Understanding Bedini AOP Improvements

[Raccoon]

So first, let me tell you, I've actually build a basic SSG Bedini based on 3 phases floppy motor (It cost me the little amount of 0.50 €)
It is not accurate as I would want, and I need approximately 1 hour to recharge 10mV in a 6V lead acid battery.
Inside this motor, there is 3 sets of 2 coils each (12 coils star pattern). I separate each set of coils, and wire one set as a trigger and 2 sets in series as powering & BEMF.

So after being recommended by TinselKoala about the MH Comparator Driver, I want to share what I could make in practice (even if it don't work), and also would probably need your helpful recommendations :
• I watched all recommended videos from TinselKoala related to the MHO circuit.
• I read partially the Topic : Self accelerating reed switch magnet spinner (which include the MHO Circuit)
• I simulate in LTSpice a design near from MHO Circuit.

What I understand (If I'm not wrong), is that this Circuit is base on a voltage comparator, which enlarge or shorten triggering impulses from the sense coil (name by Bedini the trigger coil).
I get the point that this setup would probably allow more controls about the duration timing for the powering coil, and also this will induce more Back EMF in return (this is what we want).

After reading many times good advices by MileHigh (also don't forget it's is idea), I finally understand why there's an offset on the voltage comparator.
Tell me if I'm wrong:
• Wiring the trigger coil to the mass will cause lenz effect to the rotor and should probably slowing it down.
• Maybe this will prevent unwanted triggering depending where the trigger coil is placed regarding the powering coil.

So for my case, I should mention that I have an extra coil facing the bell rotor, which could be used as a RPM sensor.
(see the little green resistor package near the star coil, it is mentioned ind for inductance !)

I've made a simulation with personal parameters :
• For the moment I only use 5V power supply
• I have bunch of LM358N

The zip file include all files needed for simulating in LTSpice.
I will let you know if this works better than the SSG with reals experiments (for my configuration).

[MileHigh]

Raccoon:

• Wiring the trigger coil to the mass will cause lenz effect to the rotor and should probably slowing it down.
• Maybe this will prevent unwanted triggering depending where the trigger coil is placed regarding the powering coil.

Both of your statements are correct.  Your simulations also look great.  You can see how you can control the pulse width by changing the offset voltage.

Good luck with your motor!

MileHigh

[TinselKoala]

Looks great, I'm glad you decided to try it! You are getting plenty of HV return too, it looks like, even though I might not have expected it from the coil arrangement you are using. Good work!

OK, so you have control over the pulse width very accurately. Now for the timing... I think that in your case you could put a 555 configured as a one-shot monostable pulser between the op-amp and the mosfet gate where you have the 270 R (why so big? 10R might be better) . You can then control the delay (timing) from the 555. For mine I simply move the sense coil to change the timing point but it would be more elegant to do it electronically with a 555 I think. You might also want to put a 1 megohm resistor from Gate to Source; this will help the mosfet to turn off cleanly.

The pulse from the comparator is already your rpm sensor ! Even when the drive power is off you will still get a trigger pulse from the sense coil and you can translate the frequency of the pulse to RPM because you know the number of pulses per revolution.

[Raccoon]

So thanks MileHigh, it help me go further.

Yep the 555 acting as a timer, I already thinking about that.

I also have fews others ideas, but don't know if it will improve or not: using one or two 555 as a timer + amount of time control (a sort of PWM control). Maybe this will be redundant with the voltage comparator (in this case it will probably not be necessary). It's just a raw random idea, need to try to know if it could be more accurate.

(why so big? 10R might be better)

I've just choose an over-sized value, I should tweak depending the power dissipation limitation an what's would be really after.

So now I try to get triggers from this little inductance (and this let me have the ability to power the full star coils).

Edit: Yep I get trigger with this green inductance thanks to the voltage comparator.
With a phase coil, I get absolute 8 or 16 impulses (rotor have a ring magnesium magnet with 8 north poles and 8 south poles)

[Raccoon]

Some Feedbacks.
Last evening, I found many possibilities to run my motor, but I couldn't do it from the "real" green ind RPM sensor :

• With N-MOS, the RPM go straight higher than "SSG", I must do it with 1 phase coil trigger and 1 phase coil power.
I get from the floating third phase coil some energy BEMF to burning leds (I have really burn one !). Also I don't know, but it seems it couldn't get any radiant energy (I've tested everywhere), I couldn't light the neon or charge cap, ect... When stopping, the NMOS fire hot energy (I prevent it by stopping), when running it get just a little hot.

• With a simple NPN, I could do the same than with NMOS, but I get straight high power into the neon.
It's far more than with the SSG Bedini. In fact I use a cheap small transistor (which probably don't have any diodes protections), it get really better than other, but really hotter. I try reduce the conduction with larger base resistor, but this kill any run. Also, the motor rotate really less than with the NMOS. With 2 phases coils in series, the neon sparks a lots of purples and blues sparks (just inside)!

After burning my finger tip (with touching the transistor), I find a little trick to know when it get fire.
With a little water syringe, I leave a drop on top of the transistor. When it get dissipated, I just shut-off the power.

So after many practices, it seems for me that there's a difference between BEMF and "high BEMF' which could be name radiant energy. The most effective thing at the moment is a simple SSG with a NPN running just a little hot.

Know, I want to try different possibilities :
• Cooling far more the NPN drive
• Getting neon light up with NMOS
• Getting running with real RPM trigger, thanks to the NE555
• Trying to run all of that with just an simple Oscillator (frequency adjustement, PWM adjustement)

Also I'm thinking about doing some sort of self starting the rotor. (Sometimes with the NMOS it get itself!)