Self accelerating reed switch magnet spinner.

[TinselKoala]

It's up to about 1600 RPM on 24 volts to the drive coil. The switching circuit is drawing 17 mA from the 12 volt supply, including the LED strobe. If I disconnect the LED the switching circuit's draw goes to about 4 mA.

I found that the drive coil works pretty well when it's strapped to one of the frame supports. In this position, the comparator potentiometer varies the duty cycle through the whole range from all on to all off. It looks like about 60 percent HI is optimum. The cap across the coil (instead of a diode) produces an interesting ringing waveform during the off time. A diode, in addition to the cap, has a big effect on this ringdown but doesn't seem to affect the rotor RPM much. A diode across the coil was a great help for the Magnetic Levitator apparatus, though, so I'm not sure which way to go.

Without the capacitor from drain to source, the comparator behaved badly. I thought it was field from the coil causing feedback, and I wasted an hour fooling around trying to get clean pulsing with the drive coil connected. Then I just touched the circuit and noticed that it started behaving better when I touched in certain places. My body's capacitance was helping, especially when I touched both drain and source of the mosfet. So... grab a random small capacitor.... and it worked beautifully!  A 10 nF , 80 V poly film cap.

[MileHigh]

TK:

Off to the races.  I am assuming that with your scope and the judicious placement of one or more tiny decoupling caps you will make the pulse firing rock solid.

If you assume as an example that firing the drive coil between 5 and 10 degrees past TDC imparts the most torque on the rotor, then by adjusting the pot and the pick-up coil angle you can position the energizing pulse to fire between 5 and 10 degrees.  To be more specific, you probably need a kind of "dwell angle" advanced timing for the energizing pulse to get some current flow through the coil ahead of time to give the push at the optimum angle.  It all depends on how short or long the L/R time constant is for the drive coil.  So by observing that on a scope you can get a sense of whether or not you need to move the pick-up coil angle to create the advanced dwell angle.  Of course, you have the scope trigger from the comparator output so seeing the current rise in the coil relative to the firing pulse is easy as pie.

Then with one multimeter showing you the current consumption and a tach showing you the RPM, and the flashing LED showing you the live pulse angle, you can play with the pot and the pick-up coil angle and pickup coil radial distance to position the energizing pulse right at the sweet spot.  You may want to maximize the RPM or you may want to try to get maximum RPM per watt of input.

MileHigh

[TinselKoala]

Current version:

[TinselKoala]

But...but.... what about all that Bedini action, HV spikes and neons and cap charging and all of that action? This motor is like from Dullsville, man... or is it?

I've now added a NE-2 neon across the 10 nF capacitor from Drain to Source of the mosfet. With S1 closed, bypassing the 1n914 diode, I get extreme spikes of over 250 volts from the drive coil, and these can be utilized in the usual ways, like lighting the neon or charging external caps/batts/whatever. In this mode the rotor actually accelerates even more, and I can trim the potentiometer setting for a bit more ON time in the duty cycle and get substantially more RPM than with S1 open.

Speed right now is 2140 RPM. The scope trace is at 2 ms/div horizontally and 50 V/div vertically, and there are 4 pulses per revolution, so the frequency in Hz shown on the scope x 15 = RPM.

[TinselKoala]

Last night when I made the last few posts I was pretty tired. The capacitor is as shown in the schematic, not "across the coil" but rather between Drain and Source of the mosfet, i.e. across the whole power section.

Putting a recirculation diode in the usual position across the coil completely stops all the spike action and makes the Drain signal a clean square wave. But this also decreases the speed of the rotor.

I don't have the inspiration right now to do much more solid mechanical build, to find the real optimum coil positioning and mounting. The comparator set-pot works perfectly to control the duty cycle through the whole range, although the setting is a bit narrow. It might be good to use a smaller value pot here for more spread on the settings.

The thing is running "hot straight and normal" right now, very quiet, brilliant NE2 and about 2150 RPM. I haven't looked at the current draw from the 24v run batteries yet, but as I said before, the switching section uses very little power, especially if I kill the LED.

The LED autostrobe isn't really satisfactory due to the jitter and the fact that it fires on every magnet. It would be better to fire it only once per revolution of the rotor, or have a rotor magnet placement that was precise enough to not jitter. I may try a one-magnet rotor just for grins.... then the strobe would work properly.