[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: those folks at MIT (fwd)



---------- Forwarded message ----------
Date: Wed, 13 Jun 2007 17:13:46 -0600
From: Terry Fritz <terrellfone@xxxxxxxxxxxxxxxxxxxxxxx>
To: tesla@xxxxxxxxxx
Subject: Re: those folks at MIT (fwd)

Hi Jim,

Soljacic and his ever growing team are theoretical optical physicists.
Playing with coils and electricity is for them, what playing chess is
for an olympic gymnast. If they don't seem to have any idea what they
are doing, it is because they really don't ;-)) If you read
Soljacic's papers on his normal subject mater, he obviously knows what
he is doing there.

http://www.mit.edu/~soljacic/papers.html

They know as much about Tesla coils as we do about.... theoretical 
optical physics :OD  Like a fish out of water, they flop around the best 
they can...

In this case, they reinvent a lot since they really are not trained
in this subject mater at all. They seem to think the energy transfer is
from induction rather than displacement currents. The magnetic coupling
is k = 0.0026... One only needs to read this to understand how much they
know about magnetic or capacitive coupling in this case ;-)

http://en.wikipedia.org/wiki/Evanescent_wave_coupling

http://www.sciencemag.org/cgi/data/1143254/DC1/1

Nice tube in the left in figure 4. ;-))

If you have paper access, you may want to see these:

http://arxiv.org/ftp/physics/papers/0611/0611063.pdf

http://arxiv.org/abs/physics/0611063v2

http://dx.doi.org/10.1016/j.aop.2007.04.017

http://dx.doi.org/10.1126/science.1143254

I did write up an analysis of their system the best I could figure out
and it really should work fairly well.

http://hot-streamer.com/temp/WiTricity!/WiTricity.pdf

It is probably not real safe to be around for a long time and you don't
want to touch anything. Apparently it will radiate enough RF to give HAM 
radio folks a major cow...  Not sure if it is real practical... But it 
is cool!!

I updated the paper...

I just added the last page is all. I resolved that the problem where
their capacitance seemed so low was do to their lack of a ground plane
(there system is floating in free air) so the parallel capacitance to
the coils is about 1/2 what Medhurst would predict.

The power supply source impedance and load impedance are coupled to
their respective coils at about k=0.4. That drastically reduces the
"Q" to about 15 to 20! Having a "Q" of 1000 on the bare coil does not
matter... This very low Q gives them a bumpy 2.5 MHz bandwidth as
shown in the last figure in the paper I wrote. I bet that "dual hump"
thing really confuses them at MIT if they haven't figure that out yet
:oD If not, there might still be time to try and convince them that
the finicky efficiency is due to "effervescent ethereal turbulence" or
some other hilarious cause >:O)

I have been working to make a much more modest ~4MHz WiTricity system
simply run from a common 50 ohm signal generator input and a 50 ohm
resistor at the far end. Waiting for the glue to dry and I hope I can
slide the coils off the form... ;-)

http://hot-streamer.com/temp/WiTricity!/P6130604.JPG

It should run LEDs and be super
easy for anyone to reproduce without the need for 5 MIT Ph.d's to help 
out %:o)

Cheers,

    Terry

BTW - Terry mostly works on technical airsoft issues these days... 
After the PIRANHA SISG coil, I just can't imagine anyway to make a 
better Tesla coil anymore...  But the WiTricity thing did get me winding 
a few coils again ;-)



Tesla list wrote:
> ---------- Forwarded message ----------
> Date: Tue, 12 Jun 2007 11:02:41 -0700
> From: Jim Lux <jimlux@xxxxxxxxxxxxx>
> To: Tesla list <tesla@xxxxxxxxxx>
> Subject: those folks at MIT
>
> Now having read the entire paper, I have some generalized comments:
>
> 1) For such a huge crowd of folks working on it, they didn't actually 
> do much.  Was this, perhaps, something like a senior project with a 
> team of students?
>
> There are a remarkable number of simple quantitative aspects that 
> have been left out.  Lighting the bulb to *nominal* 
> brightness?  Incandescent bulbs are notoriously non-linear. Couldn't 
> they have used a ammeter and voltmeter?
>
>
> 2) They sort of went to a lot of trouble to rederive some pretty 
> standard electromagnetics.  After all, they came up with a different 
> way to estimate the inductance and self C of their coil, when they 
> could have looked in any standard handbook (or used Wikipedia) to get 
> something like Wheeler and Medhurst.  They cited some guy's thesis 
> from 1951 as an example of no closed form equation for inductance of 
> a finite solenoid. And gosh, the coupling of two inductors is 
> something that has been known for decades, if not a century.   How 
> could they not do a first order analysis with something like Ampere's 
> law and Biot-Savart?
>
> Yes, they used a different conceptual model and different 
> notation.  Seems almost like they didn't ever look at the RF 
> literature, at all.  Their reference for the oscillator is a 50s book 
> on vacuum tube oscillators?
>
> Why did they not give the estimated inductance and capacitance (since 
> they obviously needed them)?
> Why did they not give the Is and Id (since they said they measured them)?
>
> Their calculation of "efficiency" is a bit hard to follow.. They 
> calculate power as Gamma*L*I^2, which is sort of working backwards..
>
> How did they measure Q? by measuring the bandwidth with a coupling 
> loop? Did they allow for the coupling of the coupling loop?
>
>
> 3) They sort of don't really understand skin effect, much less the 
> effects of turn to turn interaction in a solenoid.  This is sort of 
> basic NBS Circular 74, Grover, etc.  stuff.
>
> 4) They sort of handwave on the effects of lossy and/or dielectric 
> materials in the vicinity.  Uhh. if it changes the resonant frequency 
> of the transmitter coil, you could change the transmitter frequency 
> with a feedback system (as they mention), but somehow, you'd also 
> have to "remotely" adjust the receiver's resonant frequency to match, 
> or the coupling goes away (as they note).  Likewise, the receiver 
> would have to adjust itself to maintain a constant resonant frequency 
> that matches the transmitter.  This is non trivial with high powers.
>
> 5) RF exposure safety..  they cited the ANSI standard, but I don't 
> think they read it, or understood it, because they make a spurious 
> comparison between cell phones and this system.  They are comparing 
> radiated power (for the cell phone) against field intensity (for the 
> coil system).  For that matter, at 10 MHz, the limit for general 
> population uncontrolled exposure (which this would be) is 0.219 A/m 
> and 82.4 V/m.  They calculate *20cm from the receiving coil* 1.4kV/m 
> and 8A/m.  (let's put this in context.. we are talking about  the 
> field 8" from a 2 foot diameter, 8" long coil)
>
> Being over by factor of 37 for magnetic field and 17 in E field 
> (contrary to their analysis which cites the E field as the problem) 
> does not sound like something that is a minor matter for those 
> engineers to fix up.
>
> 6) They assert that the coils don't have to be the same size, and 
> that as long as the product of the sizes is constant it works.  This 
> is one of those "oops, practical real world losses bite you" 
> problems.  Sure, you can make the receiver coil much smaller, but in 
> order to get the same power out, the losses will need to be 
> reduced.  But hey, I'm sure they're going to consider room 
> temperature superconductors <grin>... That will make the adaptive 
> tuning thing a bit more important, because the Q will be much higher.
>
>
>
>
>
>