Usefulness of lower-impedance, faster-response electromagnet

[Eukonidor]

Given an electromagnet which has a drastically lower impedance than other, comparable electromagnets (same diameter, same gauge wire, same field strength), and, consequently, a much faster response time, what uses do you see this being good for?

I'm looking to put together a patent application, and would like some more suggestions for applications to which it could be put.

I have tested the response time: My test coil, made from speaker wire, went from 0 to full current in under 20 ms, while the reference coil (from the same batch of wire) took nearly 200 ms.

 

[Baluncore]

I'm looking to put together a patent application,

I think it is a bit early to consider a patent. There is much prior art in that field.

The current flow in an inductor rises asymptotically towards a theoretical maximum.
You need to time the rise to ( 1 – 1/e ) = 63.2% of theoretical maximum.
(Theoretical maximum current) equals (applied voltage) divided by (coil resistance). I = V / R.

It might be good to first analyse why there is a difference in rise time.
What length of wire and how many turns did you have on the two coils?
What core did you use ? Did they both use identical cores ?

 

[Eukonidor]

I know why the rise time is different; it's part of what I'm working on. Please excuse me if I seem unwilling to share too many details. I'm not highly trusting of the general populace, and have been frustrated several times in the past, though I admit I have yet to have anything stolen, just beaten to the punch.

However, to answer part of your questions, the total length of wire and number of turns are identical (I believe it was 100 total turns, though I don't remember for certain), and both magnets shared the same steel rod core.

There will need to be more tests, surely; for one thing, while I suspect that longer coils (not the wire length, necessarily) will benefit more as far as response time is concerned, I would like to test that. I also need to know how the overall power draw compares to normal coils. If I can get hold of a good signal generator, so that the impedance can be correctly calculated, I will see if I can develop a relationship between them.

 

[jim hardy]

same core ?

iron is iron. Well, not really. Check into Allegheny 430 solenoid grade.

 

[Eukonidor]

It's what I had available. That also means it was same diameter, etc.

 

[Baluncore]

As a Devil's advocate, I know that there has been an enormous amount of work done optimising electromagnets over the last century. The theory is now well defined. While it is possible that you have discovered something new, it is statistically quite unlikely that it will be novel and patentable.

The problem now is how we can identify what actually causes the rise time difference, without you telling us what you think improves the current rise time.

Maybe the magnetic path is different in your two test measurements. If the air gap is slightly different, then you will get a different inductance and hence a different current rise time to 63%.

There is no question that faster rise times are better in most situations. Rise time is greatest when a high voltage can be provided at turn-on. Rise time is fastest when inductance is lowest, so wind less turns, use a lower voltage but with a proportionally higher current to get the same (ampere * turns).

Since inductance is proportional to the square of the number of turns, to get a rise time that is 9 times faster you must reduce the number of turns by a factor of only 3, but increase the current by a factor of three to maintain the magnetic field.

 

[Eukonidor]

As a Devil's advocate, I know that there has been an enormous amount of work done optimising electromagnets over the last century. The theory is now well defined. While it is possible that you have discovered something new, it is statistically quite unlikely that it will be novel and patentable.

You're right, and I've been thinking about that. It's a ridiculously simple thing, but I've never seen or heard of this being done, but maybe someone here has, and can point it out, because I also cannot find a reference to it anywhere, despite some fairly comprehensive searching.

Basically, all it is, is a set of flat coils around a common core and wire them in parallel. In the test I performed, I made 5 disks, with the wire spiraling into the center and back out again (or, if you prefer, it spiraled out from the center in opposite directions). To test the two wrapping methods, I simply connected them in series and tested, then connected them in parallel and tested again. I also did measurements of individual coils, as well as two, three, and four in both configurations. The graph below is the responses of all 5 disks connected. The top is the series hookup, and the parallel is the bottom:

https://www.dropbox.com/s/bb8xz66zyxeytz1/MagnetReponse.png

 

[jim hardy]

inductance is flux X turns divided by amperes flowing through all those turns..

All of the current flows through every single turn in your series connection?

You haven't said how the current divides among the turns in your parallel connection.

What happens when you approximate your inductance from your observed L/R time constants ?

Baluncore pointed out inductance is in proportion to turns² . Do your observed data agree?

 

[Eukonidor]

Hmm...

1. Yes, each coil was wired to the one adjacent to it, output of one to input of the next.

2. I'll have to go back and check that. I was testing voltage drop across the input and output with a cheap oscilloscope set to capture single events. I'm assuming that you are asking if the current through each coil in the parallel setup was the same or different. If properly constructed, the current should be essentially equal for each one.

3. Since this is based on current, I'll have to check the current at full power. I have nothing to measure current over time, however.

4. Basing it on the course at this page: http://www.wisc-online.com/objects/ViewObject.aspx?ID=DCE10304, which says it requires 5 time constants to reach the full DC current value, I get a tau of 2.4ms for the parallel setup, and 40ms, or 16.7 times the parallel number, for the series, as near as I can tell from my measurements. This is reasonably close to the ratio of 25 times, and is probably within the margin of error for my setup.

 

[f95toli]

There has been a LOT of work done on coils used in pulsing for ESR and NMR/MRI experiments; since the idea in these experiments is to "pulse" the magnetic field it follows that the the setups have to be able to generate large fields with very small rise-times (ideally microseconds or less).

Hence, before you file your patent it might be a good idea to first look at the state-of-the-art for magnets made for pulse applications; it is pretty unlikely that you've discovered something that has not been tried before.

 

[Eukonidor]

Got a link, or anything? I've BEEN looking.

 

[jim hardy]

Hmm...

1. Yes, each coil was wired to the one adjacent to it, output of one to input of the next.

2. I'll have to go back and check that. I was testing voltage drop across the input and output with a cheap oscilloscope set to capture single events. I'm assuming that you are asking if the current through each coil in the parallel setup was the same or different. If properly constructed, the current should be essentially equal for each one.

3. Since this is based on current, I'll have to check the current at full power. I have nothing to measure current over time, however.

4. Basing it on the course at this page: http://www.wisc-online.com/objects/ViewObject.aspx?ID=DCE10304, which says it requires 5 time constants to reach the full DC current value, I get a tau of 2.4ms for the parallel setup, and 40ms, or 16.7 times the parallel number, for the series, as near as I can tell from my measurements. This is reasonably close to the ratio of 25 times, and is probably within the margin of error for my setup.

Thanks ! That helps understand what you're up to...
I really like to tinker with inductance. We used it to measure position of control rods in the plant, basically a 12 foot long electromagnet with the upper part of the rod drive shaft for a variable length iron core. As the rod moves upward, more of the drive shaft moves up into the coil so an increasing fraction of its magnetic path length becomes iron and inductance increases. Indeed it about doubles at rods fully out.
But that's a digression.........

Okay now think about your L/R measurement:

you haven't said whether there's any resistance other than your wires in your circuit.
For thinking-out-loud purposes, let's assume for a minute there's not, ,,,,
and that current divides equally among coils,
and all your turns count in the inductance equation,
and that a single coil has resistance r.
So.. thinking out loud:
The resistance of all your coils in series is 5r
and in parallel it's r/5
aha ! Same 25 ratio as for turns in inductance equation.
How to separate those effects?
...

is there perhaps some other resistance, maybe in your current measuring circuit, that gets you to that ratio of 16.7 instead of 25 ?

It's really cool that you have the interest to do this.

 

[Eukonidor]

I plead primitive testing conditions. No other resistance yet, so I will be adding that for next round.

 

[jim hardy]

we learn best by doing .