Akula0083 30 Watt Self Running Generator.

[MileHigh]

Lost_bro:

I changed the value of the upper left hand resistor to 67ohms (originally 1 ohm).  With the original value of 1 ohm, the LEDs self extinguish when the stop switch is shorted:   Now (with the 67ohm resistor) when the stop switched is 'Closed' or shorted, the LEDs will stay lit to a certain degree and there is a partial charge returning the the 'Battery' (according to the sim.).

I am not a builder but the wisest course of action would be to replicate to the schematic and parts list as closely as possible, for starters.  I am assuming that eveybody except Itsu will have the same PCB.  Changing a resistor value to see your desired goal is a cart-before-the-horse way of going about it.

I would suggest that if there is an issue with the PWM circuit in the schematic and it needs a small mod, then you all agree on the mod and everybody does exactly the same mod.  So everybody at least starts out with the same functional device to test.  You really don't want anarchy setting in right away.  If you all find that the LEDs don't light when the Stop switch is short-circuited, then you can collectively decide what to do next.  There may be no need to do a kludge to force the "Stop" switch to not shut off the LEDs.  My instincts are telling me that the Stop switch is supposed to stop the LEDs from lighting up when it is a short-circuit.  If that's the case, then there is no need to make a sudden move to hack into the boards.

It looks like a bit of extra enthusiasm without examining the circuit led to the belief that the Stop switch had to be shorted for proper operation.  It's more like you can view the Stop switch as a supplement to the On-Off switch, sort of like a temporary way of turning off the LEDs while the rest of the circuit still runs.  My reasons for saying this are outlined in an earlier posting and the simulation backs the analysis up.

There is no need to doubt the simulation's ability to simulate this circuit.  However, it's sometimes necessary to have some tweaking done on the simulation, simply to get it to more accurately model the real circuit.  Once that is done, the simulation will be deadly accurate.  From what I can see, none or very little tweaking will have to be done to see the signals in the simulation be nearly identical to the signals that you see on your scope display.

If there is "magic" then you will expect to see something out of the ordinary.  If not, then what you see when you probe your boards will be very close to what you see on the simulation.

MileHigh

[wattsup]

@MH

It will be easier for me to answer in bold inside your great post. Thanks for the insight. We will know soon enough though so the only recourse is to stay open to all possibilities.

Wattsup:

Did you play Can't You Hear Me Knocking?  lol

I am so mad about one thing. Yesterday I go buy a Tascam Linear PCM Recorder just to able to record our complete jam session. I was certain is was recording, then when you jam, you never stop so when it was time to check the recording, it did not. So I will have one next time.
*(*(*%*&?*$%#$(&?*%@(*&#_+)$&?*@?&$@(&_?**$&_&@)@+
Some stones but not that one.


For the transformer, one or more gaps will significantly lower the total magnetic flux through the core for a given ampere-turn excitation.  That's why the flyback will be even harder.

I know there is a belief or hope that cutting one or more gaps will do something.  I will suggest something that I have mentioned several times before.  The best thing to do is make A-B comparisons, uncut and cut.  Or if you only have one transformer make a set of orderly measurements before and after you cut the gap or gaps.

So I view the cuts as degrading the performance of the transformer.  That's me questioning the design with my critical eye.  Verpies is the expert on magnetics and he is hoping for a big score so you never know!

I saw that the pulse waveform is somewhere between 16 kHz and about 120 kHz.  That means the transformer will be operating either in signal mode or pulse mode, depending on the state of the Stop switch.  You guys will have to do your builds but if the dots on the transformer in the schematic are accurate then things look pretty grim for the circuit with the Stop switch shorted.

When the Stop switch is shorted, the transformer acts in signal mode.   That means a voltage square wave on the input becomes a voltage square wave on the output.  You can reread my earlier posting but the secondary output of the transformer doesn't output to the LEDs.


Please don't forget that most of the technical or commercial usage of the e-core is with primary using magwire and a good number of winds to achieve some coil resistance. A primary of 60 turns or more of AWG 21 is not uncommon. In the Akula usage the primary is wound with a plastic coated conductor (not magwire) of few turns. Not your typical usage for standard simulation. So question is asked. Does a low turn, low skin effect primary provide a form of dielectric in the primary, since the primary is the side that is driving the load. Analogy: Mosfet open, load the gun (primary and load), Mosfet closed, shoot.

When the Stop switch is open-circuit, the the transformer operates in pulse mode.  You fill the "bucket" of the magnetic core with a pulse of energy, and then the magnetic core dumps the stored pulse of energy into the LEDs.   It's just Bedini revisited.

