[FoRK] Annihilation from within...

Dave Long < dave.long at bluewin.ch > on > Tue Dec 12 03:05:46 PST 2006

> Metal vapor magnetometers are being uses in archaeometry, and the MEG
> application is eventually expected (=I don't know, that's what the 
> researchers
> claim) to achieve a better sensitivity than SQUIDs.

That's interesting[0], because I'd thought SQUIDs, as they're measuring 
stuff down near the quantum level, don't have a lot of room underneath. 
  On the other hand, better sensitivity than SQUIDS, at temperatures 
above 4 K, is certainly plausible.

>                                I know that these things are being 
> operated
> in the wild (in the field, literally), so I presume you can remove 
> magnetic
> field background (50/60 Hz, etc.) by DSP.

If the background can be removed by DSP *after* the measurement, that's 
a big advantage.  The problem operating a SQUID (that I tried to 
handwave last post) is that one needs to have an accurate feedback at 
all times.  It's a precision instrument: if it isn't kept close to its 
null point, it quits measuring[1].


:: :: ::

[0] and an interesting application.  It turns out one can decide fairly 
well whether a ring of stones was used as a fire-pit or not, because a 
fire will tend to bring them above the curie point and cause them to be 
at least partially remagnetized according to their alignment in the 
ring.  (the paleolithic version of hard drive tracks?)

[1] too much flux, and the ring goes resistive.  No superconductivity, 
no quantum interference, and instead of a S.Qu.I.D. one is left with 
just a "D" -- a fancy resistor.

See "6.3  Magnetic Flux Trapping in SQUIDs" from the excellent

Mr. SQUID User's Guide

(see also: Figure 5-10 Magnetic field levels of various sources)

However, the MSUG also mentions that it's possible to use gradiometer 
configuration flux transformers to reject the low frequencies -- 
substituting geometry of superconducting materials on the input side 
for DSP on the output side.

:: :: ::

>> guess as to why the pictures in the PDF only show the device fitting
> Can you send me that paper?

It's the one Jeff mentioned:


or cf the picture at:


(both of these articles also mention the inverse problem, which is my 
other beef with naive reliance on DSP.  If the inverse is soluble, but 
tough, heavy DSP will help.  If not, GIGO means getting a false 
solution faster.  This is why I was thinking of steep gradients, with 
the idea of using a synthetic-aperature approach: the better the focus, 
the easier the inverse.

There is said to be a very strong "oops!" signal that fires before one 
even consciously realizes that things have gone awry -- but it comes 
from the vicinity of the brainstem, and hence isn't very accesible to 
surface or exterior sensors.  Is it possible to focus a sensor array to 
pick up subcortical activity?)

More information about the FoRK mailing list