10 replies to this topic

[Snake]

I recently saw Rare Rides thread about LIM/LSM launch and thought I would give it a try. The need for precise timing of the coils seemed to difficult for the scale of knex. Tyrant had timing down to the millisecond and still struggled. I thought of a way to remove the timing all together. Base the system off a rail gun.

A rail gun is very simple. You have two parallel rails with current running through them both. A conductive object is placed between the rails and dead shorts the rails, creating a large amount of magnetic flux. The flux pushes the object out at high speeds.

A rail system could be hung over the track and a piece of metal could be attached to the car (think of it as a copper spoiler ). The metal would contact the rails and it would launch the car. A capacitor bank would store the power and release it all in an instant. The capacitors would recharge when the car was moving through the layout.

To try the idea I made a small proof of concept. I hooked up a bridge rectifier to a 120v outlet in my house. I then added a small transformer to act as a current limiter so I would receive less than an amp output. A diode was added in parallel with the transformer to stop inductive kickback from damaging the rectifier or capacitors. The capacitors were in parallel to the two leads coming from the rectifier.

When testing the capacitors were charged at 165v and had a total capacitance of 1690uf, or about 23j of energy. Large sparks could be made by touching the leads together. When a small nail was stet on the rails it would weld its self to the rails. A small nudge broke it free but it quickly re-welded it's self, Shooting out some sparks with it. A thin piece of aluminum places on the rails was blown off the rails as soon as contact was made.

A rail gun is not a very good choice for a launch. Firstly, It sprays sparks on every launch so it is a fire hazard. Secondly, the object has to be really light or it is welded to the rails. If the object enters with speed it will not be welded but then is it really even a launch? And thirdly, there are exposed high voltage rails that can deliver 1000's of amps in an instant. Way more power than is needed to kill you.

So a rail gun isn't a good choice. But a single coil LSM is. One coil would be easy to time. It would allow the whole system to be insolated also to prevent shocks and be more safe. A large capacitor bank would have more than enough power to move the train with only one coil. I still wouldn't recommend it for most people as it can be deadly if you don't know what you are doing.

[Britfag]

I don't think the railgun concept was one that would be suitable in this situation..
timing to the millisecond is easy, I mean, if you have components all capable of quicker than millisecond response time.. jobs a gud'n.

I would consider using multiple IR gates, to control each coil, they respond at the speed of light, and the microprocessor responds in a thousanth of a second..

[TheSUCKCrew]

The thing that worries me about the whole magnetic launch concept is the inaccurate fit of the cars onto the track. When you actually have to leave a margin of 2 millimeters in both sidewards and upwards directions, it's hard to make an efficient launch.

If I would ever attempt this, I would go for an AC design.
Although the calculations might be a bit more complicated beforehand (there will be no sensors so the math behind it needs to be exactly right, and you can't easily compensate for flaws in your design by changing intervals), I think it would end up being the cheapest and most efficient way.

[Snake]

Britfag- Say you have 3 coils. You would have 5 time's that would activate and deactivate the coils. say you have narrowed it down to a 10 millisecond window for each time doing some math. That would leave you with 100,000 possible combinations. I don't think I would have time to test all of them. A single super coil would have only 1 time is needed. If it was made really long like 50cm or so it could launch the car pretty well. The power needed would be really high though. Around the order of 150-200v and a few hundred amps.

Suck- AC would be nice, but the precision needed would be crazy. But it would be extremely efficient.

[CW5X]

I'm not a particular expert with electronics but I have a little bit in my time at school. I'm not quite sure if this would be appropriate but if you had something like a light sensor to sense where the train is and write a short program to get these to activate and deactivate the coils instead of doing lots of maths then it might make it a bit easier. It would involve the purchase of light sensors and overall might not be quite as accurate if you don't position them correctly though. Sorry if I made myself sound like a complete idiot there but that's kinda my understanding of the situation. Just throwing an idea in. Might be completely wrong because as I said, I am no expert but I would be interested to know if this can be cracked.

Edited by CW5X, 07 June 2013 - 05:18 PM.

[Maxlaam]

The problem with railguns is their high voltage and ampere requirements. Unless you have some converter or good capacitor isolated and not connected to any power socket you might be onto something. But if you were to use it from a plugsocket you'll blow your fuses in an instant.

[Snake]

I'm not a particular expert with electronics but I have a little bit in my time at school. I'm not quite sure if this would be appropriate but if you had something like a light sensor to sense where the train is and write a short program to get these to activate and deactivate the coils instead of doing lots of maths then it might make it a bit easier. It would involve the purchase of light sensors and overall might not be quite as accurate if you don't position them correctly though. Sorry if I made myself sound like a complete idiot there but that's kinda my understanding of the situation. Just throwing an idea in. Might be completely wrong because as I said, I am no expert but I would be interested to know if this can be cracked.

