Hi John, welcome to CR4! It sounds like you spent a decent amount of time trying to reverse engineer Tinkerbell's engine. I've never seen that toy but here's what I think is going on beneath the epoxy. There's a circuit inside known as an oscillator or pulse generator that is constantly switching the voltage to the coil, either on-off-on-off or plus-minus-plus-minus as long as the capacitor that is across the solar cell is being charged by the light. There's probably a small printed circuit board that has a few more components on it whose job it is to tell a chip known as a 555 timer or its equivalent how high and how wide to make the pulse that goes to the coil. It's cheap and well known:

http://en.wikipedia.org/wiki/555_timer_IC

I hope this saves you from trying to remove the epoxy!

RAM, thank you for the response. I thought that might be what was happening and maybe it is, but the 555 and any support RC would have to all fit under an epoxy blob that is 3mm in diameter and 1mm tall. The entire PCB is a 1/2" by 1/2". And that's it. Do you think there is a tiny 555 under there with more caps and resistors? Or is there some other way to accomplish this. The PCB is quite thin and shining light thru it shows the four mentioned leads going in and nothing else. Thanks again, but I have to believe there is som very small 4 lead chip in there, no? John

Yes, they actually sell a chip that can be surface mounted under a blob like you describe, takes up very little room with no costly socketing, plated through holes, etc., just the chip flow-soldered to a small board and 4 printed circuit holes.

Here is a link to some pics that might make it clearer. I agree with RAM, that if a 555 based oscillator were under that epoxy that would do it. I am looking for a way to do it now with discrete parts so as not to have to use an IC. Any ideas? Thanks, John Photos: http://gallery.me.com/jkostelac/100071

John,

This toy is using a technology known as "COB", or "Chip-On-Board", where the IC die is mounted to this small PCB, then wire-bonds are added that attach the IC die to the board, then the whole mess is encapsulated in the epoxy "blob". Trying to figure out the exact details of this chip is going to be next to impossible. There's a very good chance the IC die in question is purpose-built for the job it's doing, and it's not a standard "off the shelf" IC design such as the common 555 timer IC. It can be expensive to develop a purpose-built IC die such as this, but when you going to divide that cost across millions of "toys", the price-pre-unit comes down considerably.

As far as developing a circuit to do the same thing, it shouldn't be too hard to do. I would start by Goggling "coil driver" or something like that. You basically need to deal with two functions; one is the oscillator that controls the speed of the "flapping", and the other is the "driver", that will allow you to drive enough current through the coil to make the magnets move. Your driver will also have to deal with the collapsing field of the coil as it switches from one voltage extreme to the other. The voltage spike that is generated when a coil is switched can destroy the driver if not dealt with.

I hope that helps, I know it doesn't address your desire to come up with an equivalent circuit, but I hope it gets you on the right path.

Tom D.

Tom, Thank you for your explanation. It also explain some of what's going on. One question that has not been addressed, but was not specifically asked is what is going in in the coil as the IC sends a current pulse? More soecifically, there ate two magnets that pass over the coil just a hair's breadth from it in an single left or right strike. The two magnets are mounted with flat faces parallel to the flat surface of the coil AND they are mounted oppositely. That is, one magnet has South facing the coil while the other has North facing the coil. They are side by side in line with the swing across the coil. So, the IC sends a pulse that repels the first magnet while attracting the second, but also the magnets are producing back-EMF as they pass over the coil, no? How could this info be used to re-engineer what's going on in the blob? Or is the back-EMF just current limiting as in any motor? I find it amazing that so few folks are curious about how and why their toys or tools do what they do. Thanks, John

Don't underestimate "folks" curiosity; nearly all those on the forum are the type that look at something, and think to themselves, "I wonder how that works?" I tore apart many "toys" in my youth in search of the answers . As I got older, I was even able to put them back together, and sometimes, they still worked .

I did a quick search for "electric pendulum circuit", and the first site had what is probably the simplest circuit to accomplish the feat. They use a separate winding to "sense" when the magnet passes, and a single transistor drives the coil. There's a pretty good theory that accompanies the circuit, as well as some captured waveforms that do a good job of showing what's going on with the circuit. Check it out.

http://www.sciencetronics.com/geocities/electronics/projects/pendulum.html

Hope that helps.

Tom D.