PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2018-07-23, 05:03:08
News: A feature is available which provides a place all members can chat, either publicly or privately.
There is also a "Shout" feature on each page. Only available to members.

Pages: 1 2 3 4 5 6 [7]
Author Topic: The final answer...  (Read 25215 times)

Hero Member
*****

Posts: 1872
Mike,

Here is a test with Circuit 2 running with a 10v dc supply.  Again notice the lower operating frequency verses Circuit 1.  Probe ID's are same as before.

Regards,
Pm

Thanks PM

That is interesting, well I think so.

So we have a step charge of L1+L2 at the parallel LC frequency with C1 when the mosfet is high.
When the mosfet goes low L1 discharges through the diode and L2 and C2 discharge at their LC frequency back to source it seems, hence the negative current oscillation.
Now the discharge to the diode is like a rectified positive pulse with hash. If you could expand that we could see better.

Regards

Mike 8)



---------------------------
"All truth passes through three stages. First, it is ridiculed, second it is violently opposed, and third, it is accepted as self-evident."
Arthur Schopenhauer, Philosopher, 1788-1860

As a general rule, the most successful person in life is the person that has the best information.
   
Hero Member
*****

Posts: 517
Thanks PM

That is interesting, well I think so.

So we have a step charge of L1+L2 at the parallel LC frequency with C1 when the mosfet is high.
When the mosfet goes low L1 discharges through the diode and L2 and C2 discharge at their LC frequency back to source it seems, hence the negative current oscillation.
Now the discharge to the diode is like a rectified positive pulse with hash. If you could expand that we could see better.

Regards

Mike 8)

Mike,

Sorry to be late in responding but have been taking care of household duties.  I think you may have meant "mosfet gate" in your analysis above because the action you describe is correct.

What would you like to see expanded in the discharge phase of the circuit?

Pm
   
Jr. Member
**

Posts: 63
Well I have not had much luck with tank circuits yet, but I did get a good handle on 90deg ferromagnetic coupling. ^-^

Several systems of various types and sizes were tested, it should not be hard for anyone here to replicate.  Just wind several turns of insulated Fe/Ni wire in a loop and wind many turns of copper wire over it in a spiral.  Pulse the copper, extract output on the Iron.  Output in these setups is predominantly DC (with a little hash >:-)).

Because of the frequency ranges (~1khz-50khz+) and other factors, it does look to be based on magnetostriction.  This means all flavors of ferromagnetic/paramagnetic metals should work fine.  Nickel battery strip used for chaining LIPO batteries works great.  Just insulate with a layer of Kapton tape before coiling.
https://www.amazon.com/Breynet-18650-Nickel-Welding-Battery/dp/B07CVVZ1Y9

Efficiency is not the best(yet).  I hit a wall at 50-60% efficiency when driving these.  Perhaps I need to let the wire hang more loosely, or add a nickel coating to iron, or maybe add a runner line of copper parallel to the iron wire in order to couple into low-loss copper.  More likely I just need to move on to resonating pairs of these in tank circuits and let the Kunel magic happen.


The neatest feature besides DC output is that within audible frequencies these things literally sing.  Loud.
https://www.youtube.com/watch?v=I2yo-2yFP4g
   

Hero Member
*****

Posts: 1872
Mike,

Sorry to be late in responding but have been taking care of household duties.  I think you may have meant "mosfet gate" in your analysis above because the action you describe is correct.

What would you like to see expanded in the discharge phase of the circuit?

Pm

Hi PM, take your time, my wife and I are a bit under the weather ATM.

The pink trace going to the diode and the mosfet drain.

Have you put a resistor across the output cap to create a load? just wondering what difference there is to the pink trace with and without.

I think we don't quite yet have the iron core/coil right, hence my question about an infinity magnetic current loop on the other thread ;)

Regards

Mike 8)

« Last Edit: 2018-06-29, 12:14:57 by Centraflow »


---------------------------
"All truth passes through three stages. First, it is ridiculed, second it is violently opposed, and third, it is accepted as self-evident."
Arthur Schopenhauer, Philosopher, 1788-1860

As a general rule, the most successful person in life is the person that has the best information.
   
Jr. Member
**

Posts: 63
I'm finally starting to see the different properties for grounded vs ungrounded systems, as it relates to tanks.  Grounded systems simply do not ring as well.  Well, for RC circuits they're fine, but ground seems to snub any radiant/electrostatic impulses.  Too much capacity I reckon. :-\  I suspect that strong HV would overcome that issue, but I've smoked too much equipment to risk diving back into that field quite yet C.C.
Anyway, this has led to a couple questions:

Would it be advantageous to simply float my entire bench (power supply, scope, oscillator, etc) with a large 12v battery bank and an AC sinewave inverter?
or would the battery present so much capacity that you'd lose the benefits of floating?    I do have an off-grid solar rig set up, but relocating it to the lab area would take much effort.
Was looking for some advice before I go break apart and relocate that monstrosity. :P :P


Perhaps it's just better to power devices and oscillators via battery, and use magnetic/dielectric coupling to take indirect scope readings of the tanks?
   
Hero Member
*****

Posts: 517
Mike,

OK, I finally got back to the bench and ran the following test on circuit 2 with C1 = .022uf and C2 = 6800pf with the supply at 10v dc.  Probe ID's are same as before with CH3(pnk) alternated between the AC and DC output and CH2(blu) being the supply voltage.  The load resistance used was a non-inductive 50 ohm film resistor which provided the best efficiency at the frequency of 594.1kHz.

