Self accelerating reed switch magnet spinner.

TinselKoala · October 03, 2013, 01:33:05 AM

I don't think there's time for this year's contest, and I'm wrapped up in something else right now anyway. But when I have a little time and benchspace I'll see if MH's idea can be made to work, using the rotor/axle assembly from my Bedini SGM.
I especially like the "autostrobe" idea. I am using something like that in my Arduino Pulse Motor driver program, but that's still a work in progress and it's simmering on a back burner right now.

The site www.fasttech.com has all kinds of breakout sensor modules for Arduino, and I think I saw a couple of reed switch breakout boards listed there.

MileHigh · October 03, 2013, 12:39:40 PM

TK:

Just for fun... You take it to the next level. lol You have your favourite 555 oscillator. Then we go high tech and get a quand NAND gate chip, the 74XXX00. The most useful gate chip of all!

De Morgan's law, very important: "not (A and B)" is the same as "(not A) or (not B)"

http://en.wikipedia.org/wiki/NAND_gate

If you were hard core you could build a gate with transistors, the schematic is right there!

You actually want an AND gate. You do this: Output = [Comparator] AND [555_timer]

The output from the AND gate drives your IGBT/MOSFET/transistor. So now when the output of the comparator is TRUE the coil driver is being switched on and off by the 555 timer signal. This takes the "slicing" of the energizing of the coil to the next level. You can slice and dice to your heart's content and control the average amount of current that goes into the coil before you collect the back spikes. It's arguable that this allows you to reduce the current into the coil to reduce your resistive losses in the coil itself and collect a purer form of "radiant energy."

MileHigh

TinselKoala · October 03, 2013, 02:38:49 PM

Uhhh... OK..... but for that degree of complexity I'd probably go with the Arduino for its programmability. The idea of chopping the signal could be implemented by using one of the Arduino's PWM outputs, and gating that thru the external mosfet with another of the Arduino's digital outputs.

Meanwhile I implemented your comparator idea with a basic circuit using a TL082 opamp and a single supply. I can trigger the comparator without difficulty using a magnet swung past a coil.... but I have already identified two issues.

First... of course the voltage generated by the sense coil will depend on the speed of the magnet's passage. This means the comparator's setpoint will determine the speed of the rotor necessary to even start firing, and as the rotor speed increases the pulse width will actually increase, since the "window" of firing voltages will occur over a wider space during the magnet approach to the coil.

Second... the coil I used had to have a core. I couldn't get it to fire easily with no core. So the rotor magnets will be interacting with the core of the sense coil, which, as I understand your idea, would be mounted in a different place than the drive coil(s). This will cause mechanical problems, I think.

Now, using a ratiometric Hall sensor instead of the sense coil would allow the comparator to flip and fire the mosfet at a given magnetic field strength, which would eliminate both of the above problems I think. The firing point (translating to dwell or "on" time, I think) could still be adjusted with the level control on the comparator, and the precise timing could be done with sensor positioning as you envisioned. (That's the way I timed my Marinov Slab and my Orbo replications, by adjusting the position of the Hall sensors.)

MileHigh · October 03, 2013, 04:01:35 PM

TK:

It might be more fun to do it without a microcontroller for some people. Can you change both the frequency and the duty cycle of an Arduino PWM output without resorting to any software tricks?

QuoteFirst... of course the voltage generated by the sense coil will depend on the speed of the magnet's passage. This means the comparator's setpoint will determine the speed of the rotor necessary to even start firing, and as the rotor speed increases the pulse width will actually increase, since the "window" of firing voltages will occur over a wider space during the magnet approach to the coil.

Note that a conventional Bedini motor setup also exhibits the same behaviour where the relative pulse ON time per rotation increases as the rotor speed increases. With the comparator circuit addition you have the luxury of slicing the positive hump output from the coil at any potential you want, allowing you more control over the pulse width.

