romerouk..

keep in mind your reed switch is drawing an arc at this power level.
your magnets are flying by this reed gap and there may be a quenching effect provided by them to the arc duration.

Hi Laurent,

I appreciate your kind words, thank you very much!

Very good informative video again.
I attached a corrected schematic on your double FET switch and Hall sensors because I think there is a drawing error in it: I show you a blue wire which is needed to connect the 9V negative battery point to the common source electrodes of the MOSFETs) and you need to cut the wire where I show a gray line and text, ok? Also, I added black dots where connections are needed. Please check how your actual circuit is wired, I believe it is wrong as you show in the video.
When you short the generator coil by the FETs (where, at what circuit point do you activate it? at the 9V battery?) so when you activate it then all the nice 80-90V_{peak to peak} sine wave voltage seems to be shunted to a lower than 10V or so. I cannot see any spike when you activate the two FETs, some spikes appear later when the gen coil (i.e. rotor's RPM) starts to recover (video time 6:03 - 6:05 or so).  Please switch on and use the Trigger cursor to stop the displayed waves running, this way the spikes will be seen much better.
(Your first or second video on this coil shorting topic showed very nicely the big spikes at the induced voltage peaks when the reed switch was activated at the right time.) 

Thanks,  Gyula

EDIT: The 22 kOhm resistor between the gates and sources is a bit high: a bit speedier switching should happen with 4.7 to 5.6 KOhm or so (also depends on the max output current ratings of the Hall switches). This way the switch-OFF time of the FETs improves. The best would be to use a dedicated MOSFET driver IC for the job, Doug Konzen just showed such schematic and tested it successfully 1-2 days ago).

...
let's go on and
my last video

http://www.youtube.com/watch?v=L0l_zAC-m90

Thanks Gyula for great answer

Hi Laurent,

I also must finish for today but what I quickly can tell you is that first try to use a reed switch just to control the two MOSFETs, ok? I attached the schematic I also showed to Doug Konzen last August. (it is in EVGRAY Photo folder http://groups.yahoo.com/group/EVGRAY/photos/album/1586264660/pic/list ).
When you have this reed control going ok then you may wish to use a single Hall switch first, on one of the sine wave peaks, then with a second Hall on the other peak.

Good night,  Gyula

EDIT you may even leave out the series capacitor too and insert it later when you see useful.

Hi FreeEnergyInfo ,

I believe it correct, I also say first use it without the series capacitor (I mean the 0.47uF) and when you see the spikes at the wave peaks then try the right cap value,  Konzen can give you more info on this latter.

Gyula

The size of the series cap and its use is essential to ensure correct operation found by tuning. The cap acts as high pass filter where the condition of repetitive shorting MODULATES the carrier frequency and provides intense high frequency energy..  The LF carrier is the frequency of the generator pulses. Anything you do at this frequency will immediately  lug back to the generator. However the modulation layer can be syphoned off without disturbing the carrier. This is because the event which created the HF energy no longer exists within this dimension so there is no Lenze effect and thus no lugging after the high pass filter. However the energy recovered is real and often OU.

BTW you don't need a mechanical generator to do this. If you can do all this with a solid state magnet coil setup and a pulse driver then short the pulses. Shorting or quenching is not new Armstrong discovered it about 1922 as Super Regenerative as coil interrupts increase the Q to > million to extract weak RF signals.

Hi Laurent,

The IRF2708 has a bit low V_DS rating, only 75V, the two in series would give 150V maximum peak to peak AC wave shorting, now you have a 80-90V peak to peak generated voltage, if you manage to short it, the peak to peak voltage may exceeds this and burns both FETs.
Here is a selection of the V_DS=200V power MOSFETs from irf.com to some choices, try to pick the lower R_DS types:
https://ec.irf.com/v6/en/US/adirect/ir?cmd=eneNavigation&N=0+4294841672+4294873618

But it all depends where you can order, perhaps you may wish to select from a component seller list? If you show a link to a seller I can choose the best available from them.

The IRF540 has V_DS=100V, so a bit better than the IRF2708.

Gyula

Hi Gyula

I tried your correction on the circuit ( that i got from  the diagram of Doug Konzen) and it works , but i got this very pure sine wave without any spike but with peak to peak 80 and more voltage.

than i tried to modify the hall sensor position to get the spike and nothing happen. Than i simply disconnected the 2 Halls and the scope trace stayed the same.

Than i tried different AC caps and the result is each time very different.

So i have have to go sleeping and perhaps tomorrow will bring some better idea of this really complex thing.

By the way i use now the IRF 540 mosfet  and i intend to order some IRF 2807 with very low on resistance  any advice ?

good luck at all

laurent

Had some interesting results using a Bedini SG motor type setup and the coil shorting.  Initially I was using the SG motor just because it has 4 magnets N-S as Romerouk said is needed around the rotor.  While it has only 4 I used 2 reed switches 180° apart and offset 90° from the coil (of MO fan).  I put a full wave bridge across the output of the fan coil (very heavy duty high volt high amp unit so probably a lot of loss) and fed that back to the run battery.  Had a scope on it and was seeing nice spikes.  Initially the battery was running down slowly losing a hundredth of a volt every few seconds while charging a small 9 volt nicad on the Bedini SG.  Run battery was a 12 volt Nicad drill motor battery pack.  Nothing spectacular there.

   Then I grabbed a 5ufd 400v cap (not electrolytic) and put it in series with the output from the FWBR.  I noticed the motor run battery stopped losing voltage.  But this did not happen until I disconnected the O-scope!  Fluke 87 used and was reading to .01 volts.  It just stayed at 12.32 volts and I charged two 9 volt Nicads and one 9 volt alkaline (yes they do take some charge) without losing any voltage (or at least less than 1/100 of a volt).  The charges were brief but enough to get the 9 volt batteries from abou 7 volts up to around 9.7 volts.  I then disconnected the 12 volt and ran the Bedini SG off the 9 volt batteries to make sure it was a real usable charge.  These worked for the first time in a while as those particular 9 volt nicads were weak and had not been able to run the motor for a long time.  So no real proof of OU or anything but the sudden jump from losing voltage to running the motor and charging batteries with no drop in voltage looks promising and since I was just grabbing parts I think with some fine tuning it can do a lot better. 
   Also the transistor on the Bedini SG was blown a couple days ago so I just scrounged one off a computer power supply to replace the blown one.  The one on there now is an E13007F2.  Probably far from ideal and I had to lower the base resistor to it down to almost nothing to get the motor to run. 
   I also tried feeding the Bedini output for charging a battery back to the run battery through a FWBR and cap but as soon as anything was hooked up that way the motor slowed and voltage started dropping.  I think there may be a way to do this but I haven't figured that out yet.  If there is a way to do that maybe the combination of the coil shorting and looping the charger back to the run battery might be enough to make a self runner...  I'll admit I mostly shooting in the dark here as I'm not nearly as sharp on this as many people here but this test was exciting to see.  I think the coll shorting holds a lot of promise.