Self accelerating reed switch magnet spinner.

[TinselKoala]

The ratiometric Hall sensor works almost like an "ideal" coil in its response to the magnetic field, but it operates on the strength itself not the rate of change of the strength like the coil does.

So you have an increasing voltage as the rotor magnet approaches the sensor, peak voltage at TDC and decreasing, but still positive, voltage as the magnet recedes. So the ordinary comparator just turns on at a certain field strength == distance from TDC and this doesn't change with speed, then turns off at the symmetrical distance after passage of TDC. If you want to turn off at a different strength, or at TDC (max field, max output from sensor) you use a window comparator, a tiny bit more complicated but still within the single chip dual opamp TL082's capability.

http://www.allegromicro.com/en/Products/Magnetic-Linear-And-Angular-Position-Sensor-ICs/Linear-Position-Sensor-ICs/A1324-5-6.aspx

[MileHigh]

TK:

That newfangled hardware!  You kids these days.

I will make some comments rooted back in the Old School analog days.  Also a disclaimer:  I am not trying to get you to build anything, this is just fun stuff to talk about.

Going back to your setup, you are using a diode to protect the op-amp input when the pickup coil swings below ground.  I suppose that you have a series resistor in line also?  Hard shorting of the pick-up coil output through a diode would cause Lenz drag on the rotor.  I still prefer the idea of "floating" one end of the coil at +6 volts.  That gives you the option to set your comparator voltage lower so that you can dial up an "extra wide" pulse that starts firing before top-dead-center (assuming that you are not offsetting the moveable pickup coil from the "zero angle" position).  Kind of like a dwell timing thing.  Something tells me you are probably an expert on that subject.

If the "ground" end of the coil was set to "float" at +6 volts, adding a series resistor and a set of clipping protection diodes would be very wise and I forgot that.  No point in blowing your op-amp input like a newbie.

Going back to a conventional Bedini motor, if you had one that chopped the switching signal because of the drive coil to pickup coil coupling, I would want to scope the pickup coil output waveform and examine it closely.  You seem to be indicating that the transitions would be sharp, but I am not convinced that things would "snap" like in a Joule Thief circuit.

Now here is something from the analog trick bag:  When you are looking at the pickup coil voltage waveform you will see a waveform generated from the passing magnet and the drive coil coupling.  But you still won't exactly know what kind of switching the main drive coil is experiencing because you can't reliably "see" on the output side of the transistor because the drive coil itself affects what you will see.

Supposing your DC resistance of your drive coil is 20 ohms.  Suppose also that adding some extra loading to the pickup coil output is not going to alter things drastically (to be verified on your scope.)   So you connect the pickup coil to a second base input resistor of the same value as on the motor, connected to a second identical transistor that's driving a 20-ohm purely resistive load.  In other words you copy the coil drive circuit and substitute the coil with an ordinary resistor.  When you look at the voltage waveform at the junction of the collector and the 20 ohm resistor you will be able to see how fast and clean the switching is for the drive coil.  Doing this lets you see how the transistor itself is switching on the output side.  (Note there is no need to do this when you use the op-amp comparator output - you know the transistor switching waveform will be fast and clean.)

Anyway, I know that sounds a bit crazy, but half the fun is supposed to be in the investigation.   My gut feel is telling me that when you get this oscillation phase when the drive coil is being energized, it's not necessarily with a high slew rate waveform.  Hence, you are burning off power in the transistor, which is undesirable.  It may not even be significant, I am just playing a virtual game - the challenge being to upgrade a conventional dumb Bedini motor setup so that you have more control and the switching is always fast and clean.  And do it all with analog components and do it on the cheap.

There was a time just a short while ago in the overall scheme of things when there were no microcontrollers running nearly everything and we got along just fine.  Analog electrical and analog mechanical computers and machines did just fine and nobody knew any better.  Today, it would almost be the end of the world if suddenly every microcontroller stopped functioning.

MileHigh

[MileHigh]

TK:

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.

I really don't think that would be an issue as long as you can do a reasonable rotor build were the four magnets are indeed 90 degrees apart.  I also recognize it depends on the particular type of magnet and the magnet supplier.  However, it seems reasonable that a good magnet supplier would produce cylinder magnets that are quite consistent from one to the other.  Especially if they are from the same batch.

Since the "jitter" would be regular and repeat every four ticks, I don't see this as causing any observable issues.  Suppose you have a conventional Bedini setup and the pulse fires at +2 degrees past top dead center for all four ticks.  Then you have a modified comparator-based setup and the the pulse fires at -2, 0, +3, and -1 degrees from TDC for every rotation.  I just can't see that being an issue because of its consistent repeatability and the approximately same torque "slow impulse" applied to the rotor for every rotor magnet passing.

The big deal is you can all of a sudden start playing with the pulse timing like never before.  You are free from the "chains" of having the pickup coil being coaxial with the drive coil.

MileHigh

[TinselKoala]

Well, OK, so I tied the end of the coil to 1/2 Vcc thru a simple divider of 2 x 22K resistors across the supply. Now it works much better, in terms of trigger sensitivity and pulse broadness, just as you said!


Thanks for the suggestion. I'm no op-amp expert that's for sure... but I did find a copy of the IC Op-Amp Cookbook by Walter G. Jung in the bookstore a couple months ago and I've been studying up....

Next will be to see if I can actually implement the physical arrangement to get a motor turning. I may need to wind some coils, and cut some wood...

[TinselKoala]

"The big deal is you can all of a sudden start playing with the pulse timing like never before."

Well... no, not exactly like never before, not for me anyway. I commonly use Hall sensors in variable position mounts, bias them with magnets, etc. The Hall sensors I use can switch a mosfet directly so the circuits are ridiculously simple... as long as you don't look inside the Hall package, he he.