Muller Dynamo

[mariuscivic]

I think i got something here.
FINALLY i managed to charge a cap with no lenz at all . The gen coil is not affecting the rpm. i  tooked out the hall and connected a signal generator that gives pulses every 5 microseconds.This way the sine wave is shorted every time, not only at the peak. ISN'T THAT CRAZY??? The cap that i charged is 22uF and and not to fast too. 20V in 10 sec. But i'm sure that after i'll tune this circuit there wiil be much  more power out

[gyulasun]

Just finished testing parallel mosfets.I'm usingIRF840 wich have 0.8ohm(400V / 8A). I put 10 of them and the result was not good.The voltage droped from 150 to 60V.Just like the IRFS640 wich have 0.18 ohm.(200V / 9.8A). Maybe i must change the coil...thicker wire....i dont know.

Hi Marius,

When you connect MOSFETs in parallel, their input and output capacitances also add up, in case of IRF840, Cin=1300pF  from data sheet when drain-source voltage is 25V and this is voltage dependent and increases when you use 12V drain-source voltage ( http://www.datasheetcatalog.org/datasheet/stmicroelectronics/3013.pdf )  and when you parallel 10 IRF840s, the input capacitance increases tenfold too, to 13 - 14nF  so I believe you need to use dedicated driver IC (TC4421 or 4422 etc) to control the FETs.  (Doug has mentioned this too.)  Because the tenfold increase in input capacitance makes the ON and OFF switching times much longer than in the case of a single FET and the induced voltage at flux collapse is inversely proportional to the switching time:  V_induced =L*dI/dt  where L=your coil inductance, I= the current at the moment of switch-off and t= switch off time.  The "secret" to switch MOSFETs on and off very fast is to be able to charge and discharge their input gate-source capacitance and this needs strong current source and current sinking capability from the driver, this is what the dedicated driver ICs are designed for.
So the slow switch off time explains  why you get only 60V with your 10 paralelled MOSFET assembly.  When you can obtain / use such driver ICs, you can compare the results in the recovered bemf voltages.

Gyula

[konehead]

Hi Mariu
have you seen the video of the "stargate" motor - its a conventional DC motor and he stacks neomagnets around it, and it goes super super fast (but I bet the cores get hot, besides the bearing-sleeves):
http://www.youtube.com/watch?v=5Xv-req4U8U

Anyways, I'll take some pictures of my romero-variant machine, with all the magnets in place - I'm not so good with videos right now...
basically now that I installed two facing coils in series for about 6.4ohms in coils resistance, the motor goes around 250-300rpm wihtout any magnets and with magnets, it now goes 960rpm which is huge difference -  only magnets doing this increase in speed!....draw to motor while it goes 960rpm is 60ma and 12V ...which is really low draw plus I have 5 coils on each plate, with ferrite cores too...Ihave only 4 magnets in rotor, going to put in 4 more for 8.....

[konehead]

Hi Mariu
Looks great, the signal-generated shorting...this is all new territory so whatever you can come up with is great......that is very small cap at 22uf if you can do this with say 500uf or 1000uf or 22,000uf cap then it will really be dramatic.
What Gyula jsut said about your paralell-mosfets must be what it is!  (try some drivers!!)
you can get the 4421 chips on ebay really cheap - about 1/2 the price of getting them through digekey or newark or mouser...

[mariuscivic]

Hi Gyula!
Thanks for the info on the mosfets. As i said before i'm no good with electonics;especially with details, so any info is welcome!
Still playing with coils and shorting. The gen coil in the picture produces 3V at 3500rpm.(the magnets are very weak). With this kind of shorting , voltage rises much more than i expected. I have connected to the mosfets a simple flyback driver circuit. Now lenz is here again but not so strong. Pulses are around 5 microseconds .If i could give only one short pulse per sine wave i'm conviced that we could charge caps, large caps, with allmost no lenz