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

[gyulasun]

--------------------
I played with:
1) A 1 to 1 small trafo (measured Henries: about 0.350 H primary and secondary).
2) A signal generator (see picture) (and also an home made 555  +  Mosfeet square wave gen).
3) And old hammeg HM207 scope.

I have noticed that, with *square* waves:
These square waves "morph" into nice sine waves when you approach the resonant freq of the
trafo (about 170 KHz in my case). A trafo as also some (small) capacitance. No?
When I reach this very freq the 'peak to peak' voltage is multiplied by more than ten.
Beware: no load here! No 'OU' claimed.

Anyway, I did not know that you can transform square waves into (more apparent 'voltageable')
sinus waves with a mere 1:1 trafo should you reach the right freq. Did you?
Of course, with a sine wave you can also observe a voltage multiplication at resonance.

Very Best from Brest,
Yann

Hi NerzDishual,

The same "phenomena" happened with the MEG setup when the builders tuned the input pulse generator to find the highest output voltage amplitude and they found a sinus or very much sinus-like waveform as the peak output across the trafo's secondary coil.  I agree this occurs at the trafo's resonant frequency i.e. putting this more precisely, it occurs at one of the trafo's coils resonant frequency.  (of course if trafo has several different coils then there can be as many resonancies as coils.)  In the MEG case the secondary coils were in the Henry range, (5.7H with Naudin's metglas core coil, http://jnaudin.free.fr/meg/megv21.htm ) and the coil's self capacitance (a few pFs) constitutes resonance with its self inductance and you can consider the secondary coil as an LC tank kicked by the input pulses via the primary coil which has much less number of turns and acts like a coupling coil to the LC tank.  And at resonance, the voltage and current is sinus wave across and inside the tank provided the loaded Q (figure of merit) is reasonable high like say 8 or 10 or higher.  Away from resonance the voltage and current is not sinus wave across or in the tank of course.

Greetings,  Gyula

[JouleSeeker]

  Good points, Nerz and Gyula.

'Why not using at least DC input? With a DC bat and an inverter, for ex."

  As usual, reliably measuring Pinput and Poutput is crucial to our progress.  (Sorry to sound like a broken record, repeating myself here...)

[wattsup]

@JouleSeeker

Thanks for your PM. I will send you an e-mail tonight from my home office.

As for testing with the two recent transformers, up till now I have not seen anything out of the ordinary. I tried connecting in many ways but again nothing special besides what I showed in the last video, that is basically running the transformer at near short circuit with good output, but not OU.

Also tried by putting a 250v 35uF capacitor in many points but again nothing but increased input consumption.

I don't mind about buying the transfo's because I will be using them in the TK tests as well to simulated TK adding his looping transformer.

The only real curiosity is how this is working in terms of applying a AC input then jumping the output of the first transfo to the second and the effect this has on the second output. This can be shown with scope A and B at various places to show the phasing differences. If there is anything really interesting in that, I'll make another youtube to show it.

But with the new tranfos my efficiency rate is never as high as when I used those toroids that were 2 x 120/2 x 12. This is pretty incredible when you thin the 120 to 12 step down is giving me better results then a 1:1.  @JN did mention that the core material may be important and I can see that part of his reasoning to be correct since these standard laminated cores are not giving anything near the toroidals.

So, I cannot at this point certify that @JN is just full of really hot beans. lol

But let's just say the pot is simmering in that direction. lol

wattsup

[TinselKoala]

People were doing reliable power measurements on circuits with complex waveforms long before anyone had digital oscilloscopes, or even digital cameras and calculators.

As I think .99 can affirm, a manual power computation procedure using an analog oscilloscope can produce results that are typically within a few percent of those obtained by digital scopes or even directly from simulators. And of course the computation of instantaneous power produced in this manner -- by multiplying instantaneous voltage by instantaneous current -- fully accounts for power factor.

It's all a matter of resolution and attention to detail. If you are willing to sweat and concentrate for a couple of hours, you can get results with a few hundred dollars of gear (a good analog scope and a digital camera) that are _as accurate and precise as your measurement technique can produce_ even using an oscilloscope that costs ten or _a hundred_ times as much. Of course the DSO will give you the result in milliseconds.... but is it really worth it, to get a fifth or sixth digit of precision, when you cannot possibly do better than three with your standard probes and wires and component tolerances and such?

@wattsup: try putting a good fast diode in series with your capacitor, and repeat your tests looking at the DC voltage across the cap.

Output power measurements can be reliably performed using a "photocalorimeter". Choose a suitable incandescent light bulb for your power levels, and make a light tight enclosure for it. Use a CdS photoresistor and a regular resistor to make a "potentiometer". Feed the voltage output of this potentiometer to one input of a 741 op amp configured as a single-supply voltage comparator. Use a real, 20 turn precision pot and a turn-counting knob for input to the other input of the 741. Have the output of the 741 turn on an LED when the comparator sees the voltages on the inputs equal. Calibrate the system with known DC power input to the bulb.  Input your unknown, note where the knob position is when the LED changes state, feed known DC in until you get the LED to change state at the same knob position.
Less than ten dollars worth of parts and milliWatt accuracy if done right.

[TinselKoala]

I'd also like to "inject" a theoretical point here. In a transformer, the voltage of the secondary depends significantly on the rate of change of the voltage, hence the current, hence the magnetic field of the primary. So for the maximum voltage rise in a secondary, you want to get the maximum rate of current increase in the primary: drive with as square, fast risetime pulse as possible, and you will get the maximum voltage in your secondary at the resonant frequency. The importance of a fast risetime pulse for maximum voltage is masked because of core materials. They do not allow the benefits of fast risetime and virtually unlimited magnetic field strength because of saturation and viscosity effects. This is why air-core coils, trafos if you like, are used for extreme HV purposes like radio transmitters and Tesla coils.