Choosing the right magnet - stuck

[vsage]

Hello,

I am an engineering student doing my first real engineering project with a group of other engineering students. To give a background of my issue, I am charged with the task of designing a system which is able to detect when a rotating shaft spins too fast and send an appropriate signal to a device that will disengage this shaft from its driving source. I've decided to use the rotating shaft to generate a small emf using a wire loop in a constant magnetic field, or at least am using this model, as a voltage source to a circuit which will then interpret the signal and detect when the signal gets too large. I've already simulated the circuit, but what's getting me is how to actually build the voltage source when it comes to prototype building time:

The area I have to work with isn't all that small - maybe a cubic foot or so around this rotating shaft - so I have some leeway on how big I can make the wire loop or how large the magnets can be, but I can't seem to get good info on how to determine what magnetic field my wire loop will be seeing if I place opposite poles on either side of my wire loop. My electromagnetic fields book says that the field can be treated as constant, but I'm not so sure.

With that being said, the only listings on websites regarding magnetic field are measurements taken at the surface of the magnet. If I place a wire loop between two similar magnets spaced say half a foot apart, I infer from my book that on a direct path between the two surfaces of the magnets the field will be the same as on the surface of either, but this seems very wrong to me. I think the book may be assuming the surface area of the magnets is much greater than the length between them, and I will not have that sort of convenience. Am I just going to have to do calculus on each length of wire based on the surface shape of my magnets, or can I do something easier with the information given on websites that I'm looking at to purchase magnets from?

 

[waht]

The real thing will probably be a lot different then a simple voltage source,

small emf will generated in a one loop wire, wind it with lots of turns. The stronger the magnetic field, more emf will be induced.

Dependind on the configation, you might actually have AC current induced so you would have to rectify it will a diode bridge.

Otherwise you will have a pulsed DC voltage coming out your loop.

 

[vsage]

Sorry I should have been a little more specific on some things. I predict currently that I will get a sinusoidal voltage proportional to A$$\omega B_0$$ where $$B_0$$ is the "constant" field going across the loop, A is the area of my loop factoring in turns, and $$\omega$$ is the speed at which my shaft rotates, but my problem is that I do not understand how to derive $$B_0$$ from surface magnetic fields listed on websites from which to purchase magnets.

Most of the cheap magnets I can find (one in particular I was looking at was http://www.gaussboys.com/ndfeb-magnets/B1201.html ) don't have enough surface area though, which makes me think the calculations for an accurate induced emf will be difficult if they are being placed like 20cm apart.

 

[waht]

It's not very practical to calculate it all out. Even if you do you won't take the fringes into consideration due to even more difficulty.

Take some sample magnets, hook them up in your machine and measure the magnetic field. If you are in the ballpark range great, if not make some corrections. Or design an additional adjustable voltage compensation amp for the wire loop.

hope that helps

 

[vsage]

I figured as much. I suppose there are easier things to adjust beside the strength of a magnet. Thanks for your insights.

 

[es1]

Also, since you said 'cheap' magnet I'll bet the field strength is not well controlled meaning you'll get a solid amount of variation from magnet to magent. So I would just assume some minimum magnetic field (based on your measurements) and make it work for everything greater than that. In this case you don't care what the exact magnetic field is, only that it is greater than a reasonably safe minimum. So no need to calculate. :)

 

[marcusl]

If you sweep a coil through a constant magnetic field, you'll get a sinusoidal signal. You won't be able to "interpret the signal and detect when the signal gets too large" because it's frequency, not amplitude, that will change. Make sure your control circuit is frequency based.

You might get a hold of an "electronic" automotive distributor from the late 70's or 80's. A star-shaped soft iron wheel with 8 points, called a "reluctor," is attached to the shaft. A magnet and a multi-turn pickup loop are arranged near the points of the reluctor, separated by from it by a small gap. Usually the magnet is located inside the pickup housing. See Fig. 4 below
http://www.procarcare.com/icarumba/...urcecenter_encyclopedia_ignition.asp#Figure5"
There is no coupling except when a tooth sweeps around into alignment, coupling flux from the the magnet to the coil and giving a voltage spike. You'll get 8 pulses per revolution, without having to engineer the whole thing from scratch. You can get get the parts from aftermarket suppliers like MSD
http://www.msdignition.com/1distributors.htm"
(reluctors and pickups are at the bottom), or go to a junkyard and get a complete distributor for, say, a 1975 Chrysler V8, for probably $10.

Finally here's an industrial pickup for this application.
"www.woodward.com/pdf/ic/02010.pdf"[/URL]

 

[vsage]

Hello. I noticed your post two days ago marcusl and I have to say thank you for going a little tangential from my initial question :) I ended up redesigning my detection device completely and incorporating the reluctor/mpu to check speed. I originally had hoped I could discern the speed of my part from the amplitude of the signal I was getting (Faraday's Law and such) but decided ultimately that wouldn't work after seeing how simple reshaping those pulses I will get from the reluctor/mpu would be to turn into a digital signal that can be counted.

I do have a question along the same lines at a different point in my project, and after looking at several websites I am still confused: How do I excite a relay? Right now my idea is that I send one small square pulse to a relay's input and if my thinking is right, the relay will change output states from closed to open or from open to closed. Is this an accurate perception of any particular type of relay?

 

[NoTime]

I do have a question along the same lines at a different point in my project, and after looking at several websites I am still confused: How do I excite a relay? Right now my idea is that I send one small square pulse to a relay's input and if my thinking is right, the relay will change output states from closed to open or from open to closed. Is this an accurate perception of any particular type of relay?

There are stepper relays more or less like that, but normally with more states than just on/off.

However, the average relay requires constant power for closed.
Not a pulse.

You can set up a latching arangement.
Still requires a constant power input for relay closed, but the control trigers can be a pulse.

What are you trying to do with the relay.

 

[berkeman]

I thought I'd heard of latching relays before (with zero hold power required), but I've never used one myself. So I googled "latching relay" and one of the first hits is a great one from National Semiconductor:

http://zone.ni.com/devzone/cda/tut/p/id/3960