Light gates would work, CW5X, but the timing problem would still exist. The smallest gap possible between knex pieces is still too large cut out an electrical timer.

The problem with railguns is their high voltage and ampere requirements. Unless you have some converter or good capacitor isolated and not connected to any power socket you might be onto something. But if you were to use it from a plugsocket you'll blow your fuses in an instant.

Rail guns need high voltage to get more amps to flow through the system. If the system was superconductive, very low voltages could supply massive current with much less power. To double the amps you must double the voltage. Coils are the same as a rail gun. High voltages are needed to push as many amps as possible.

A schematic of the circuit I was using is below. The black box is the rectifier, the blue box is the transformer(which is being used purely for resistance), the green box is a diode( to stop kickback from the transformer), and the purple circle is the capacitor. The input is 120v AC from the wall and the output is 165v DC.

[Britfag]

I'm not a particular expert with electronics but I have a little bit in my time at school. I'm not quite sure if this would be appropriate but if you had something like a light sensor to sense where the train is and write a short program to get these to activate and deactivate the coils instead of doing lots of maths then it might make it a bit easier. It would involve the purchase of light sensors and overall might not be quite as accurate if you don't position them correctly though. Sorry if I made myself sound like a complete idiot there but that's kinda my understanding of the situation. Just throwing an idea in. Might be completely wrong because as I said, I am no expert but I would be interested to know if this can be cracked.

this is what I suggested.. it would be a pretty decent way of working it.
Tell me how to work the launch and I'll do the electronics

realistically, using a single microchip and a few relays, the times could be dialled in one afternoon.

I don't think the timing issue would be that bad.
Considering the forces you're dealing with and the mass you're moving, the acceleration is going to be insane unless your coils are teensy.. I reckon you could dial in the timings using a decent potentiometer circuit and an algorithm to adjust all the times based upon the first gap and a bit of maths to predict acceleration.

I'll.. acquire a copy of something like Yenka, and produce a PCB capable of controlling the coils.

Edited by Britfag, 07 June 2013 - 07:36 PM.

[CW5X]

this is what I suggested.. it would be a pretty decent way of working it.
Tell me how to work the launch and I'll do the electronics

realistically, using a single microchip and a few relays, the times could be dialled in one afternoon.

I don't think the timing issue would be that bad.
Considering the forces you're dealing with and the mass you're moving, the acceleration is going to be insane unless your coils are teensy.. I reckon you could dial in the timings using a decent potentiometer circuit and an algorithm to adjust all the times based upon the first gap and a bit of maths to predict acceleration.

I'll.. acquire a copy of something like Yenka, and produce a PCB capable of controlling the coils.

Sounds like a plan.

[Jogumpie]

Let's say you want a top speed of 30 km/h or 19 mph (that should get you up to 3.5 m or 11 feet considering no friction, so with friction you could probably get it up at least 2.0-2.5 m or 7-8 feet). That's equivalent to 8.3 millimeters per microsecond. Now assume you have a sensor / magnetic coil combination every 2.54 cm or 1 inch, your maximum polarity switch frequency would be 328 Hz. I wonder if you can get the coils to switch polarity that quickly. Of course you can increase the spacing of the coils or decrease the desired maximum speed. It's just to give you an idea.

[Britfag]

Let's say you want a top speed of 30 km/h or 19 mph (that should get you up to 3.5 m or 11 feet considering no friction, so with friction you could probably get it up at least 2.0-2.5 m or 7-8 feet). That's equivalent to 8.3 millimeters per microsecond. Now assume you have a sensor / magnetic coil combination every 2.54 cm or 1 inch, your maximum polarity switch frequency would be 328 Hz. I wonder if you can get the coils to switch polarity that quickly. Of course you can increase the spacing of the coils or decrease the desired maximum speed. It's just to give you an idea.

Surely increasing the coils just a smidge making them perhaps 1.3 inches apart, but increasing the launch length by a lot (looking at 2x length) the acceleration could be slower, but the top speed could be the same?

Also, could it be possible to have larger or longer coils the further down the launch you got, so that your timing accuracy didn't have to be to the microsecond?

Add to this that the chips I am used to run at a matter of 800 MHz, which paired with a set of High-current mosfets or relays (slight delay, but minimal..) would be able to knock out the issue of the switching polarity.

Edited by Britfag, 08 June 2013 - 04:57 PM.