Steorns December 2009 Demo

[poynt99]

This should further help to explain the blue trace, and why it is actually "folded-over" in terms of the amount of energy it represents and where it is going.

.99

[poynt99]

A comparison of Steorn's original blue "net electrical in" trace, and two cycles of the same trace unfolded to illustrate a more realistic energy profile supplied to the circuit.

By this, it's quite clear that there is far more energy going in to the circuit than what the pickup coil is generating.

.99

[poynt99]

I sent a message to TK via his Youtube account, and this is an excerpt from his response.

Yep, I just looked at your scope sims. You are right, of course, and you are basically saying the same thing that I have been saying from the beginning. You are saying it much more clearly and your sims are very convincing, where I was just trying to get a qualitative description clear in my own head.

The problem, though, is that Steorn is claiming that all that heat power is not supposed to be counted, when comparing input to output, so they are subtracting it from the integral. As long as they are allowed to claim that the rotor is being powered by (V*I - I^2R) rather than simply (V*I) the total input power, they will claim OU, since the rotor "logically" is only powered by that input energy that does NOT go to heat, either as direct resistive Joule losses or the more indirect dissipation of the flyback pulse that you have explained so well.

But I maintain that you cannot separate these two portions of the input at all, because the I^2R loss part is a necessary overhead, and there isn't any way to just drive the rotor by injecting some (V*I - I^2R) remainder into it somewhere. You need the whole shebang, even if most of it is lost to heat.

Keep up the good work; there are folks on moletrap that would also like to see this so I will post a link.
--TK

TK, all,

Is there any evidence that Steorn is in fact subtracting the I²R losses either in the real-time scope trace or afterward in post-processing?

I have seen this mentioned at OU as well, but I really doubt that they are indeed doing this. Why? Sean clearly indicates that by extrapolating/comparing the two traces shown on the scope, the COP can be approximated to be about 3, AND, I have been able to closely duplicate that blue trace wave form without performing any other math on the trace.

If my hunch is correct about Steorn measuring the coil voltage rather than the power supply voltage (and I believe I am), then they are in fact measuring ONLY the I²R losses, and of the coil alone. This represents the majority of the energy being sourced by the supply, the remaining energy is what is being dissipated in the MOSFET switch (assuming they are using one) and the circuit wiring as I²R losses, and some Joule heating in the coil cores. The output energy from the rotor momentum and the heating generated by the pickup coil, represent a relatively small portion of energy compared to what the power supply is sourcing to the circuit.

What Steorn is presenting and claiming about that blue trace is incorrect. It is not the "net electrical in" at all. In fact, it is part of the net electrical out to be correct.

If analyzed properly, it becomes obvious the device is grossly underunity.

.99

[poynt99]

A message to Max:

Get yourself away from Steorn as far as possible, and as soon as possible. OU may be out there, but it's not with Orbo.

You appear to be a young, bright and aspiring electronics wiz. Use your common sense and decide if you're on the right path. Filter through all the nonsensical techobabble and embrace some healthy objectivity.

The posts here and TK's videos have hopefully given you something to think about.

Cheers,
.99

[MileHigh]

Hey Poynt,

It's me!  lol

At this juncture do you see Steorn's proposition as being if you take away all of the Joule heating (pre and post switch-off) that they are somehow imparting more rotational energy in the rotor than is "supposed" to be there.  Then they can drain off that energy in a pick-up coil and claim "over unity"?  That's what it looks like to me.

During the pulse there is a "sliver" of (i * v) tha flushes out the "badly aligned" domains from the core.  That's ultimately what imparts energy to the rotor.  They are claiming the area of the sliver translates into more rotational energy than expected in the rotor.

Of course this is all Steornian Orwellian steornspeak, because for years we have all been thinking that the claim was "true" over unity.

MileHigh

[poynt99]

Hey Poynt,

It's me!  lol

Nice to have you back  

At this juncture do you see Steorn's proposition as being if you take away all of the Joule heating (pre and post switch-off) that they are somehow imparting more rotational energy in the rotor than is "supposed" to be there.  Then they can drain off that energy in a pick-up coil and claim "over unity"? That's what it looks like to me.

I actually have no idea what Steorn's true position/argument/claim is to be honest. Everything about them is "muddy". What you have said there could very well be their position indeed. What is the point though? That input Joule heating energy still has to be expended to "free the magnets", so in the end what is gained?

What I believe Steorn is precariously placing all their money on, is their misinterpretation of their misrepresented input energy measurement, nothing more. They truly believe that:

1) The blue trace is a measurement of energy supplied by the source, and
2) The blue trace is proof that the net energy expended (apparently noted at the end of each cycle) is 3 times less than what their pickup coil is burning off in its DCR.

