allowing that much THROUGH. Still, I'd like to see YOU pass 2.5KA continuously. [noparse];)[/noparse]