Multiphase Resonant Circuits

[sourcecharge]

It's hard to believe that I'm to only one doing the research on core resistance, does anyone else have similar experiences with the legs equation calculation of core resistance vs under load conditions?

[Thaelin]

High Sourcecharge:
   Actually they are discussing this  in another thread. In kapandzies cousin, they are realizing that some of the cores do have a lower ohmage than standard ferrite. I have noticed that in a few cores that I have played around with. Grum made mention of it so that is where I found out of it.
   Dont give up the race. Follow your heart and keep your eyes on the goal. That is what fires the persuit after a cause. Because I have always had a strong background in mechanical, I have been keeping my eyes on a specific area of interest. But I never let a possible idea go by with out looking into it.

be well and over look my club fingers.   thay

[sourcecharge]

High Sourcecharge:
   Actually they are discussing this  in another thread. In kapandzies cousin, they are realizing that some of the cores do have a lower ohmage than standard ferrite. I have noticed that in a few cores that I have played around with. Grum made mention of it so that is where I found out of it.
   Dont give up the race. Follow your heart and keep your eyes on the goal. That is what fires the persuit after a cause. Because I have always had a strong background in mechanical, I have been keeping my eyes on a specific area of interest. But I never let a possible idea go by with out looking into it.

be well and over look my club fingers.   thay

Yes, this is mainly because of the permeability of the ferrite cores, and it's ace properties....

Magnetic Measurements at Low Flux Densities Using the Alternating Current Bridge
By Victor E. Legg

Legg's equation states:
R(core) = u*L*(a*(Bmax)*F + c*F + e*F^2)
Where:
u = reference permeability
L = inductance (H)
F = frequency
a, c, and e (ace) are properties of specific types of material and they vary under different Gauss conditions..
a = Permeability-magnetizing force coefficient
c= Residual resistance coefficient
e = Eddy current coefficient of core
B(max) Gauss= V(rms) * 10000 / (sqrt2 * pie * N * A(e) )
Where:
N = number of turns
A(e) = effective Area (cm^2)
V(rms) = voltage rms across the inductor

This equation is used with MPP cores from a manufacturer formally known as Arnold's Magnetics.
Arnold's Magnetics is now Micrometals, but they used to have a datasheet that listed all of the ace values for all of their different cores' permeability.
MPP cores are the most efficient in core resistance, with the exception to air cores, which really isn't a core...
They only come in toroids.

My problem is somewhat different....
I am able to design, calculate, and measure an inductor using these equations for specific Q.  Last one was for about 250 and measured with my mastech 5308 LCR meter which fluctuated between 230 and 300 Q.

Here is the thing, the equations within my 2nd post are known to work for the Vp(out) of a series resonant circuit if all resistance is known....tested under B2 spice...and observed using air coils with air capacitors....
So when I put in a certain Voltage in, and get an amplified Voltage out, I can specifically state that there must be a exact amount of series resistance within the circuit within an accuracy of the digital oscilloscope...

Most of the resistances can be measured and at only 10khz, the only resistance that would be of significance that cannot be measured is the core resistance. 

AC wire resistance and parasitic capacitance of the inductor should be very low....

First because the frequency of operation is at only 10 khz, and AC wire resistance does not even measure when below 50khz with 28 AWG wire:

Second, I cannot believe that a progressively wound core with polyimide Heavy insulated magnet wire with a parasitic capacitance of only 6pF (64100 hz Self Resonant Frequency) could have that much effect at 10khz and is not variable to voltage increase. 
Polyimide has a dissipation factor of only 0.002....that's very low.....so, a capacitor does not decrease in efficiency well below it's dielectric breakdown point, and in this case, Heavy polyimide magnet wire breaks down at 4000 V, where I'm only generating <1500V with 5 or 10 V inputs ...

So when you measure a core at 230 - 300 Q and the output voltage of the LC circuit shows a resistance of a corresponding Q of only 120, the backward calculations of the magnetic field show that in order to generate the amount of resistance within the core, it would have to be at 167 Gauss but the magnetic field was calculated and measured to have only about 6.5 Gauss...

In conclusion, the input actual voltage was 5V, but backward calculating the required input voltage to generate 167 Gauss, would equal to an input voltage of 129 V.!!!!!!!???!!!!

WTF

[sourcecharge]

Just wanted to correct myself because I was posting of the top of my head instead of checking my documentation, and post some of some really hard to find pdfs...

First:

B(max) is Flux density....
B(max) = V(rms) x 10^8 / ( sqrt(2) * pie * F * N * A(e) )

so ya, a little different, but my spreadsheets were correct, I simply didn't remember them exactly...

Second, Micrometals is not giving these out anymore, and they are 10+ years old but they still work great....

The intro pdf has all the formulas that anyone would need to produce high quality inductors...

Funny that it's no longer on the web...

[sourcecharge]

Over the coarse of last year, I wound 20 powdered iron type toriod cores with my toriod winder.  They were all very similar and were wound with exactly 1010 turns.

Each core has different properties, even when purchasing the same core.  That means all 20 core were not exactly the same inductance.  At which point, I adjusted the number of turns to account for this variance.  After which, the output of each series resonant LC circuit was different for each other, due to slightly different resistances from core loss and copper loss.  I introduced a potentiometer between the capacitor and inductor to balance the output voltage peak of each series resonant LC circuit.

Results:  I have found that the EPRR (or the first ratio) using 20 phases was about 5.3.  This is significant due to the fact that B2 spice is showing anything above 4.8 would output overly efficient output.  The circuit was designed to only measure the output and not to achieve an overly efficient output due cost restrictions, and inventory availability of MPP cores.  Another words, I knew this was only to check to see if spending the money on expensive cores that would work in my automatic toriod winder would have a desired output.

The output voltage level was higher than predicted from B2 spice. 

There was a problem though.

The current measured equated to a 91% Power Factor.  Single phase measurements of the same core as well as every other LC circuit under the same circuit and conditions showed around a 80% to 76% Power Factor.

I am unsure if this is due to the introduced pot between the L and C, if it is due to the lossyness of the iron powdered cores, or simply because of the actual circuit result using 20 phases.

I noticed that when using 20 phases, if the ground was disconnected from the capacitor, but all of the capacitors were on the same ground line connected together, there was no observible output or input difference.  This led me to use LCL at 10 phases, and found the same result.

This test circuit was not designed up to the ability of the test requirements.  Meaning, there are many varibles that need to be accounted for. 

Future tests should include air core inductors to limit the core differences.  This may increase the complexity of the test circuit, but it may be cheaper than buying large numbers of MPP cores.

Finally, I would like to say that if the Power Factor continues to increase as the number of phases increases, I'm theorizing that it will not ever go above 1.

Therefore, even if the number of phases increases the Power Factor, the EPRR ratio should exponentially increase up to the number of phases used.  This would overcome the Power Factor increase, but such an experiment would require ALOT of phases to overcome it.

I might be wrong though, the Power Factor may go above 1, which would be crazy btw, so at which point, no matter how many phases are used, there would never be an overly efficient output.

Oh, and, if this works, and the increase in Power Factor is only due to the lossyness of the iron powdered cores, I have theorized that the output load would have to be time dilated, and possible gravitational discrepancies may arise.

This means that the time dilated loads would experience more time than the input circuit which should be able to be measured by mechanical clocks, and digital weight scales.

I still have not wound the 20 MPP cores by hand, (not even started) but I have made the tools to do so.  I don't think I have ever procrastinated as much as I have been procrastinating winding these cores.  I am really hoping that I think up an automated way for mass production.  Doing it by hand seems very amateurish.