Wrong on both accounts!

During the pulse there is a "sliver" of (i * v) that flushes out the "badly aligned" domains from the core.  That's ultimately what imparts energy to the rotor.
They are claiming the area of the sliver translates into more rotational energy than expected in the rotor.

Of course this is all Steornian Orwellian steornspeak, because for years we have all been thinking that the claim was "true" over unity.

MileHigh

I believe the process is very simple:

1) The rotor magnet is attracted to the core (and the core becomes partially polarized).
2) At TDC the coil is pulsed which fully polarizes/saturates the core and allows the rotor magnet to wiz on out of TDC and merrily on its way to the next coil.
3) The coil pulse is released between 20º to 30º after TDC, allowing the core Weiss domains to partially "relax" and re-orientate.
4) Repeat.

2a) Cores with high permeability and high "squareness ratio" are ideal for saturable core applications, and with the core already partially polarized by the rotor magnet, full saturation occurs quite quickly at the beginning of the pulse. The coil tau goes from a value of perhaps 200ms to one of about 30us in a very short time, and this accounts for the very minimal "inductive rise" shown in Steorn's scope shots demonstrating this effect. Of course along with this abrupt decrease of inductance comes an abrupt increase of supplied current to the coil, and full Joule heating begins immediately.

3a) Upon release of the pulse, the coil current begins to diminish and the core's Weiss domains return to a more "neutral" position, causing a gradual increase in the coil inductance in the process. Here we have a tau transition that works in reverse, beginning with a very short value and ending in a very long one.

I hope my simulation will be able to answer any and all unanswered questions, and confirm what is now known about Orbo.

.99

[vpo20]

Hello everybody. Steorn motor works with no so called "back EMF" simply because it uses parametric method of driving. Namely the magnetic permeability of the ferrite core is deeply changed due to magnetic saturation caused by applied coil current and superimposed orthogonal magnetic field from flying magnets. If you saw all presentations you may remember them measuring the inductance of coil in absence and presence of magnets. It is 1 Henry and 40 microHenry. Principle of driving is the same as driving children's swings / pendulum when rider stands quickly up each time when swings passes middle point and sits down down at extreme points of deflection. Electrical resonance circuit also can brought into oscillation if capacitance changes twice per one period of oscillation. It called parametric resonance look for instance in http://en.wikipedia.org/wiki/Parametric_oscillator. Standing up and down changes centre of mass / length of physical pendulum at frequency double of resonance frequency.
The Back EMF rule is not a low and it only true for linear networks with fixed parameters.
Here exact orthogonal orientation of magnetic field in respect of soft magnetic core totally de-couples traditional magnetic link but still retain parametric coupling. On approach magnets are drawn toward soft core but at the lowest position ferrite core become not magnetic  due to circular field saturation caused by current switched on at this moment so that magnets are not attracted any more and easily fly away. One may easily make not a rotating motor but a swinging pendulum driven the same way. Then the analogy to children swings become very accurate.
There is no gain of energy there I am afraid. Energy is spent to magnetise core and loads of it wasted. If you load motor to get energy out then rotation speed goes down and as a result you has to keep magnetisation longer then without load thus current to be kept longer  and waste all energy you gained from rotor  with excess. No miracle. Sorry..

[poynt99]

Hi vpo20,

I believe we are saying basically the same thing.

.99

[poynt99]

This is from the demo Steorn did showing the flat top current and voltage proving no bemf. They also showed that there is no delay of the current, as it really is an apparent delay.

Here are my voltage and current wave forms so far to match theirs. This is made with a modeled core material with 400 turns and a 8 Ohm DCR. That little "nib" at the top left of the blue voltage trace is more evident while the Orbo is getting up to speed, and its preponderance is a function of the variable resistor setting.

Starting to get close  .

.99

[darkspeed]

Dam poynt99! that is getting close.. good work!

But your missing the magic finger in the scope shot 

[poynt99]

My Orbo Simulation circuit as of 2010-02-23:

.99

[poynt99]

Another slight variation showing the negative energy excursion.

This is with a linear (i.e. air) core, L=20mH, DCR=0.5 Ohm, Rlimit=2 Ohms, Vsupply=12V, and flyback diode not present.

.99

[poynt99]

For comparison sake, the above with the flyback diode present.

.99

[poynt99]

Another, just showing the integration trace itself.

Air-core and no flyback diode.

