Appears to be overunity Cicuit

[linda933]

P.S. My circuit works better WITHOUT the cap in parallel to the
input coil.
Then there are more bigger spikes and thus the input
power also is lower.

But I must still try to do experiments with a different transformer,
which has a lower impedance in the input coil.
My 2 x 22 Volts coils in parallel still has a too high resistance...


It is amazing to see, that when you short out the 2 diodes at the output
with 2 small cables, that the voltage waveform
at the 2 x 22 Volts input coils is changing dramatically and you will be drawing
more input power.
When you don?t short out the diodes and leave them in the circuit,
the input power will fall and there will be a spike induced back,so the input
current falls due to this back spike, but the bulb does not change its brightness
much lower...so this back induction pulse seems to help the circuit save input power....

Regards, Stefan.

@Stefan

What you have drawn is a pretty common circuit.  Depending on which way you hook up the primary (low V side), you have either the essentials of a Forward Converter or a Flyback Converter.  When you short the diodes, you have both at once.  Usually, these converters will have an output capacitor and the forward converter usually adds some means to reset the core if a single transistor approach is used but, in essence, you have shown the heart of these basic circuits here.

As you will find in any textbook on switching power supplies, the forward converter is characterized by having its output rectifier(s) conducting forward power while the primary-side switch is on.  The flyback converter is characterized by its output rectifier(s) conducting while the primary-side switch is off.

The two diodes are in series in this particular case and it will not matter whether they are split one on each end of the transformer or just put in series on either side.  They could just as well be a single diode with a high voltage rating.  You would have the same operation and lower overall losses.

I bet your neon protectors never light up when you have the output diode(s) shorted out because, in that mode forward/flyback, the magnetic field always finds a load in the bulb when it is collapsing.  That is why the input waveform looks quite different (less spikey) in that mode, because you have a resistive reflected load for both polarities with the diodes shorted.  The only inductance to provide a spike is the small amount of leakage inductance; the result of slightly less than perfect magnetic coupling in every transformer.

They also probably don't light up when you are in flyback mode, since the collapsing magnetic field finds its current path through the diodes and lamp when the primary is hooked up that way.  Note that you can change between forward mode and flyback mode by either reversing the primary winding polarity or the diode polarities.  Either way gives the same end result. 

In forward mode, there is no means for resetting the core except your neons.  In a one-transistor drive scheme, resetting the core is just another way of saying "allowing the field to collapse through a current path" and it is something that is very important in one-transistor forward converters.  Often, a third winding and catch diode is used, returning the core reset current to the power supply.  In flyback mode, the resetting of the core is inherent, as the collapsing field is what delivers current to the output circuit.

I think all the behaviors you have observed so far are normal, predictable and quite conventional.  You might want to freshen up on your switching power supply basics...the PDF app note file below is interesting and basic...might help you figure out a better way to reset the core and return the power to the input supply...seems like you have an extra winding available...

Also I have added a nice basic PDF on the flyback converter and its topology...

Linda

[hartiberlin]

Hi Linda,
thanks for the 2 PDF files.

Well, I think, there is still going on something else here.

Otherwise, how could the lamp stay bright all the time
up to around 100 Khz, when I am still using
a old iron transformer only build for 50 Hz ?


I will experiment further starting about in  2 hours from now and
then post a couple of pics and a video later on.

[linda933]

Hi Linda,
thanks for the 2 PDF files.

Well, I think, there is still going on something else here.

Otherwise, how could the lamp stay bright all the time
up to around 100 Khz, when I am still using
a old iron transformer only build for 50 Hz ?


I will experiment further starting about in  2 hours from now and
then post a couple of pics and a video later on.

Old iron cores, especially those with the thinner laminates, have no problem going up to frequencies well above audio.  The limiting factor is usually core losses, but you must realize that you are only putting a few watts through this, so it is probably operating well below any core loss limitations even at 100KHz.  Another frequncy response rolloff mechanism is "leakage inductance" due to finite magnetic coupling coefficients.  This appears as a series inductance to the transformer and can often be compensated with a resonating cap to extend transformer frequency response.

As you increase power throughput, I'd expect two problems:  First would be core saturation because of DC bias.  Unless you add a symmetrical flux resettting method, your forward converter will tend to push its operating point way over to one end of the B/H curve and this brings on early saturation.

The secod problem would be heating of the core due to core losses if you operate at high frequency and high power simultaneously on a physically small core.

Linda 

[Spewing]

hey harti, is this your findings? i mean did you make this site?

http://jnaudin.free.fr/html/overtep.htm

[hartiberlin]

hey harti, is this your findings? i mean did you make this site?

http://jnaudin.free.fr/html/overtep.htm

Well, I had the basic idea back then,
but did not have the time to fully check all different setups out.

There was some capacitive coupling via the gate of the Mosfet
where additional power came into the circuit.

Will also try again this circuit soon.
Sorry for not yet providing the videos,
but I am still testing things out and don?t want to rush any
conclusions yet.

Linda is right, that the core can easily go to around 30 to 50 Khz in my case now,
without having any big impact on the output light at this power level
around 20 Watts I am now doing these experiments.
Also have to let it run longer to see, if the iron core
of this big transformer gets anywhere hot.

I still have to test a few more caps with this setup,
but in the moment it looks like this is not overunity
or at least.,what I put in comes out there at the bulb.

The first video user Spewing
posted had a very to light sensitive camera,
probably his mobile phone, so the 60 Watts bulb looked
very bright, but it probably was not in real live.

I also can light up a 25 Watts transparent bulb pretty much,
but I also put in about 14 Volts x 1.3 Amps,
so it is around 18 Watts and the filament looks also like about around a
20 Watts glow.

As user Spewing has already calculated his output at around 36 Volts was lower
than the input power, so I still wonder, if the 60 Watts bulb
still lighted really up as it would be  under a normal 60 Watts bulb at the grid,
or if it was lower and only the videocam overexposed it ?

I also have the effect, that when I don?t connect the lamp, the cap
at the output charges up to around 150 or 200 Volts ( depends on the configuration)
and then when I connect the bulb, it is bright in the first moment and then goes
back to normal glow only.
But I also had used only a 220 uF ( 400 Volts type)
only at the output !
User Spewing,what cap did you use ?

This is pretty dangerous, cause if you accidently touch this
cap the high voltage DC can kill you...
ALso if you accidently short this cap it gives you a real loud
bang !

ALso this is not "Cold" electricity as I once touched accidently
the pulsing output and it gave me a "very nice jolt" !
( Fortunately I did not have the 220 uF cap connected during this ! )

Regards, Stefan.