First a bit of a story. Our optical pumping apparatus has a pair of Helmholtz coils. (shown here,
formatting link
Something like 5% of the coils just don't match well*. (This can be for various reasons, wrong # of turns, mechanically mis-aligned, or other.) It's not that hard to find a turn mismatch. (~400 turns per side so a small resistance difference) But sometimes things are just a little off for no apparent reason. (These coils go on the shelf.. to be played with when I retire. :^) Anyway last night I was testing some and again had some 'just a little off' coils. I was thinking of trying a circuit 'fix'. Shown here,
formatting link
For which I need floating current sources. I think first of a three terminal V_ref, resistor, opamp (to V_ref common) and load. Like this,
formatting link
Is that going to work? (Battery powered or something.)
George H.
*At high fields (several Gauss) the signal line width is determined by the field homogeneity of the coils. Excess line width means smaller signals and un-resolvable signals.
Right, I should just do something simple to make sure the idea works. So maybe a floating power supply and resistor on one coil.
When I'm feeling finicky, I'll trim the local field inhomogeneity by waving a little magnet around and looking at the signal. Once I find the 'right' orientation I'll tape it to a wooden chair. This is only a temporary fix.
George H. (I have to take these to my 'country estate' for testing. B fields in the city are all over the place and changing constantly.)
In the past, people were better than automation at shimming NMR magnets. I think the algorithms have got better lately.
I might get one of those surfboard log-periodic antennas and a portable spectrum analyzer to do rough EMI testing. But I'll have to do that up in the mountains; the city is an EMI nightmare.
Cell phones work badly up there, which is a good sign for EMI surveys.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Years ago I took the optical pumping (OP) out into the middle of my field, (yard) to get away from 60 Hz fields. There seems to be this 'real' effect that the zero field* OP signal gets smaller when you get really close to zero B field.
George H.
*OP depends on the Zeeman splitting of the ground state electrons. If there's no B field there is no splitting, no pumping.
That is with and without a current stealing resistor. It's a bit of a cheat, because the much wider 'without' is the second in a string of Zeeman levels... but it just happened to be close, and so the width is right but the first is ~twice as deep.
Here's the 'full' spectrum
formatting link
There's a lot going on there. The four big wide things are the main transitions (over driven and RF power broadened) in between are three double quantum transitions (DQT). There's also two smaller signals at the beginning.
The depth/ sharpness (of DQT) is what I want. TBH it's never been better.I took some crappy coils and made 'em the best ever.
I'm going to call it Jan's current stealer in the news letter. (unless you object) Hmm, it may also be tuning out local field gradients... doing more than just 'fixing' the coils.
I bought an "estate groomer" (triple mower deck that runs off my tractors PTO) from the local Massey-Ferguson dealer. (Java Farm Supply) and since then I figure as long as I keep grooming it, I've got an estate. :^) (well mostly trails, some yard.)
(I'm just trying to flatter you so you'll keep feeding me ideas. :^)
No, The coils are driven from a current source. I'm guessing it's just slight differences in how the layers of wire go down. Maybe there is some wire tension difference or something.
Anyway this is all a very small effect. The (theoretical) field inhomogeneity for Helmholtz coils goes as C*(x/R)^4. Where C ~1, R is the coil radius and x the distance from the center. For my coils R = 6" and the cell size is 1" so x ~1/2"... (x/R)^4 ~ 20,000 the Field is ~7 Gauss so best inhomogeneity ~0.35 mG (milli Gauss) I typically want to see less than a 1 mG linewidth, And for the coils with Jan's current stealer the I measured 0.4 mG!!
Right if I ever start another thread with some complicated circuit fix and you think, "Gee George, how about a capacitor instead?" Then please do speak up. :^)
Oh yes... which is what makes working in the city hard.. the field won't hold still. So there are two sets of coils, one vertical and one horizontal. We first orient the Axis to lie along the Earth's horizontal component of the field. Then use the vertical coils to tune out the vertical field. And then sweep the horizontal field.
Cars, trucks, elevators, forklifts.. any hunk of iron moving around. The iron in buildings, I-beams, rebar and such also guides the local fields in weird ways. (There's very little horizontal B in my lab... and it goes in many directions. Old building.) And then there are the 60 Hz fields, those vary but a few mG is common.
Yes, this is how 'I' think it works, and what made me think about the parallel resistor:
Being in a large part a RF guy, I was thinking about Q (no not from the bond movies).
Any L has associated with it a capacitance, and the impedance is always complex. So you have 2 LC circuits here.
Since you are driving with a triangle wave, the harmonics go way up into the clouds, no matter how slow your driving frequency is (Fourier remember).
Basically Q = w.L / Rseries and bandwidth is B = f/Q
Putting a compensating resistor is series would work too, but then you need a very low value potmeter.
But the relationship between Rseries and Rparallel makes it possible to use a parallel resistor to make Q equal.
So it is basically (edit your paper) a 'Q balancing' circuit.
'If it was me', and I am curious, I would, and there recently was an interesting thread here about measuring very low resistances, measure the Rs of those coils.
Hmm, well maybe it looks like that 'cause the signal got so much sharper. But this is a DC trick. There is some residual magnetic field gradient in the z direction (axis of coils) the gradient means that spins on one side resonant at a slightly different field* than spins on the other. (Spins are electronic spins of Rb atoms on different ends of the 1" Rb cell) Stealing a little current from one coil causes a compensating field gradient. And viola a nice sharp resonance. (Stealing the current also causes the transition to move a bit in terms of the absolute current... which is what I really measure/ plot.)
George H.
*I'm sweeping the field at fixed RF frequency, one can also sweep frequency at fixed field.
Hey, I think you said you are a Penrose fan. I saw he was on Joe Rogan's podcast.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.