Typically the 'Guard' electrode surrounds the two active electrodes to
effectively
short out surface leakage currents and improve signal/noise ratio.
Typical separation
is 2 cm. If placed in between the electrodes it is to ensure the best
possible balance
in the noise pickup to maximize the CMRR.

You can get readymade electrodes (as used for myoelectrically controlled
prosthesis)
from Otto-Boc, or other electrically powered prosthetic makers
(Schmidel?). Check with
your local prosthestist or hospital rehab department for local sources
and options.

The Otto Boc electrode has a single switched output (with adjustable
thresholds) for
controlling open/close of an electric hand so you may be able to use it
directly
as input to mouse switches or the 'mousekeys' program (Microsoft
accessibility option
pgm) with a suitable scan code stuffer.

A now defunct local company (Leaf electronics) made an interface that
allowed
a single EMG electrode to control both opening and closing with a single
muscle site
by use of a graded muscle contraction. A light contraction would open
the hand, a strong
one would close it. I presume that this is now commercially available
elsewhere,
but I haven't checked.

Another local company specializes in building devices to assist
handicapped individuals
with computer control.

http://www.madentec.com/

There are many other companies out there, and most of their engineers
are quite
happy to answer questions about their technology.

Robert.Rolf@UAlberta.ca

> >  Could you explain to me why you position the Ground-wire between the
> >differential inputs ?  I would think that those should be as close together
> >as possible (to get the same effect of outside disturbances on both lines),
> >wedged between two Ground-lines.
> >
> >Most differential inputs cables I've seen where either Coax (two wires with
> >a Grounded shielding), or a simple twisted three-wire cable.
> >
> >Regards,
> >  Rudy Wieser
>
> It is mostly a matter of convenience. The three wires are twisted
> together to reduce common mode noise. The "ground" electrode then
> goes in the center to maintain electrical symmetry.
> You can also surround the two signal wires with a ground, but this
> is a bit messier to implement in the real world. You can
> obviously use coaxial wiring instead of the twisted 3 pair
> wire, but for the target application you do not really gain
> much for the extra expense. Remember that the target application
> is simply to get a *useable* signal from a muscle movement. If
> we were interested in the exact waveform (as in electrocardiographs),
> then the extra expense is warranted.
>
> In regard to the spacing between the active electrodes, it depends on
> the signal source. For the heart you would want them separated
> by at least a few cm since the heart is located rather deep.
> For muscles located nearer the surface you can reduce the spacing.
> Note that if separation of the electrodes is too far, then you get
> a mixture of several muscle signals, and that is fairly useless.
>
> For the target application (handicapped person), you would want to
> find a muscle that the subject has good control over, and then
> isolate that signal as much as possible. One subject might use
> a finger muscle, and another a jaw muscle. It also helps a lot
> in the real world if you can make an electrode carrier that
> can be readily applied and which is comfortable to the user.
>
> Fr. Tom McGahee

> >> The electrode interface begins with an isolated instrumentation
> >> amplifier with two "active" differential inputs and a third
> >> "guard" electrode that establishes the ground reference.
> >>
> >> The "guard" electrode is physically located between the two
> >> differential inputs. This helps to cancel out much of the
> >> common mode signal.