Self accelerating reed switch magnet spinner.

[MileHigh]

I have to do a new revision of the design because I am still really bugged by this headroom business, but I want it to still be a dual op-amp design so the whole thing can be done with a single TL082 chip.

For starters, we are going to blow off the unity gain voltage follower.  You don't really need it, it was just there for "show."

Here is the new pipeline:

[pulsing output from source battery] -> [current sense resistor] ->  [inverting amplifier with reduced variable amplification] -> [RC averaging filter] -> [inverting amplifier with fixed 4X amplification] -> [multimeter on DC voltage with non-inverted probes]

So, now you connect the inverting amplifier to the one-ohm current sensing resistor.  However, what we are going to do is lower the gain of the inverting amplifier by a factor of four.  That means if we were going to program the first stage of this analog computer (consisting of the single inverting amplifier and the RC filter connected to the amplifier output) to output negative eight volts representing eight watts, now we are going to program the analog computer to output negative two volts representing eight watts.

This gives the inverting operational amplifier four times the head room to deal with "taller" pulses being sensed by the current sensing resistor.

Now, we have a job for the spare op-amp that remains in the chip.  It's trivial, we just want it to amplify by a factor of four to bring the voltage output back to eight volts.  Note that there is a 1 Kohm resistor that is part of the RC filter.  We don't want this second op-amp to load down the RC filter too much.   Therefore we are going to increase the values of the input and feedback resistors.  Let's say that we use a 50 Kohm input resistor and a 200 Kohm feedback resistor.  The output from this second stage will be eight volts representing eight watts.

So, sorry to throw a monkey wrench into the works, but there you have it.   Revision two is a much better design and has four times the headroom for coping with the current sense pulses.  That also means that you may also get away with using 9-volt batteries again.

[TinselKoala]

http://www.youtube.com/watch?v=pyuFehBV_uA

[TinselKoala]

The schematic currently in use:

Changes:
- eliminated the diode in series with sense coil
- added 2 ea. 22K resistors, one on either side of the Setpoint trimpot to spread out the setting range
- added Secret of DPDT to enable easy reversing coil polarity to run in repulsion mode
- changed mosfet to IRFP450
- changed recirculation diode to 1n4007

[MileHigh]

TK:

- added 2 ea. 22K resistors, one on either side of the Setpoint trimpot to spread out the setting range

Astounding!  Visualize the voltage dropping like a multi-stage waterfall!

- added Secret of DPDT to enable easy reversing coil polarity to run in repulsion mode

Hmm.... Think about this one TK.  I looked at your latest clip and I was looking for the switch-on at TDC and I did not see it.

You changed the polarity of the drive coil so that it would push instead of pull when energized.  However, the sense coil is still turning on the drive coil BEFORE TDC (assuming a "zero offset" sense coil angle), when in fact you need to have the sense coil turn on the drive coil at TDC.

I think that you need to invert the sense coil logic also.  Notice that you don't necessarily have to contemplate wiring in yet another switch - just turn the sense coil around by 180 degrees in its mounting.  I am not sure if that will be easy or hard for you to do, I would have to scrutinize your clips again.

MileHigh

[TinselKoala]

Actually, if you think about it a little more, you will see that it is only the Timing control, the physical positioning of the sense coil, that actually relates to the moment of turn-on and turn-off wrt the TDC position.

Yes, reversing the sense coil also, does produce waveform phase relationships like you would like to see... but the timing must still be changed, although not as much, and here's the kicker: the picture from the strobes is the same, once you have found the "sweet spot" of timing and dwell. Here's the scopeshot of the motor running in repulsion mode, with the sense coil also reversed.

What changes is the position that the setpoint pot has to be turned to, in order to get the dwell angle right. The sense coil needs to be moved the width of the "off" time in order to get to the same timing position wrt the drive coil itself, depending on the polarity of the sense coil.

See the following photos. The first two are repulsion mode, sense coil reversed per your suggestion, motor running at best RPM. The scopeshot and the timing position are shown. The second pair, still in repulsion mode, ditto but with the sense coil in my normal polarity.

ETA: The images are "flashless", at a shutter speed of 1/4 or 1/8 second, so the strobes don't show up properly. You can see the strobe indications much better on the video camera; I'll try to show them later on this afternoon if I have a chance.