Ya but who did Bedini get it from. NT TOP. No need to go more.

As far as resonance goes, I don't expect that you will see that.  If you are seriously going to pursue that then you will need to define what resonant mode you are talking about, or observe it and figure out what mode it is.  Be forewarned that resonance can be a power hog, sucking up battery power with no real output.  For example, that's what pick-up coils in resonance on a pulse motor will do.


Hmmmmmm. I would have thought the contrary as at resonance usually amp draw goes down and voltage goes up. When I say resonance it also includes slightly off the resonant frequency often produces better output then when the waveform rises out of the screen so there is a balance to find. We are talking specifically about the Akula E core and how Akula would be using this component with his circuit while we can presume his instructions are not 100%. Lower turn count means higher frequency. No core model in the diagram means just choosing the e-core material will not be an easy task and will require either some luck or a few good coil designers that can use the material tables and other targeted specs, but what are the specs needs to be defined first. To do that you have to fully understand the intended circuit usage so it becomes a viscous circle. hahaha

When you talk about the rebiasing, I don't quite get it but I would suggest to you and the gang to investigate that issue by making a timing diagram by observing multiple scope signal traces at key points in the circuit.  The issue is bridging the gap between the verbal speculation and the bench testing.  If there is any "swing" then you should be able to confirm or deny that by recording and analyzing the timing of the operation of the circuit.

I am putting two pdfs, the second really hits home and may be a set method to help define the  core working parameters. Whatever builders do, please use a known e-core, not a no brand no models type off ebay or one you may have at home without any model identification. You need to work with a known e-core otherwise you will be turning in circles. Once one builder gets some good results, others that may have less good results could consider changing to the same model and if the results improve or equal the other, then you will know that working smart has its merits. 

If there is no verbal speculation that makes any sense, it's senseless to build it. You have to work out all the details before you build, not while you build. You build to produce the effect you expect to produce, even if you expect out of the box, it still has to make some sense. Hard way to live indeed but someone has to do it.

For the copper strip, I suggested the A-B comparison testing.  I don't view the coper strip as being real.

Well, it's in the photo and there is a 5th terminal wire so we have to give it full credit for props hence it should qualify as real enough.

I know that there is a lot of enthusiasm so perhaps a builder and tester will prove me wrong about the transformer and the copper strip.  If the gang could keep the same enthusiasm level going on the testing side and analysis side that would be great.  It's all about scoping the input and output signals for the transformer and then following the signal path to the LED array.

MileHigh

Thanks again for your post.

wattsup

[Hoppy]

Lost_bro:

I am not a builder but the wisest course of action would be to replicate to the schematic and parts list as closely as possible, for starters.  I am assuming that eveybody except Itsu will have the same PCB.  Changing a resistor value to see your desired goal is a cart-before-the-horse way of going about it.

MileHigh

I much prefer to initially build my own experimental circuits / replications using breadboard or strip-board. I have all the necessary components except LED's and just need some time for building between my electronics work assignments. My intention is to begin with a lighter LED load and build-up to the full 30W load when I have the full compliment of LED's. I should be in a position to knock-up this circuit sometime this week and will keep as close as possible to Akula's design and component choice, which is obviously very important when replicating. Having said this, Akula has shown some rather odd component values and placements on his circuit schematic, which suggests to me that he is may be playing games! Purely from an observation point of view, I consider that Wottsup's suggested inductor winding details is a good starting point. I will lay a copper foil underneath the winds.

My interest is not so much to discover if this circuit self-runs - which I'm quite sure it will not - but to hopefully find out if there is anything in its operation that may have excited or mislead Akula into believing that something special worth publishing.

[energia9]

what could happen inside a core or bundle of wire to give back more ,
it never ever gives back more than it was put into it, you can not argue this one..

[verpies]

what could happen inside a core or bundle of wire to give back more?

For example McFreey's beta current pulse or Gunderson's acoustic Villari node

The two conditions (v<0 && dv/dt<x) of TL494's negative feedback loop are consistent with the former, since the physics of beta current pulses makes their polarity negative.
Also, the copper shield would come in handy in collecting the deccelerated fast-electrons, which would otherwise cause heavy negative charge buildup and associated HV insulation breakdown.

Indeed, the usual electronic operational principles prevent the Akula's circuit and transformer from exceeding unity O/I power ratio, unless something unconventional happens in that transformer. 
The conventional behavior of the transformer in this circuit is pretty accurately illustrated by the simulations posted in this thread.