Registered users: Bing [Bot], Google [Bot], Google Adsense [Bot], MrRobot
I just won an all expense paid trip to the Intel ISEF, and I have a few weeks to do more research. I have a few specific questions.
First of all, I am researching electromagnetic induction, and I am using a ring launcher to do that. basically dumping a capacitor bank though a pancake coil. Google for more.
Anyway, I am going to get acceleration vs. distance from the coil. At the affiliated fair where I won the trip, I was talking to some engineers, and they suggested an LED on the projectile, blinking at a few khz, and taking a long-ish exposure image of it. I thought it was a great idea, and gave it some more thought. I am trying to decide whether to "measure" the length of each streak left from the LED, or the distance between streaks. How many velocity points do you think I need to generate a decent graph?
Also, I realized that whatever circuit I use will be exposed to the ridiculous magnetic pulse, and must be hardened against basically getting EMPed. Somehow, I don't see a Faraday cage working for this. Any help would be great. Estimated rise-time of the current pulse is 200 microseconds.
Also, since I will be testing during the day during spring break I need to do this in my basement (for darkness), I need to absorb around 200J of Kinetic energy from a projectile that is moving straight up, and not have it drop on my cap bank, or my head. Any suggestions?
Last edited by ramses on Sat Mar 20, 2010 9:35 pm, edited 1 time in total.
My first thought is the frames per second you can record..then the LED freq can be varied either way. I expect a match would be ideal but obviously the led can be much faster. I would make the LED flash as fast as possible and still be caught by the cam. I think that will give the best measurement. I'm guessing but I would go with the on time.(start of a streak).The off time would be more varied imo
Good luck not blasting a hole in anything important.
no, bullrees, you don't understand. He is going to take a single, long exposure, in a dark room, of the LED strobing. Like with a camera, not a camcorder.
Just randomly, I would think 10-20 data points over the course of the acceleration would be more than sufficient, but I don't know how feasible that is. I don't see why a cage wouldn't work. Just make sure to counterbalance it with something on the other side. You also need to make sure that your image will have enough resolution. If you are strobing that fast, you need the ring to travel more than the width of an LED each strobe, otherwise you will just get a blurred line.
Also, I would suggest defusing the LED, so that it doesn't lens flare in the camera.
Do not look back, and grieve over the past, for it is gone;
Do not be troubled about the future, for it has not yet come;
Live life in the present, and make it so beautiful that it will be worth remembering.
Thanks. The you are correct with the long exposure thing. The reason I didn't think the cage would work is because the circuit will still be well within the magnetic field. IIRC, the magnetic field will pass through the cage, and still induce the current. My lens seems fairly resistant to flare, so I may end up leaving the LED. I will use green because the pattern has 2x green pixels as either R or B.
6MP should be enough resolution.
Additional problem: I need to submit an abstract before I have a chance to do these experiments. Any advice?
a cage will work, solid shielding is better. The material (higher permability the better) makes the magnetic field go around the material:
A higher and higher permability material is needed for higher frequencies. Hymu-80 is the best you can get, but steel will probably suffice.
Won't it saturate?
Thats the key, finding the right material.
Even if it does saturate, it still offers a lot of protection.
Even wrapping in aluminum foil helps, a lot of high end labs use aluminum foil for low level shielding.
A cheap and dirty trick I have used in the past is now much cheaper to do than before due to inexpensive disposable cameras.
Using a quartz 4 pin ocsillator module, of a common value such as 1 or 4 MHZ, cascade a couple cmos decade counters such as 4017.
Use a decade where the time gives a good interval such as 4KHZ for a pulse ever 1/4 ms. Use the individual outputs to drive a set of disposable camera flash units. Tie the trigger of your coil with a opto coupler to the reset pin on the last decade counter. When the coil is fired, the camera flash units fire in 1/4 ms intervals or other interval you program. Use an apropiate high speed high voltage mosfet to interface each camera flash with each trigger line from the 4017.
As a bonus, each flash can be color filtered to track when each fired.
Set a film camera for the flash intensity for only one flash. the moving ring will be properly exposed while stationary components will be overexposed. Proper placement of the strobes can be used to properly expose a yardstick for example by only permitting one flash to illuminate it.
Congratulations. If you can get some LN2, freeze your projectile prior to launch. You will like the result.
An energy dissipation target for the projectile is recommended. A gallon milk jug full of sand works well at scattering the KE of your projectile. Suspend it directly above the launcher.
I don't think either Al or Cu superconducts at LN2 temperatures...
I also just realized that to measure voltage, I need to ground my circuit. I don't think I should use my safety ground, but I don't feel like I should use the "neutral" (wide) part of the plug either.
I am spreading this discussion to 4hv.org
Neither Cu or Al superconducts. The change in resistance is dramatic and makes a huge difference in launch power.
Look up the Temperature Coefficient of both materials in the LN2 vs room temp range.
If you search the online videos of the launcer you are building there are several videos posted showing shots with room temp vs cooled rings. I can't take the time now to find them.
Copper and Aluminum have the same coefficient.
Edit: found the video with a very quck search. 9 ft vs 20 feet cooled.
I suppose I might attempt that. Provided I can find a source of LN2 and the time to do twice as many tests. Then I need to fill in my safety form for LN2, as well. It is crowded as it is.
Try making your cage out of aluminum and ground it well. Aluminum has the tendency to repel magnetic forces whereas iron and steel accommodate it. Bismuth will repel very well, also. It is why an aluminum ring works so well being propelled by a magnetic force.
yeah, the circuit will probably be wrapped in AL foil, or steel mesh. As to grounding it, it will be ON THE PROJECTILE!, meaning no ground. The reason AL is repelled is because of EM induction, not paramagnetism.
Time to find a supplier of LN2... I wonder if my school has a Dewar flask...
Just a quick note on flying electronics exposed to an EMP. Don't use a ground plane with multiple connections. The induced current and distributed resistance can cause voltages in the plane. Use twisted pair or coaxial for all signal, and power connections. Use a single point ground. You have to avoid ground loops due to the EM coupling. Using twisted pair or coaxial provides a high degree of common mode noise rejection. Using a single point ground prevents induced currents in data, signal, and power lines. You must use good construction practices or the EMP will kill it.
This means all power leads from the battery is twisted pair. All LED leads are twisted pair back to the source. All grounds tie together in only one place, not scattered on a ground plane. Use a ground plane for shielding. It will have high induced current in it shielding the electronics above it. It is tied to the single point ground with only one short ground wire. The single point on the ground plane can be the single point ground for the entire circuit.
For your launcher, the leads between the work coil and capicitor bank needs to be as short as possible. The forces on the wires between the cap bank, switch, and coil is very high. Use very stout mounting for them.
PM me if you need help on the subjects of ground loops and common mode noise rejection.
So, why use a paramagnetic material for the projectile and HAVE to use LIN to make it work well? Why not use a ferromagnetic or even a diamagnetic projectile?
Who is online
Registered users: Bing [Bot], Google [Bot], Google Adsense [Bot], MrRobot