QuoteSecond... the coil I used had to have a core. I couldn't get it to fire easily with no core. So the rotor magnets will be interacting with the core of the sense coil, which, as I understand your idea, would be mounted in a different place than the drive coil(s). This will cause mechanical problems, I think.

In theory the comparator gives you more sensitivity. Did you tie the opposite end of the pick-up coil to +6 volts? If the "ground" for the coil is +6.0 volts and you set your comparator threshold to +6.1 volts then it should be very sensitive.

I envision a pick-up coil that is mounted on a right-angled stand made with wood or something. The number of turns and the geometry of the coil and core or no-core give you a lot of options. The "problem" with a typical Bedini setup with a single drive-pickup coil is that the receding rotor magnet induces positive EMF into the pick-up coil which switches on the transistor. However, when the transistor switches on that induces negative EMF in the pickup coil which switches the transistor off again. So you have an undesirable (in my opinion) negative feedback oscillator effect taking place during the firing pulse. "Negative" in the sense that I would rather have complete control - a full pulse or a chopped pulse of my own choosing and with guaranteed clean switching. Depending on the particular Bedini motor sometimes you see the self-oscillation, sometimes you don't. Sometimes you see it but then it disappears as the rotor speeds up and the "regular" pulse width decreases.

When you get your reference timing signal that is 90 degrees displaced from the main driving coil, the magnetic flux from the drive coil does not cut the pickup coil very much. In fact, just with a scope and putting a sine wave into the main drive coil, you could tweak the angle of the pickup coil so that there is near-total self cancellation in the small amount of flux that passes through the 90 degree offset pickup coil. You can imagine the toroidal magnetic field pattern generated by the drive coil and adjust the pick-up coil angle such that it has a low cross sectional profile with near full flux-self cancellation relative to the curving magnetic field pattern of the drive coil.

MileHigh

TinselKoala · October 03, 2013, 07:02:27 PM

The setup I'm using has one end of the coil tied to ground and the other to the noninverting input of one amp in the TL082 through a 1n914 diode. The inverting input gets the signal from the wiper of the 10-turn, 10k trimpot, whose legs are at positive rail and ground. I'm using a single supply, just Vcc and 0. For the test load I'm using an LED and a dropping resistor. So with a blue LED, I'm using 1.8k and about 10-12 v Vcc for a brilliant LED when the comparator flips and sends the supply voltage to the output. The FET input of the op amp is so sensitive I can flip the comparator by touching a little "antenna" on the Pin 3 noninverting input, same place as the diode from the coil goes. The 10k trimmer provides precise adjustment.

But I still see the issues I raised as problems. The Bedini motor, as you point out, is self-quenching, so firing the coil at the timing point and then recirculating the shutoff spike provides a sharp "bang" to the rotor with a narrow duty cycle. This will not be the case with the speeding rotor magnets unless the comparator has extra circuitry. The comparator will stay "flipped" until the magnet starts receding again and the sign of the induced voltage changes (and is clipped by the diode.) So for increasing speed it will flip earlier and earlier. I think it will still turn off at the same place in the cycle though. So maybe this is analogous to the Bedini self-quenching. I'll just have to go on to a whole build to find out, I guess.

Two words: ratiometric Hall effect sensor. Wait, that's four words. But for simplicity and immunity they can't be beat. Allegro Microsystems rules.

ETA: I think the PWM of the Arduino normally has a fixed frequency and the user varies the duty cycle. But I think that the frequency can also be changed, by Real Programmers. Know any?

ETA2: there is another problem I just recognized. If your sense coil is 90 degrees around from the pulse coil, (4-magnet rotor) then you are triggering on a different magnet than you are driving. This means your magnet positions have to be precisely equal around the rotor, or you will have a weird repeating jitter pattern. A big advantage of the Bedini system and a Hall trigger system is that you can trigger on the same magnet you are going to drive with the pulse, so you don't have to worry about precise magnet positioning around the rotor. My Bedini motor uses a peanut-butter jar lid as the rotor and the magnet positions are ...er.... let's just say they are not to "MYLOW" levels of a hundredth of a millimeter precision.