RESONANCE EFFECTS FOR EVERYONE TO SHARE

[gotoluc]

Hi Duff,

great post

Thank you for being the first to post a replication attempt.

First thing I would say is your primary is much longer than mine (just about twice) and I think this could be a major factor and maybe also wire gauge could have an effect. The next thing is maybe get the same bulb I'm using as load.

Please PM me your address and I will see if I can afford to send you a complete primary and secondary with bulb for you to test.

What I will also do is go to my uncle's computer engineering business and have him check my measurements with his equipment just to make sure my scope is reading things correctly.

That is about all I can think of at this time.

Luc

[duff]

@all

I want to post a simulation of Luc's circuit (last video) in an attempt to show what I believe is something unusual about the results that Luc is getting.

The simulation shows approximately 0.66Vpp output across R_Lamp, which has a value of 66.6 ohms and was calculated from the scope readings in the last video.

My test results were the same as the simulations.

Luc was getting 3Vrms across that bulb, which is 2.767Vrms more than the sim.

-Duff

[Edit]

The red numbers in the schematic are the node numbers.
Look at the waveform Legend to determine where the wave form was taken.
Single numbers in the Legend, example V(1), is node 1 with reference to ground

[gyulasun]

Hi Duff,

Would you mind running the simulation with a small modification, just to check the following:

R_dummy=1e12  (and not 1e-12)   and connecting it between node points 5 and 0.

Maybe you will get the same results in the simulation but normally you would not put a short circuit (1e-12) between nodes 5 and 3 in practice,  3 being the most sensitive point of the resonant circuit.

In Spice based simulations the dummy resistors are of GigaOhm values, that is what I used to include, to get rid of the ground independent nodes in a circuit.

Thanks,  Gyula

EDIT:  just thought of it: why there is no sinusoidal waveform (or nearly sinusoidal) across the lamp?  Values of L1 and C1 do give indeed 0.5MHz resonance but your pulse generator gives out 54.7kHz??   (PER = 18.28us)  Is that so?

[duff]

Hi Duff,

Would you mind running the simulation with a small modification, just to check the following:

R_dummy=1e12  (and not 1e-12)   and connecting it between node points 5 and 0.

Maybe you will get the same results in the simulation but normally you would not put a short circuit (1e-12) between nodes 5 and 3 in practice,  3 being the most sensitive point of the resonant circuit.

In Spice based simulations the dummy resistors are of GigaOhm values, that is what I used to include, to get rid of the ground independent nodes in a circuit.

Thanks,  Gyula

EDIT:  just thought of it: why there is no sinusoidal waveform (or nearly sinusoidal) across the lamp?  Values of L1 and C1 do give indeed 0.5MHz resonance but your pulse generator gives out 54.7kHz??   (PER = 18.28us)  Is that so?

Hi Gyula,

Thanks for your input. You have a keen eye...

Yes, I should have used R_dummy=1e12, my mistake. I was thinking inductors rather than transformers...

Also I do need to adjust the period.

Right now I experiencing some quirks with ngspice that I need to work through but I will update the above posts when I get that worked out.

Thanks,

-Duff

[EDIT]
I can no longer modify the above post so I have reposted the revised data below.

The updates look more in line with what Luc in reporting
He did not and probably could not give the waveform across the primary due to the USB Scope input limitations.
However, I was not getting that kind of voltage, so maybe that's why I was not getting his results.

[EDIT]
Update a mistake in netlist - changed R_Lp from 0.25 to 2.5 ohms
This really didn't make a difference in the waveforms - yes probably affected the circuit Q.
Although not shown below I did try decreasing the coupling between the coils to 0.60. that resulted in an increase in amplitudes which is less believable.

[gyulasun]

Hi Duff and all,

Thanks for the modifications,  now simulation is improving.   

First thing is I would expect Luc to confirm your schematics in that he also connected his components exactly like that. He did not mention though he objected your schematics I assume it is the same as he built it.

Two other thing in your simulation: 
1) Luc uses a signal generator that I assume has got a 50 Ohm output impedance and I wonder if you included it in your simulation as the inner impedance of your generator?
2) The coefficient of coupling (k) between your L1 and L2 can significanly influence the output towards load resistance (i.e. the bulb)  so I wonder what value you used in the simulation?  Seeing the photo of your coils I would say 0.5 - 0.6 for (k),  a rough guessing of course.  You (and probably Luc too) may have noticed when moving a little one of the coils away from or closer to the other, the output power (hence the brightness of the bulb) is changing accordingly.

My understanding on this setup in general is that the problem is how to utilize reactive power circulating in resonant circuits and the answer is not that easy at all:  at the moment I can only suggest studying Hector's rotoverter setups because the problem basically is the same, there are already so called "peak extraction" circuits that seem to be able to do it (at least in paper) but no any report on a successful VAR power extraction yet... 
Member esaruoho at energetic forum has addressed this same problem just today, his posts #235 and #236
http://www.energeticforum.com/renewable-energy/1507-roto-verter-8.html

Regards,  Gyula

EDIT:  just noticed you use R_lp=0.25 Ohm for the primary coil copper resistance and Luc stated 2.5 Ohm, right?  This has significance in simulation only if you did not use a 50 Ohm generator inner impedance because the figure of merit (i.e. the Q quality factor of L1)  comes out much higher in that case.