Hi Mr uu
 
It depends on lots of things besides the widths of coil and widths of magnets - such as aircoil, or coil with core in it - when you have core, then the core gets the "brunt" of the magnet, and the core is what gives the "signal shape" to the coil, and its not a direct magnet-coil thing going on, like in an aircoil.
yes I agree its the edges of the rotor magnets are what induce the peaks in an aircoil...But wha is interesting to me at least, is  it makes no sense when using hockey-puck shaped magnet that there are two "edges" to magnet, since you can twist the rotor-magnet around, and it doesnt make any difference to the signal-shape(this makes no sense to me at least!)
While, if you use a rectangular shaped brick or block shaped magnet, then theer is an obvious N-S polarity to the ends of the magnets, the sides, and the faces.....(why Beardon alsways calls magent "dipoles" since there are 6 sides to dice, and 6 sides to magnets)
Also the speed in rpms of the rotor makes a difference in the signal-shape too in how huch time the magnet gets to react to the coil...
I personally like to string two "adjacent" coils on a coil-plate  connected together  "backwards in polarity" and in series too, and then have one magnet in rotor pass by each, and so not have a single magnet for each coil, ( this is with aircoils) this gives a perfect AC sinewave as shown....this means explained in another way, if 8 magnet rotor, you would have 16 coils on coil plates each side,  (this mullergen style sort of ) wiht 8 of those coils face-on with the magnets to get the perfect sinewave shown here with photo of AC signal from coils indeuced rotor of 8 all-N magnets and it has 16 aircoil positioned each side:

Hi Marius,

Here is a 555 timer circuit with independly variable duty cycle and frequency adjusting possibility:

http://www.overunity.com/8597/solid-state-orbo-system/msg234095/#msg234095 

The output of this can go directly to the MOSFET gate as shown in your above link or if you could use a MOSFET driver like TC4422 or 4421 etc it would give a faster switch.   Also if you could obtain the CMOS version of the NE555 like LMC555CN from National Semiconductor or TLC555CP from Texas Instruments, then these types have much less fall and rise time output pulses then the old NE555 bipolar process type have, so they are able to switch faster (you could omit the 2N2222 driver but just test this, it is better to have an "isolating" stage between the MOSFET and the output pin of the 555.  And the CMOS version is fully pin compatible with the old NE555.

Gyula

Hi Gyula,
    Nice simple circuit, what cap value connected to pins 2&6 would generate a 2mhz frequency, would it be a higher or lower value? sorry for the dumb question, i don't have an ocilloscope to check.


    Thanks!

Sorry, my comment was not pertinent after well experience speaking of you, Kone 

hi Mariu
here is my pulsewidth adjsutment circuit again: (you need driver chips with it)
 
 

Crazycut,

Well,  you have consider the frequency adjust potmeter as an R1 resistor value, besides the capacitor.   Generally an RC time constant is to be considered, the smaller the R  or/and the C, the higher their (1/R*C) value i.e. the frequency. For the 555 timer there are equations to determine the output frequency or there are charts to figure out the approximate frequency. 
Here is some charts to see some R and C values: http://www.williamson-labs.com/555-tmg-charts.htm  There is a tutorial on the 555 in their pages too.

IF you really need 2MHz frequency, then you have to use the CMOS types, like LMC555C or TLC555C because the old NE555 can run up to 120-150kHz only.   Here is a data sheet for a CMOS type, LMC555 and it turns out this type can run up to 3MHz in the astable mode and the capacitor is 200pF (minimum recommended value) and the R1 or R2 resistors are some hundred Ohms only.
see Page 5, maximum frequency test circuit:   http://www.ti.com/lit/ds/symlink/lmc555.pdf   

IF you happen to have a digital multimeter which is able to measure frequency directly, you may use such instead of a scope's frequency check.

Gyula

EDIT:  here is an online circuit simulator, showing a low duty cycle 555 circuit, slightly different than what I showed above but still practical and you can check the frequency too.  And if you edit the circuit resistor or capacitor values inside the Java applet, you can see immediately their effect ( just right-click on the component you wish to adjust).
http://www.indiabix.com/electronics-circuits/555-low-duty-cycle-oscillator/

Notice:  in this latter circuit when you vary the duty cycle or the frequency, then BOTH the duty cycle and the frequency changes.

Hi Gyula,
    Thank you for the reply, and this info surely helps alot....
Cc

OK Konehead; that is  interesting what's going on there...
Lets say that you are running your dynamo without the backing magnets. When you start adding them, does the input power goes up and then stabilise or stays always the same?
If the input power stays the same then you really have something there( i dont think anyone  had this effect before) (then I'll kindly ask you ALL the detaills of your dynamo  )

Hi Konehead,
   
     I beleive that romero did a reversed or vice versa polarity backing magnets (or maybe not), to compensate the cogging, im thinki'n lets say coil set#1 has a north polarity on top & south below, (N-NS-S), as the rotors magnet aproaches the coil with core, the attraction is nullyfied (with correct adjustment) then coil set#2 has a south polarity on top & north below, (S-NS-N), as the rotor approaches it is being attracted towards the coil with core (again with correct adjustment) experiment this all around the coils to get best performance in rpm and power in vs power out, my point is its making a push and pull action thus helping the rotor to spin more freely, and maybe this confirms your findings on your motorgen. don't know if it would be the same when the coils are loaded.
   
     Romero did mention (don't remember what post is it) that faster does not mean its better! what do you think kone?

Hi Crazycut!
Thanks for sharing your ideas.
May I ask you what means "cogging" here. i've checked in translation dictionaries on web and they say it's about "teeth-wheel", there is no here, so could you help me to understand in which mean you were using it please?
Cheer.