RF prototypes/ ARRL handbook

It looks cute but the pictures show a lot of slop in the magnet position so I doubt it will achieve anywhere near that field map. Not that I can eyeball sub-part per thousand differences from a map like that but I've spent some time playing with arrays just like that and testing small position and tilt errors. Also, given the size of the "bore" opening compared to the length of the magnets pretty much only a sub, sub millimeter thick slice across the center will be homogeneous. The distance from the closest side of the homogenous region needs to be at least two (three is better) bore diameters from the end of the magnet segments to limit the effect of field falloff as you approach the ends. Again, percent is easy, sub part per thousand gets hard. You could probably shorten the magnet length with some form of end cap but then you lose the clear bore access. After doing a lot of comparisons I decided that just using two discs and having side access perpendicular to the magnetic field gave the same or greater field strength and homogeneity over an appreciable volume for the same cost and much easier construction. I'm not a magnet physicist, this is worth what you are paying, YMMV, etc., etc. :-). Other things I played with: making the discs thicker doesn't get that much more field, using a yoke of flat plates across the outer faces of the discs with one or two side panels gave more field but added a lot of weight and didn't really change the simulated homogeneity so it was a wash compared to just buying thicker discs, and trying a few simple pole cap shapes helped the homogeneity but lowered the field since the magnets were now further apart for the same clear gap. I was bored, curious, and had a pc and FEMM so I explored :-).

--
Regards, 
Carl Ijames

Maybe one could solder a bunch of unshielded drum core inductors to a PC board and glue a couple of them to the front or back of your disk magnets. Make an active shim system. That would allow George's system to be a shimming trainer too.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Now _that_ would be fun. If I ever retire, I hope to do that sort of stuff--though you'd probably be limited by nonuniformities in the permanent magnets, which wouldn't have the nice smooth variations of an air core electromagnet.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Yeah something like that could work.

The signal is proportional to the polarizing field, so the bigger the better.

Thinking more.. How about we figure many users will have their own B-field. And we make a cheap gizmo to stick in the field and see an FID*. I was thinking I can drive it with a Sig gen in burst mode... maybe without an intermediate amp? And then I have to amplify the ~1uV signal up to scope levels.

George H.

*free induction decay

I was just thinking about the yoke as that big hunk of iron that closes the magnetic loop around the pole area. like this.. my first google hit.

GH

Nice! I like the side by side 180 out of phase samples! (the 180 out of phase cancels the interference noise.) We sell an Earth Fields NMR

Instead of side by side coils it uses one good sample coil, and a bigger, (but of the same turns area) co-axial cancellation coil. seen in pic. The sample is a plastic bottle, beer can size.

George H.

rote:

ks.

t t

I don't think so... but could be wrong. The figure of merit is field homogeneity over field strength. There are electromagnetic solutions for like a cylinder cut into magnetic material... Which implies I could give up total field for homogeneity.

George H.

Right... So I want to see a signal on a 'scope. Say a 10 MHz freq. with a field homogeneity of

1ppt 0.1% over the sample I'd get 10 kHz... uggh! I guess I'd like the signal to last ~100 times longer than the Q of the circuit. (So it's harder confuse ringing with signal.)

George H.

I

nd

ort

mit

so

eter

e

y,

.

e

s is

a

side

e

er

t

).

This is probably Gauss's theorem #8, but I find that you can cancel any inhomogeneity to first order, by adding the right dipole field (magnitude and direction.) In practice this means you can wave a magnetic around in space an find one of the two sweet spots for it. (given you have some signal that changes with homogeneity.)

George H.

The Geller kit is inspired by the proton precession magnetometer article in the Amateur Scientist section of Scientific American of years gone by.

Apparently it managed 20 nT resolution, and used tubes.

If your physics type is interested it's in the Amateur Scientist CD:

I think this is the book the CD is based on- it's searchable online:

...or you can download it, which I just did.

Mark L. Fergerson

ks.

d

rt

it

r of

in

ding

nd

high

for

air

lume

but then you'd need to support multiple frequencies

I designed the controller for a big FTMS.

The preamp was designed by some chemists,

and they gave up 30 dB of s/n or so. It was ghastly.

I thought it might be possible to detect a single ion in orbit, but we never got to try.

Varian acquired Ion Spec, then Agilent acquired Varian, then Agilent kill off the NMR and FTMS product lines.

I have some planned preamp schematics around here somewhere.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

This is pretty good:

Very nice write-up! I will distribute it to my RF Lab class.

I didn't look at all the pics.

I've been 1/2 thinking about this. For a single coil system, I'm thinking I need some sort of switching between the two circuits, flip the spins and then see the signal. Which is doable, but another thing... A dual cross coil system has some advantages. Flip with one, sense with other. The two functions are naturally apart. That makes it easier to test when something is not working. (which if you are contemplating some marginally functional kit is very important.)

George H.

Saved, thanks. I will not disparage your lack of electronic posts/content for X-months.

Got any other good stuff/ links? George H.

something like this works ok,

tweak Cmatch and Ctune for resonance and 50R match using a VNA

you can do the quarter-wave with a length of coax or a CLC pi section for moderate powers 1n4148 diodes will work

The down side is inter-penetrating coils are hard. so the flip coil is often put outside the other, and thus has to be bigger...

You could make some water sample that flowed through two coils. That would be fun... but more complicated.

If you are thinking of two coaxial coils, I think it's a bad idea... there's all this 'transformer/ flux' pickup in the sense coil*. You still have some of that in cross coils, and back to back diodes across the input will limit that.

*manufacture-wise two coaxial coils right on top of each other is dirt simple. Maybe I just need the right protection diodes across the sense input?

GH

Awesome. I think that might be what Norman did on our latest nmr thing. Norman was part of W-map at Princeton. Smart guy!

The current thing has two caps to adjust, ones called coupling and the other tuning. Tuning it up is a skill, takes time... and it's not made better by crappy tuning caps.

George H.