Sharing ideas on how to make a more efficent motor using Flyback (MODERATED)

[conradelektro]

Some thoughts about the capacitor in parallel with the "fly back coil":

During my tests I got the impression that the capacitor has to be chosen according to the energy in the spikes.

If one uses a low DC resistance and low impedance drive coil (like Luc or woopy), the capacitor should be 1 µF to 10 µF (because it has to store a considerable amount of energy, e.g. there is about 1 Watt in the spikes if about 50 Watts are fed into the motor or circuit).

If one uses a high DC resistance and high impedance drive coil (like I do in my equivalent circuit), the capacitor has to be very small (e.g. 47 nF), because the energy in the spikes is only about 10 mW if about 500 mW are fed into the circuit.

No capacitor is bad, because the fly back spikes are not stored (and not flattened). Too high a capacitance is also bad, because the spikes are flattened too much. In my case I use a fly back coil (the primary of a step down transformer) which would like 220 Volt (because it is a step down transformer from 200 V to 6 V). Therefore the spikes should be flattened to about 220 V in the ideal case. (Remark: I only manage 80 Volt in  my circuit).

Just some very empiric considerations.

Greetings, Conrad

[verpies]

And with "non-SiC Schottky" you mean something like this:
Gallium Arsenide Schottky Rectifier
http://pdf1.alldatasheet.com/datasheet-pdf/view/166395/IXYS/DGS20-025A.html (250V)?
or this:  http://ixdev.ixys.com/DataSheet/L652.pdf (600V)?

No, because the forward voltage drop (V_F) of these diodes is 1.2V or more
A good Schottky diode for this purpose would have a low voltage drop around 0.3V

[verpies]

The "capacitor alone" is a better and cleaner approach for a theoretical power calculation in the spikes. Practically it will be difficult to empty the capacitor after each spike in a clean way (without effecting the storage of the spike energy).

Not really - e.g. Itsu did that once with a relay 
However, the simplest way to empty the capacitor is to connect a 1MΩ resistor in parallel with the cap and give it a very long time between the "flyback" pulses, so this resistor has the time to discharge the cap meanwhile.

I like the idea of harvesting the spike energy with a coil. After all we are dealing with a magnetic effect and why not try to get back a magnetic force. Putting a coil in the fly back path is a new idea, which I have not seen before (or which I have not recognized before).

I also like the idea of recovering the remaining energy stored in inductor (L1) and transferring it into a second inductor (L2) for doing extra mechanical work.

...but using a direct connection between L1 and L2 is the least efficient method for transferring this energy.  You are not doing this fortunately, but your transformer idea is only a little better since it does not maximize the time that the current flows in the second inductor. 
You actually do not even have a 2nd inductor (L2) that could attract a piece of a rotor in a motor (a transformer is not an inductor, unless the flux coupling coefficient is below unity).  Maybe you were planning to put the 2nd inductor in place of that 10Ω resistor but that would change the behavior of the circuit, entirely and render the intermediate transformer extraneous.

If the goal is for L2 to do as much mechanical work as possible, then the circuit below is the most efficient way to do it ...while observing these notes.

The ultimate efficiency is realized with Synchronous Rectifiers ...and second best - with low voltage drop Schottky diodes.

[verpies]

During my tests I got the impression that the capacitor has to be chosen according to the energy in the spikes.
If one uses a low DC resistance and low impedance drive coil (like Luc or woopy), the capacitor should be 1 µF to 10 µF (because it has to store a considerable amount of energy, e.g. there is about 1 Watt in the spikes if about 50 Watts are fed into the motor or circuit).
If one uses a high DC resistance and high impedance drive coil (like I do in my equivalent circuit), the capacitor has to be very small (e.g. 47 nF), because the energy in the spikes is only about 10 mW if about 500 mW are fed into the circuit.

Actually the resistance of the coil just wastes energy and has little bearing on the capacitor parameters, unless it wastes so much of it that the capacitor barely charges up.  The final current and inductance matter much more.
The ideal match of the capacitor to the coil is defined by the equation ½Li² = ½CV², where i is the current flowing in the coil when the MOSFET opens.

No capacitor is bad, because the fly back spikes are not stored (and not flattened).

In your circuit, the energy contained in these "flyback" pulses is not stored but burned up in that transformer, but if you deleted that transformer, then this energy would be stored in that capacitor with utmost efficiency.
Actually, there is such a thing as a "bad capacitor" in this circuit.  For example a capacitor that has a voltage rating lower than V=i_L*(L/C)^½, where i_L is the current flowing through the coil when the MOSFET opens.

Too high a capacitance is also bad, because the spikes are flattened too much.

There is some truth in that when the capacitor is too big, because when the transfer current flows too long ( time is equal to t=π(LC)^½ ) then it gets more time to be converted to heat by the diode's forward voltage drop and the circuit's resistance (including coil's resistance).

[gotoluc]

May be Luc could tell us where exactly he is looking for "more energy"?

I never claimed more energy!

I am not criticising Luc, I just want to understand and to measure the energy in the fly back spikes. Where is the magic? Or better said, where does Luc hope to find the magic?

I never claimed there was magic either.

Better move along as I don't think this is for you.  No magic or more energy in what is suggested.

Regards

Luc