Pix 1 shows the output DC voltage of 11.93v mean on CH3 which results in 2.85w output.  The input power seen in the Math channel is 3.865w mean.  Note the gate drive voltage distortion and the high peak current at the mosfet drain.

Pix 2  shows the AC output waveform at pin1 of the Schottky diode on CH3.

The pix 3 Math channel shows the input at 3.069w mean with no load.  The AC output with no load is seen on CH3 .

Regards,
Pm
   
Jr. Member
**

Posts: 63
Mike,

OK, I finally got back to the bench and ran the following test on circuit 2 with C1 = .022uf and C2 = 6800pf with the supply at 10v dc.  Probe ID's are same as before with CH3(pnk) alternated between the AC and DC output and CH2(blu) being the supply voltage.  The load resistance used was a non-inductive 50 ohm film resistor which provided the best efficiency at the frequency of 594.1kHz.

Pix 1 shows the output DC voltage of 11.93v mean on CH3 which results in 2.85w output.  The input power seen in the Math channel is 3.865w mean.  Note the gate drive voltage distortion and the high peak current at the mosfet drain.

Pix 2  shows the AC output waveform at pin1 of the Schottky diode on CH3.

The pix 3 Math channel shows the input at 3.069w mean with no load.  The AC output with no load is seen on CH3 .

Regards,
Pm

Thanks for sharing, Partzman :)
  From your scope-shots it looks like your coils might be ringing closer to 10Mhz, which might be too high to efficiently switch with your setup.  Maybe adding some core material (ferrite/iron/nickel) will help bring that number down, or adding tank caps, or simply scaling the whole thing up to hundreds of turns instead of dozens.
Of course we're about neck-and-neck right now so it's anybody's guess ;D

If you have an LC meter, you can also try measuring the inductance of the copper coil while slowly powering up the iron coil w/ straight DC and slowly increasing the current.  It would give you a good idea how much inductance 'swing' is there, and what your hysteresis curve looks like.


I was thinking about doing this setup instead with a few dozen turns of iron wire, and several hundred turns of copper wire wrapped toroidally around the iron wire core.  That would provide 90deg coupling and make it act more like a mag-amp.
   

Hero Member
*****

Posts: 1872
Mike,
  Note the gate drive voltage distortion and the high peak current at the mosfet drain.

Thank's PM

Yes I did notice, what are you driving the gate with or are you driving straight from the SG (10v)?

Still analysing the  tests so as to alter the circuit accordingly.

Regards

mike 8)
« Last Edit: 2018-07-01, 15:34:09 by Centraflow »


---------------------------
"All truth passes through three stages. First, it is ridiculed, second it is violently opposed, and third, it is accepted as self-evident."
Arthur Schopenhauer, Philosopher, 1788-1860

As a general rule, the most successful person in life is the person that has the best information.
   
Hero Member
*****

Posts: 517
Thanks for sharing, Partzman :)
  From your scope-shots it looks like your coils might be ringing closer to 10Mhz, which might be too high to efficiently switch with your setup.  Maybe adding some core material (ferrite/iron/nickel) will help bring that number down, or adding tank caps, or simply scaling the whole thing up to hundreds of turns instead of dozens.
Of course we're about neck-and-neck right now so it's anybody's guess ;D

If you have an LC meter, you can also try measuring the inductance of the copper coil while slowly powering up the iron coil w/ straight DC and slowly increasing the current.  It would give you a good idea how much inductance 'swing' is there, and what your hysteresis curve looks like.


I was thinking about doing this setup instead with a few dozen turns of iron wire, and several hundred turns of copper wire wrapped toroidally around the iron wire core.  That would provide 90deg coupling and make it act more like a mag-amp.

On this coil arrangement, the inductance change of the copper winding L1 was ~2% with the 3 turn iron winding L2 varied from 0 to 1.5A dc.  The iron wire I'm using is typical bailing wire obtained from a farm supply store.  Very little inductance change.

Pm
   
Hero Member
*****

Posts: 517
Thank's PM

Yes I did notice, what are you driving the gate with or are you driving straight from the SG (10v)?

Still analysing the  tests so as to alter the circuit accordingly.

Regards

mike 8)

The gate drive is taken from a complementary bipolar half bridge driven by a Cmos 4049 buffer all with a 10v dc supply.  I'm sure the distortion is primarily caused by ground looping in my breadboard layout.  The mosfet used was an IRF14N05L.

Pm
   
Jr. Member
**

Posts: 63
Quote
On this coil arrangement, the inductance change of the copper winding L1 was ~2% with the 3 turn iron winding L2 varied from 0 to 1.5A dc.  The iron wire I'm using is typical bailing wire obtained from a farm supply store.  Very little inductance change.

@partzman: 
I've had better luck pushing closer to the saturation knee.

Eric Dollard mentions a propagation constant 'Ohm-Siemens' as an energy factor for each verser.  This suggests the amount of change of L or C within each quadrant results in a proportional synthesis/destruction of energy.  2% may simply not be enough to close the loop.

You might try 'paddle wire' used in floral decoration (hobby/craft stores will have it).  Or, insulated garden wire (local hardware stores should have it).  Less core and more turns = easier to saturate.


@GK
Loved your book.  Quite clever and entertaining.
Shields up!  8)
« Last Edit: 2018-07-05, 17:49:35 by Reiyuki »
   
Pages: 1 2 3 4 5 6 [7]
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2018-07-23, 05:03:08