Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    • CommentAuthorMrEntropy
    • CommentTimeDec 23rd 2009
     
    I've figured out the current (yellow) trace on the scope in talk1, and it's pretty impressive:

    Firstly, the scope shows the whole trace. This is indicated by the zero marker on the left, and (as alsetalokin noticed), by the curvy bit before the quick rise, because clipping isn't curvy.

    Given that, and what we know about how the motor works, the interpretation of the scope trace is obvious:

    a) the flat part at the beginning after the voltage is applied the inductive current ramp up (really!)

    b) the curvy bit is the onset of core saturation

    c) the quick rise is new inductive ramp up after the core is saturated.

    d) the flat part at the top is the current limit set by the coil's DC resistance.

    We learn a couple interesting things from this:

    1) The permeability of the core material is very high. It's the ratio of the slopes of parts (a) and (c), and looks to be > 1000. The core might be ferrite material W, or maybe something more expensive.

    2) The core is saturated at the top of part (b), well below the limit set by (d). If a series inductor and chopper circuit were used to limit the current to around this value, then the orbo would continue to work just fine, and would _not_ be wasting a lot of energy heating up the coils.

    My eyeball-o-meter indicates that current limiting would save about 99% of I2R (coil heating) losses, and that might make the orbo reasonably efficient.
  1.  
    What you descibe here is a very standard toroidal coil. Nothing special. Nothing efficient.
    But you are right, as I say over and over again, the coil doesn't heat up in a relevant way.
    So the 'heat' argument brought up by Mr. Orbo is crap too.
    The input power ist mainly used to magnetise the core.
    • CommentAuthorMrEntropy
    • CommentTimeDec 23rd 2009 edited
     
    Not exactly standard -- I wasn't able to find a coil like that at any of my on-line shops, although you certainly can have them made, or order the W cores and wind them yourself.

    Big coils that saturate at such low currents probably aren't too useful outside of steorn.
  2.  
    Whatever, it just doesn't help anyone.
    • CommentAuthorMrEntropy
    • CommentTimeDec 23rd 2009
     
    I think that characterization of the device under discussion is more helpful that just about anything else that happens in this forum.
    • CommentAuthorDirtfarmer
    • CommentTimeDec 23rd 2009
     
    Thanks Mr.Entropy...
    ..appreciated here.
    Goodwill,
    -Dirtfarmer
  3.  
    Well for me the people in this forum are much more interesting than any transparent spinning Lego toy.
    But I still think it's pretty easy to replicate this thing. Maybe I'll try in january just for fun.
    •  
      CommentAuthorNoSideSam
    • CommentTimeDec 23rd 2009
     
    If this is the reason for the apparent delayed current trace rise then it is possible that Sean intentionally turned up the supply voltage far beyond what was required to operate to motor in order to make the current trace appear more square. Curiouser and curiouser.
    • CommentAuthorsonoboy
    • CommentTimeDec 23rd 2009
     
    They may be Metglass cores.
    • CommentAuthorsonoboy
    • CommentTimeDec 23rd 2009
     
    To my way of thinking, they should be energizing the coils so that they are saturated just as the magnet passes center and then interrupt the current and recover as much energy as they can when the field collapses. The longer the coils run saturated the more energy they cannot recover.
  4.  
    Well, right, but they have to be powered on for a certain time to let the magnet pass without pulling it back.
    • CommentAuthorjoshs
    • CommentTimeDec 23rd 2009
     
    Posted By: sonoboyThey may be Metglass cores.
    There is no escaping the fact that the KE imparted to the rotor all comes from energy depleted from the battery. Right now their conversion efficiency is extraordinarily low. Improving efficiency by major factors gets them back to levels of very inefficient motors. Forget OU.

    Metglas would be borrowing from Goldes again.

    The waveforms indicate that they do not recover the I^2L/2 energy from the coils. They also indicate the I^2R losses are much larger.

    The scheme requires reducing the attraction between rotor and stator by more than 50% of the rotation between each stator toroid. the only thing the Orbomination can ever be is a very inefficient motor.
    • CommentAuthorMrEntropy
    • CommentTimeDec 23rd 2009 edited
     
    Posted By: NoSideSamIf this is the reason for the apparent delayed current trace rise then it is possible that Sean intentionally turned up the supply voltage far beyond what was required to operate to motor in order to make the current trace appear more square. Curiouser and curiouser.