.99

[poynt99]

A very interesting side-by-side comparison of the Orbo V, I and J traces vs. the air-core/no flyback diode version of the same. Notice the peak Icoil is about the same in each. All scales are equal for good comparison.

.99

[poynt99]

Investigating the flyback diode energy vs. the coil energy as a possible source for the big negative excursion in the integration trace...

nope 

.99

[poynt99]

Looking at the notion of the integrated trace being "folded over", this may not be 100% accurate, as per below with the air-core simulation.

Clearly the supply does not expend as much energy as what appears in the neg excursion. This isn't energy from nowhere though.

More to come on this...

.99

EDIT: See next post for correct/complete analysis of energy in the circuit.

[poynt99]

OK,

The energy balance has been reconciled, and I'm happy to report that it all works out as expected.

Of note is at kickback the almost complete transfer of stored coil energy to D1 and RLimit. That is the big neg excursion. Indeed what is left is the net energy burning up in the coil...so what? Very little of the transferred energy gets back to the source, and most of it is burned up in D1 and RLimit.

.99

[poynt99]

From TK:

Well, that seems reasonable. So are you saying that the Steorn "net of Joule heating" is equivalent to the red Jcoil trace in your simulation?

Could you give an interpretation of your findings and how they relate to the Steorn trace?

That's pretty much what I'm saying TK. Net of Joule heating in the coil. Not net of input Joule heating, or whatever Sean said regarding the input.

BUT! With a functioning flyback diode in place in my sim, this all changes, as I've showed in previous posts. Rather than the coil transferring its energy up through to D1 and RLimit, a flyback diode rectifies the coil current and it gets recirculated, as you know. The Ainslie circuit functioned somewhat similar with the flyback diode removed.

My interpretation is as I've been saying for a while here. I can reproduce something very similar to Steorn (as you can see above in the red trace), but not with a flyback diode in circuit.

What I think needs to be in place is a coil having a very large inductance to resistance ratio. This is a must if it is going to be used as an energy storage/transfer device as seems evident if my sim is providing any insights at all.

So that means a REALLY HIGH permeability core, and heavy wire/low turns. I am suggesting an initial permeability spec of at least 200,000, if there is such an animal readily available.

A place to start might be to try and make a 300mH coil with the same length of the same gauge wire you have on your 20mH coil. That would mean a higher mu core.

This still leaves the flyback diode as a problem. I'd like to see Orbo's integration trace run through a math differentiator. I hope someone is taking that up to task.

.99

[poynt99]

"The purple trace, is the net electrical energy put in to the system...ah net of Joule heating".

Those were Sean's exact words. Really poor wording, and is ambiguous as to the true meaning.

Translation: The purple trace represents the amount of total energy supplied to the system by the battery. This is also equivalent to the net Joule heating in the coils.

Well, only the second sentence of my translation is true.

.99

[poynt99]

"You can work the ratios, it's about three to one, input to output...that is simply direct electrical in to out, net of Joule heating".

Good grief!

Again, part two of that blurb is false.

.99

[poynt99]

From TK:

Great work, and thanks for the interpretation, .99 and MH. I am still going to have to think about it before I can claim to grok the fullness..er...understand completely what's happening and how to make my real system behave this way.
Today's my shopping day; I'm riding the subway downtown and that will give me plenty of time without distractions to think about it.
But...if one uses a really high perm core, then...one is still trying not only for a low DCR but apparently also a relatively low inductance as well, because otherwise the pulse rise time suffers...hmmm. Maybe Orbette 1's LTT toroids were accidentally more optimal than I thought, since they have been the lowest DCR/highest inductance combo that I have yet tried....
Sonoboy, IIRC, is using a Metglas toroid with really high permeability and I remember him mentioning that he actually took some turns off to reduce the inductance, which in his system would make the rise time faster and allow more energy in per pulse. I think.

OK, now one must ask: in spite of Sean's clever malapropisms, does his premise make any logical sense? Does the displayed trace, the red one, "net of Joule heat" in the coil, or whatever, does that legitimately correspond to the actual energy transferred to the rotor? And, further, if Sean is correct about the energy per rev out from the "sense coil", is it a fair comparison to make: (Actual energy transferred to rotor :: Energy out from sense coil) ?

So now I wonder if it's possible to model Sean's Orbo output measurement in any realistic way.

[poynt99]

TK,

You still want to go for high inductance; as high as possible for as few as possible turns/DCR. I can probably do even better than Orbo by increasing the L/R ratio of the coils. The rise time is not an issue with high inductance and a non-linear core, as long as it is driven well within its saturation range, which is part of the Orbo "secret".

The idea I believe is to make the core act as an inductance "switch" (square B-H) with a relatively low threshold (low Hc).

