effects of switching polarity on an electromagnet

Graham, we're still looking for an answer to this question. What are you working on? Maybe we can come up with an estimate.

Yep, you've got me pegged, Win. I'm a first-year grad student in a music group at the Media Lab. I've actually got more of a comp. sci. background, but am just starting to get my feet wet in the fascinating (and eminently useful) world of electronics. I realize that many of my questions are probably pretty naive and may seem obvious to the regulars here on s.e.d and so I particularly appreciate everyone's help and patience.

-Graham

Sorry for the slow reply; we've just started Spring semester and I'm still trying to wrap up a couple of projects from break...

This is actually for a force-feedback drum project that I've been thinking about recently. The idea is to seat the magnet just behind the head of a drum, pulse the field direction rhythmically such that it attracts and repels a permanent magnet embedded in a drum stick tip. The original motivation was purely creative: see if this arrangement could be used to produce a sort of 'player drum' effect (e.g. you strike the drum once, but at the same time the magnet is used to rapidly attact and then repel the permanent magnet in the drum stick tip, say in the pattern of a fast triplet). Lately, I've been thinking that, if I can get this thing to work, it would be interesting to look at pedagogical applications. If you could 'quantize' a beginning player's sloppy timing you might affect their learning rate by giving them a feel for the correctly played rhythm. There are also some musical concepts such as 'playing behind the beat' that just about impossible for a teacher to convey verbally, so 'learning by feel' might be a real advantage in some situations. In any case, this is obviously all very early stage stuff.

The short of it is that, while I'm not sure, I don't expect the mechanical load to be significant (unless, of course, someone *really* took a swing at the drum!). I do know the maximum pull of the permanent magnets (14.2lbs), but this would seem to be largely overwhelmed by the added force of the player. Does this seem like a reasonable assumption?

Thanks, Graham

Very interesting, and highly relevant to your design. I was very concerned over the issue of coil heating, because 192 watts (8A*24V) will overheat and destroy a small coil like yours in the end. You may not notice trouble during short tests as the heat is taken by the copper's thermal mass, but the equilibrium temperature will get you. I envisioned that you had a steady maximum field and suddenly wanted to collapse the field and then reverse it. But if what you need is short, fast musically-timed field pulses, the average power dissipation could be acceptable. That's good.

Second, your need for as fast as a 100ms repetition rate may be for only a few cycles at a time, good, and you may be able to live with slightly reduced force pulses after the first pulse, which we saw in Tony's simulations of fast reversals.

I'm wondering if you do want a reversal. The shape of the field pulse must be a strong factor in the drummer's sensation, surely in normal use you don't want to interfere too strongly with the drum stick's natural bounce off the surface. That must be why you want to collapse the field quickly, to stop the attractive forces by the time the stick hits the surface. But do you also want a repulsive force? Perhaps a weakened one, to gain stick distance for the fast triplet beat you're thinking about, etc?

I also wonder if you want a standing full-strength field before the drum-beat event? What's the time-frame you envision for the onset of your attractive field? The natural t = L/R = 53ms time constant from simply switching your 24V supply seems pretty slow.

You can push single powerful half-sine uni-polar pulses of current into an electromagnet with a charged bipolar capacitor and an SCR.

. half-sine positive current pulse . +HV supply ---/\\/\\------, . + C1 C2 + | --- coil --- t = pi sqrt L*C . ,---|(---+---)|---+ 160mH 3-ohms . ----+---||---+--#####--/\\/\\--, Ip = V sqrt C/L . gnd |______________|/|________________| . trigger__/|\\| scr

The bipolar capacitor came up earlier in this thread. Since the value you'll need is high, it's made up from two electrolytics. As the voltage on initially-charged capacitor C1 approaches zero halfway through the pulse, the field current reaches its peak. The coil maintains a declining current flow in the second half of the pulse, charging C2. When the current reaches zero the SCR shuts off. C1 and C2 are both equal to C in the equations.

After the pulse you still have most of the energy, but it's in the second cap, and the full bipolar capacitor has a reversed voltage (diodes protect the electrolytics during the voltage reversal). This is another place a high-voltage H-bridge may be useful, to recover and reuse the energy.

To make a single-cycle bipolar current pulse, replace the SCR with a TRIAC, and provide gate drive through the first half of the cycle, plus a bit, then remove it so the TRIAC will turn off after the second half finishes. A significant portion of the charge will be back in C1, ready for the next pulse.

If you make C2 larger than C1, the attack will be fast but the middle portion of the cycle will be longer than the attack. Or if you make C2 smaller than C1, the current cutoff and reversal will be faster than the attack and decay, which might be better for your drumsticks. C2 will need a higher voltage rating.

BTW, you can get TRIACs with current ratings to 6000A. :-)

This must be a very loosely-coupled magnetic system, so the electromagnet's inductive Q will not be mechanically reduced.