Muller Dynamo

[chalamadad]

Marius, you will find that there is a sweet spot when to short your coil for highest output. When I was shorting with the reed switch I got best results positioning it somewhere between two coilsets and not necessarily exactly next to the coil. This was dragging the rotor but when I was shortin the output behind the FW bridge DC side the rotor would speed up and sort of make up for the drag.

What happens is that the reed shorting produces those large spikes on the scope. These create drag. But when the output side is shorted or connected to a load (which is not too big) or a cap the spikes will disappear because what you do is you take the energy out of the system and get a clean AC wave again that won't slow down you rotor but you get the extra energy out. But you cannot take more out than the spikes allow you to or your rotor will slow down quickly.

With the reed you can sort of shape the coils waveform on the scope. What you want to do I guess is to NULL OUT the part when the rotor magnet wants to escape the coil. So what you get is something like this:  --´\,--´\,--´\,-- And the spikes will be where the peaks of the wave are.

[konehead]

hi Marius and Chal and all
Had something successful tonight - made a adjustable-pulse width circuit using two pairs of bidirectional mosfets hooked up in series and each pair of bidiretional mosfets has its own halleffect.
One bidirectional mosfet pair has a 4421 driver, so it is normally-OFF and switches-ON
the other bidirectional mosfet pair has a 4422 driver, so it is normally-ON switches-OFF
The distance between the two hall effects, with each tripped by the same trigger-magnet, will determine the PULSE WIDTH of the coil-shorting event...
It is very important for a short pulse width to prevent lugging of rotor.
At around 60hz, I got the pulse width between 1/2ms and 1ms and the 22uf cap filled up like crazy -up to 300V and the rotor did not lug down in speed...in fact after the cap filled up some, it started to speed up a bit.
Last night I had 3.5ms p[ulse width at the same 60hz...and it LUGGED....
So figure 60hz has 8.3ms in one phase, from zero point to zero point, so in one phase of a sinewave, whatever the speed it is happening, you would want the coil-shorting pulse width to be about 1/8th to 16th of the length of one phase - so that is how much of the "peak period" the bidirectional mosfets should turn on.
Here is the circuit I tested out tonight that has adjsutable pulse width:

[konehead]

Hi all
One more thing, is that the coil short must be on the PEAK of the sinewave, not at the "zero line area" like Chal's drawing shows.....doing it at zeroline area makes very meager voltage rise in the cap - it doesnt lug being shorted in that area of sinewave since no real power is being pushed into the cap....
imagine a motor coil pulsing a magnet repulsive - there is a spot in timing at TDC where very slightly to the left of TDC will push the rotormagnet to the right, and very slightly to the right of TDC will push the rotormagent to the left...and there is a spot in timing at exactly TDC, that will not propel the rotormagnets to the left or to the right....that is the point you want to find when coil shorting and thats why you need the short pulse width, to capture only that exact TDC point where the rotor magnet does not care which way it moves from its reaction to the coil being shorted....

[mariuscivic]

Hi Konehead!
I did filled the cap without any drag from the rotor but is important not to lug the rotor when starting shorting. My new rotor will be ready next week and i'll begin experimenting with only 8 magnets all N or S. I will try your new circuit. I get the same results ( for now) with IRF840 wich is much more cheaper.
About the speed up effect: in my case it was the FWBR that wasn't working soo good when high spikes occured. Now i have a FWBR from a copmuter power source and the cap is filling faster without the speed up

[chalamadad]

Hey Kone, good work, interesting concept using the non-inverting and inverting driver. It could be useful as two-stage circuit as well using just one hall.

With a scope you can optimize the best points to short and unshort at both peaks of the wave. You just want the wave to either go up OR down (depending on the polarity) like a vertically stretched 'S'. You cut off one half of the wave but the amplitude (voltage) will stay high PLUS the spikes amplifications which fill your high voltage caps. (see below) Now one thing to try is adding more small caps in parallel to your first cap and see how they all fill at the same speed. This is not possible with just one big cap! After that discharge the caps to a load at the time when they are currently not being filled. It might be helpful to use a 1:1 transformer between the caps and the load.

Rotate this 90°:

|
~
|
~
|