There's one problem with a low DCR however, and that is high I(coil). This may indeed be why Steorn is using the rheostat and changing it when switching the drive circuit between the conventional pulse motor and the Orbo motor. This however goes against the logic of voltage drops in the circuit with a high R(Limit) value and low coil DCR. The voltage drop would be excessive. If the R(Limit) is kept low, then coil current will be excessive (assuming core saturation is reached).

This is a constant occurrence for me in trying to crack this thing. Certain elements of the design adjusted a certain way make perfect sense in one aspect, but are in direct contravention with other aspects, and vice versa.

To paraphrase Sean: "It's all in the structure and arrangement of the coils". Sounds like Steven Mark talk.

Regarding the Steorn premise:

...does that "net of Joule heat" [red trace] legitimately correspond to the actual energy transferred to the rotor?

IMO, no. That net energy is ONLY representative of the energy lost to the resistive part of the coils. Let's pretend for a moment that we have a super-conductor coil in its place and its DCR is effectively 0 Ohms. The red trace would look very similar but instead the neg excursion would return back to 0mJ every cycle. So what? It still took a fair bit of energy from the source to create the magnetic field used to polarize/saturate the core. The energy stored in the coil is then transferred to the other circuit components (D1 and R(Limit)) as before, but this time with 100% efficiency (as opposed to the 83% shown in the red trace). So in this case the "net of Joule heating" in the coil is nil. Will this translate into higher rotor RPM or torque? IMO NOPE. What it DOES mean, is we can use a little less energy from our source to saturate the core to the same level as before.

And, further, if Sean is correct about the energy per rev out from the "sense coil", is it a fair comparison to make: (Actual energy transferred to rotor :: Energy out from sense coil) ?

It's absolutely not a fair comparison at all. From the blue integration trace, I'd say the Orbo coils are achieving an energy transfer efficiency of about 95%, which is probably as good as one can achieve without bringing the wire temperature down. If we could somehow make the n=100, then Steorn would want to claim a COP = infinity. That is completely false because we know it still takes energy to create that magnetizing force. Now, if Steorn had super-conducting wires throughout, and a super-conducting SSR switch, then in combination with their super-conducting coil (and no core loss), they might have a system that was OU. However, we know that IF we had RT super-conductor wiring available to us, there are far better ways to make energy and achieve OU. In summary, no; the net Joule heating in the coil and the energy being generated in the pickup coil are not related at all. I think most agree that the only fair comparison is energy in (i.e. from the battery) vs. energy out. If this was done, Orbo would be seen as grossly UU.

Steorn has demonstrated a few different ways, that apparently the coil energy and rotor motion is "decoupled", i.e. there is no bemf, and no greater current required to energize the field with the rotor mag at TDC. If we buy into this notion (and I am willing to at this point), then I can not simulate the rotor as an energy take-off from the system. I believe I can treat the coils as if they were transformers though, and load them that way, but I don't think that's what we want to do.

.99

[poynt99]

A look at the Orbo coil energy transfer efficiency.

In my sim, it's about 83%.

Depending on which point is used to measure the Orbo coil efficiency, it is either 86% or about 93%, as per the attached graphic.

.99

[poynt99]

From TK:

To make a valid comparison following Sean's logic, one would compare the output energy per rev, measured as the I^2R power of the shorted output coil integrated over the rev, with....what exactly?
If Sean's comparison isn't legitimate, then how would one go about measuring, electrically, the actual energy transferred to the rotor per rev?

The only comparison that can be made is of the input to output energy. This is how it should be done IMO:

1) The circuit supply voltage is reduced until a slight reduction in RPM is noticed in the rotor (no load condition). This is the minimum energy required to operate the Orbo motor at full speed.

2) Take a supply energy or power measurement.

3) Measure the energy or power in your shorted coil load or loads (more than one should be tried).

4) Do the COP computation.

As I see it:

The Orbo motor is not "supply-driven" like a normal pulse motor is. Once a certain minimum energy is sourced to the Orbo, any further increases are not reflected by a further increase in either speed or torque at the rotor. A normal pulse motor operates quite the opposite; the harder we drive the coils, the more force is transferred to the rotor, and the more torque it generates. Orbo doesn't operate this way.

The only way to obtain more torque and speed with an Orbo setup is to use stronger rotor magnets, and/or higher mu cores. In each case, a little more energy will be required to de-polarize the core at TDC. This is the only relationship between the input energy and resulting rotor torque/speed. Of course with stronger magnets or a higher mu core, the pickup coil(s) will be able to generate a little more energy also.

.99