Source Impedance II

I read this article

Is he saying on page 409 that four coils if nested can provide uniform magenetic field in all three axes?

Can anyone make a commen on the paper?

jess

Absolutely.

I did this when trying to understand the effects of stray magnetic field around my "shielded" SMPS inductor. I had a 10mm sided square of wire on the end of a BNC connector, connected to a scope. (Still have it in my "probes" drawer.

Quite informative, you could wave it around the board at various distances and angles from the inductor, see how the shielding effect of copper groundplane varied with frequency for example. Or how far away a sensitive circuit would need to be. I used my 7A22 plugin for the lower signal levels. I could have used more turns instead as you suggest, but I thought the single turn was nicely representative of a "typical" loop area of a circuit and had just got the the 7A22...

Don't you know how to differentiate and integrate sin & cos functions?

You can use a different approach:

The above equations transform into the frequency domain thusly:

V1 = j.omega.L1.I1 + j.omega.M.I2 V2 = j.omega.L2.I2 + j.omega.M.I1

where omega=2*pi*f, and j is the imaginary operator, j^2=-1, j^3=-j, j^4=1

Or into the complex frequency "s" domain, putting j.omega = s:

V1 = sL1.I1 + sM.I2 V2 = sL2.I2 + sM.I2

I think you need to do some reading on elementary calculus ,complex numbers, and Laplace transforms. Essential in electrical engineering.

Correction: V2 = sL2.I2 + sM.I2

-- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)

Correction - second try:

-- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)

Hi,

My first confusion is how can I define my sources. My source is Square wave of 100KHz frequency. The MOSFETS truns ON for 2.5 usec and stay OFF for rest of the time(7.5 usecs) . RL series resonant circuit shapes the input square wave to sine wave of 100KHz. The inductors L1 and L2 gets 400 Volts peak to peak as shown in the figure

So, will it be correct to say that the source is

vL1(t) = 400 sin ( wt + 0) where w = 2x pi x 100,000 (across L1 ) vL2(t) = 400 sin ( wt + 0) where w = 2x pi x 100,000 (across L2 )

v1 (t) = 400 sin ( 628.3 x 10 ^3 t ) + 400 sin ( 628.3 x 10 ^3 t ) ( current entering form the same direction) or same dot convention)

and v2 (t) is the voltage acorss the secondary coil placed in the magnetic field generated by coils L1 and L2 together can be defined

as

v2(t) = 40 sin ( 628.3 x 10 ^3 t ) [ 40 V is peak to peak voltage across the secondary coil ]

Am I right ?

If I am right than the Phasor equations will be

V1 = 400 cos ( 628.3 x 10 ^3 t + 90) + 400 cos ( 628.3 x 10 ^3 t +

jess

This consultant uses single-turn coils to investigate stuff. I frequently re-read his papers to remind myself of some simple techniques to get things done.

Browse the site and look for probes, loops, etc. Quite enlightening.

Cheers, John S

I once posted a plea to get Win Hill to put it in the third edition of "The Art of Electronics", which does imply that there's at least one electronics textbook that doesn't include it.

E.C.Snelling never seems to have put it in any of the mountain of stuff he wrote about ferrites and transformers. I first came across it in a Siemens (now EPCOS) application note for ferrite cores

If you do the right first year electronics course - "The Art of Electronics" was written as the text-book for the "electronics for physicists" course at Harvard, so you can't rely on students being exposed to it.

-- Bill Sloman, Nijmegen

Thanks, I saw that site before and found it very useful. I have not returned for a while so found some new stuff and things I had forgotten too.

You are applying a square wave voltage drive to an inductor in series with a capacitor; if the square wave was high and low for equal amounts of time, it would allows the square wave to decomposed into the sum of a sine wave at the same frequency as the square wave and a series of the odd harmonics of that sine wave (three times the frequency, five time the frequency and so forth) of ever-decreasing amplitude (one third of the amplitude of the fundamental, one fifth and so forth) up to an upper limit set by the switching times of the MOSFETs (typically some tens of MHz)

Since your MOSFETs are are on for 2.5usec and off for 7.5usec - which ones, and in what relationship, you've yet to tell us - this isn't what is happening. But it does tell me enough to let me answer your question below in the negative.

The current going through the "resonant" circuit is going to depend on its impedance - which is going to be different at the fundamental and each of the harmonics - and you probably ought to measure it to get some idea of what is actually going on.

Note that if the voltages applied across the bridge don't have a 50% duty cycle, you will get even harmonics as well.

No.

-- Bill Sloman, Nijmegen

Are you sure? Don't you mean 2.5us on, 2.5us off? The H bridge should reverse every 5us.

I think you mean *LC* resonant circuit.

I don't see where you're trying to go with the rest. If you're trying to derive the coupling coefficient of your "search" coil, forget it, it will alter with position and alignment within the field.

If what you're trying to do is plot the field, the best you can hope to achieve is *relative* measurements.

It's an H bridge, driven at 100kHz, with 2.5 us dead time each side. That's 2.5us on, each leg, with a period of 10us, and a 5us delay diagonally. That's a PRF of 100kHz, with 50% pulse width. I think there's too much dead band, 1us should be more than adequate (80% pulse width) to avoid shoot-through.

There have been more than enough LTspice schematics posted, by at least three people, to illustrate this.

The current in a series resonant circuit depends upon its Q. Its impedance will be purely resistive at resonance.

With the claimed Q of around 100, and 2.5us dead time, second harmonic current will be approaching 40dB down, third harmonic, nearly -50dB. Reducing dead time to 1us should push harmonics below -60dB.

That's better than a lot of "instruments" I've seen.

oooh you stepped on yourself; t-period is 10 us. ?-(

will

Guess I sorta did. The way I defined the sources in my model was: period=10us, Ton= 5us-Td, Tdelay=5us on opposing diagonals. Td (dead time) made steppable. I tend to think in terms of dead time when describing bridge excitation.

PULSE(0.85 11.05 0 25n 25n {5u-Td} 10u)

PULSE(0.85 11.05 5u 25n 25n {5u-Td} 10u)

To me, that's 5us minus dead time on, 5us plus dead time off, reversing on alternate half cycles.

I suspect physicists are expected to know it from physics. It's in "Electricity and Magnetism" (Bleaney & Bleaney). Remember that? It used to be Holy Writ on your side of the Pond.

You've got to keep in mind that I've no formal training in electronics

- I wing it on the basis of first year undergraduate physics and a great deal of reading - more of it in physics/science libraries than engineering libraries.

I used to worry about it, until I started having to supervise university trained electronic engineers - the good ones occasionally lent me their lecture notes (which they no longer understood) when I was tackling something out of the ordinary but they never knew stuff better than I did. Good for my self-image, but I'd have preferred to have been able to delegate a bit more.

-- Bill Sloman, Nijmegen

(dead time)

on

Happens to all us fallible humans. Oh well.

?-)

Robots?

I think you mean *electronics* engineers ;-)

Not in my dialect.

-- Bill Sloman, Nijmegen