    They need that voltage to saturate the core quickly. If they were to just divide the applied voltage by 10, then the that initial delay would be 10 times longer. I don't think you can drive this motor efficiently with just a battery and the commutator, although you might be able to do pretty well with just a battery, added inductor (that doesn't saturate), and a freewheeling diode. Then you just make the voltage pulse short enough so that the current never approaches the DC limit.
    •  
      CommentAuthorNoSideSam
    • CommentTimeDec 23rd 2009
     
    Thanks, point taken.
    •  
      CommentAuthorAngus
    • CommentTimeDec 23rd 2009
     
    It seems to me that the Cavorbo is a motor that runs at constant average power. That is why the V and I traces do not change under load. The reason is easy to see. The power from the battery goes primarily to saturating the core of the toroid. The energy Ep used for each rotor magnet passage depends on the length of time the toroid is saturated, which is inversely proportional to rotation speed. The number of passages per second Np is directly proportional to the rotational speed.The total energy used per second is Ep*Np , which is independent of rotation speed.

    The power dissipated by a braked wheel is T*f where T is the braking torque and f is the rotation frequency. If the power available is a constant value EpNp, clearly applying T will reduce f. The duty cycle of the toroid saturation pulse is independent of rotational frequency, since it is set by the geometry of the wheel and the location of the switches. Therefore the power drain when the toroid is activated does not change with load. The V*I is constant. Since V is set constant by the battery, both are constant.

    (Initially posted in the wrong thread)
    • CommentAuthorMrEntropy
    • CommentTimeDec 23rd 2009 edited
     
    Posted By: AngusIt seems to me that the Cavorbo is a motor that runs at constant average power. That is why the V and I traces do not change under load.

    It's not, but the version used in Talk 1 is so inefficient that you can't even see the power that is actually used to drive the machine. In that version:

    99% of the power is wasted as resistive heating.

    The remainder is spent to magnetize the core. This would normally be recoverable, but Sean wastes this, too.

    Of the energy spent to magnetize the core, a (probably small) portion is spent to drive the rotor. This part is not recoverable.

    [EDIT: fixed this part]
    If the initial portion of the current trace were sufficiently magnified, we could see that there is a curve in it as it approaches saturation, because the operation of the core is not linear in the presence of that external magnet. When the magnet is closer, the saturation curve starts earlier and bends more gently. The area between this trace and the magnet-free trace represents the non-recoverable energy spent on motion, because if you recover the available energy from the inductor after the magnet has passed, the recovery trace will follow the magnet-free path.

    As long as the ramp-up is very fast compared to the rotor velocity, then the amount of energy spent on motion in each of these ramp-ups is the essentially the same, regardless of how fast the rotor is spinning, and the power spent on motion is proportional to the rotor speed.
    •  
      CommentAuthorAngus
    • CommentTimeDec 23rd 2009
     
    @MrEntropy
    Hmmm. I think we are somewhat at cross purposes. The power I was talking about was the total power in, which, as you say, mostly goes into resistive heating and the M-H hysteresis. I think that power is essentially independent of speed.
  5.  
    Posted By: joshs
    Posted By: sonoboyThey may be Metglass cores.
    There is no escaping the fact that the KE imparted to the rotor all comes from energy depleted from the battery.

    except of course any that came from Mr. Hand or his friend Señor Air Gun
    • CommentAuthorjoshs
    • CommentTimeDec 23rd 2009
     
    Posted By: duncan torus
    Posted By: joshs
    Posted By: sonoboyThey may be Metglass cores.
    There is no escaping the fact that the KE imparted to the rotor all comes from energy depleted from the battery.

    except of course any that came from Mr. Hand or his friend Señor Air Gun
    If they resorted to an air supply then they should be even more embarrassed. (Is that even possible?) Flux gate devices aren't new.
    • CommentAuthortinker
    • CommentTimeDec 23rd 2009
     
    Very good. Except that the power demand does not go up with speed. It is a constant horizontal line whereas KE is a rising line. Where they cross is where you enter lala land.

    Tinker