Simple to build isolation transformer that consumes less power than it gives out

[TinselKoala]

Your results are quite strange.

Inductances in parallel are "supposed" to follow the same rule as resistors in parallel:  1/L_total = 1/L₁ + 1/L₂ + ... + 1/L_n

And in series,  they simply add. Have you tested your coils in series to see if that result is "strange" as well? And what happens if you reverse the connections of one of the parallel coils?

As you are finding out, measuring inductances can be problematic. The very best way is to use the inductor in a resonant tank circuit with a known capacitance and measure the resonant frequency and work backwards from there.

It's possible that the "huge" inductances approaching 2 H are skewing your results, but I don't know the characteristics of the meters you are using.

Many single-frequency meters use 900 Hz as the measurement frequency and this usually will give values that are close to the manufacturer's stated value at low inductances. 2 H isn't a low inductance, though.

[wattsup]

@JouleSeeker

Thanks for your measurements. Curious to say the least.

The point in efficiency is that those numbers 99.31% for that big Metglas unit then compare it to your toroid that states 87%, but both given under what may be considered standard operating conditions (SOC). I would suspect that @JN's transformer will show maybe about 99% just like the Metgals one under SOC. The idea is to find the highest percentage under SOC to then use it under the rather non SOC method to achieve OU. I mean even if you wind your own transformer using whatever, we should not be seeing above 100% under SOC.

When a transformer core becomes biased (in one direction) and it it is followed with a biasing in the reverse direction, that then creates extreme in the state of core change that imparts to the secondary. When the setup is fed AC to primary, secondary and bulb load, the bulb load cannot act like a trampoline (I am watching the olympics where Canada just won the Gold in trampoline - YaMan).

In the @JN scheme of things, the first primary is rather standard method except that the load of the first is not a bulb but the second transformer and the bulb combined via a semi-shorting across the bulb load. This will generate a whole host of harmonics inside the second transformer that will reverberate back to the first transformer.

Now in the case of using standard laminated transformers like my last trials, seems like the core is simply not reactive enough to pick up those harmonics and reverse reverberations in order to create a more havoc stricken extreme change in bias. The core just ignores those effects. To a lesser degree the ferrite cores are doing the same thing and this is why when I did the test with my two toroidal Hammond coils, the results were far better then with the laminated coils.

So logic would have it that with a core of higher permeability, could also mean higher sensitivity to those other harmonic and reverberate effects that is run under the right conditions will provide or favor a combining effect and not a cancellation effect.

So all this time up till now just to learn that the core is the key. This is a good confirmation to now realize not only for this device but for other devices but also bad to know because this cancels the potential for standard transformers to be used and calls for the more expensive and more specialty nature of using more exotic cores.

An added realization is that if the @JN method can work with cores, it may also work with air core designs like if you simply used a four individual layer air coil and work out the connections to simulate the @JN circuit.

So now we go core hunting. lol

But before that, let's take this logic one step further. If standard transformers do not have the inherent attributes to work with the @JN circuit, then maybe the circuit should be modified to accommodate the cores "stiffness" or "lower reactivity" by combining in series between the two transformers with two more transformer that do have the reactivity.

See here this ready made choke coil that is using Metglas cores. Yes, unfortunately they only have one coil but if put in series with the two isolating transformers (IT), maybe this could act as a go-between that has the reactive skills to exchange more effects between the two ITs.

http://www.ebay.ca/itm/Valab-4-5H-500ma-Filter-Choke-Amorphous-Metglass-Double-C-core-UTC-Transformer-/251121566211?pt=Vintage_Electronics_R2&hash=item3a78030203

wattsup

[baroutologos]

Your results are quite strange.

Inductances in parallel are "supposed" to follow the same rule as resistors in parallel:  1/L_total = 1/L₁ + 1/L₂ + ... + 1/L_n
And in series,  they simply add. Have you tested your coils in series to see if that result is "strange" as well? And what happens if you reverse the connections of one of the parallel coils?

I have also noted the fallacy of the LC meters concerning those iron core transformers. Mine gives a 1/10 measurment in comparison to real impedance that comes at 50 hz.


and yes.. parallel inductances are like resistors. but not wound on same core (closed magnetic circuit) of course. .e. equal inductances in parallel config wound on same core have equal inducactance as each one. (try checking it with a bifillar coil)

[TinselKoala]

I have also noted the fallacy of the LC meters concerning those iron core transformers. Mine gives a 1/10 measurment in comparison to real impedance that comes at 50 hz.


and yes.. parallel inductances are like resistors. but not wound on same core (closed magnetic circuit) of course. .e. equal inductances in parallel config wound on same core have equal inducactance as each one. (try checking it with a bifillar coil)

That last part is interesting, thanks. It makes sense too, like simply using a thicker wire for a single winding on the core.
What about series connection for coils on the same core.... add inductances like normal?

[MileHigh]

Also note that whether two inductors are in parallel or in series, each coil is "broadcasting" its magnetic flux pattern so the orientation in 3D space of the two coils relative to each other will affect the measured inductance.   Far apart and with their major axes at right angles to each other should give you the best results with near-zero mutual coupling.  Or you might want to maximize the coupling, and have the two inductors next to each other and coaxial.

For each coil the associated inherent capacitance will start to become a factor at higher frequencies.

To establish a good baseline you can do a Bedini-type setup, use a 555 timer output to turn the transistor on and off, and replace the charging battery with a load resistor.  Use your scope and measure the time constant and derive the inductance.

For solid-core inductors you could use the same setup, and with increasing currents, find the current level that fully saturates the core.  So then you know the saturation current.