Self Running Micro TPU, with closed loop.

[EMdevices]

Thank you for your patience everyone.Ã,  Here's a visual diagram of the circuit operation.

Component Values:

In the 2.5 hour configuration, I used,Ã,  Ã,  R = 1 Mohm and C = 1.5 uF
The main electolitic is 1000 uF.

Number of turns:Ã, 
Ã, Ã,  15 for Recapture coil,
Ã, Ã,  30 for Colector coil (2 coils of 15 turns each),Ã, 
Ã, Ã,  7Ã, Ã,  for Trigger coil

The operation depends on two oscillator types in one circuit: Ã, 

Ã,  1)Ã, Ã,  RELAXATION type oscillatorÃ,  (RC time constant determines interval between pulses), andÃ, Ã, 

Ã,  Ã, 2)Ã, Ã,  BLOCKING type oscillator, for a one time flux driven triggered pulse (Monostable operation)

Uses for this can be quite varied.Ã,  But I'm sure most agree it's a neat simple and efficient way to generate timing pulses for driving other circuits, or just as a fun blinker.Ã,  However, since it's so efficient, I believe that it will alow us to find OU operation due to it's sensitivity.

EM

[hartiberlin]

Great design EM !
So you are now using only 3 coils.

Maybe just put a few LEDs in parallel so you get more current from the
spike back into the cap.

Did you try more windings on the last induction coil with the LED ?

Did you try to measure voltage spikes from the last LED coil ?


Regards, Stefan.

[hartiberlin]

Hi EM,
could you please post some scopeshots of the voltage at the base to emitter
and at the LED anode versus emitter ?

This circuit would probably work much better with a big transformer iron core
as you would get much bigger spikes from it.

Regards, Stefan.

[wattsup]

@EM

Is that a special type of transistor and also can you say what is the ferrit diameter.

I am thinking if there was a small inductor inside the ferrit core, maybe some extra mag field could be again sent back to the circuit in parallel to the big cap. May be enough to have magnet start.

I'll ask z_p_e about transistors on the Faux Pas thread.

[EMdevices]

Stefan, I played with the coils a bit, in parallel and series.Ã,  The smallest amount of coils needed is 3, but more can be useful for experimenting.Ã, Ã,  The center tap on the trigger coil was there from another experiment, and I played with that a bit as well.Ã,  The spikes voltage from the kickback rises to Vcap + Vdiode.Ã, Ã,  So a diode with the smalles drop would also improve efficiency.Ã, Removing the diode makes it work longer, but you can't see it , what fun is that. (there are still losses so it appears to also come to a stop)Ã,  If I don't connect the reclaim coil to the cap and instead to a resistor, then the voltage can go realy high, it's just the normal kickback from an inductor.Ã,  Ã,  The blocking oscillator function involved in triggering the spike, has a very sharp transition (small risetime and fall time) so this can rise the voltage quite high if not connected to anything. (or high impedance)

wattsup,Ã,  the transistor is a simple NPN switching type I bought from radioshack,Ã, Ã,  It has aÃ,  beta = 200,Ã, Ã,  so small base current x 200 gives the collector current.Ã,  Ã,  pese was suggesting using germanium transistors because they have lower Vce drop and lower Vbe activation voltage.Ã,  Ã,  I would also say to use a transistor with as high a beta as possible.Ã,  Ferrite diameter is about 1.5 inches (about 4 cm)

I'll do some scope shots tonight when I get home.

EM

P.S.  In my picture of the circuit, you will see a resistor and diode connected to the main electrolitic, that's a zener diode at 5.5 V and the resistor about 400 ohms.    I placed that there for safety, just in case I had a runaway even,  shows optimism doesn't it ?   LOL