Re: [buildcheapeeg] Re: modular eeg work report

From: Chris (e9300298_at_student.tuwien.ac.at)
Date: 2002-07-11 13:51:56


sleeper75se wrote:

> > After the downloads (eagle-cad, avr-gcc, SP12), printouts, PCB,
> > orderings (I got most of the parts here in vienna, 1 capacitor 0,3u
> > and 4 'Spulen' 0,22uH have slightly different values) I am finished
> > with my
>
> Do you mean 22uH?
>

yes ;ö)

> > I have problems with 50-Hz humm (I can reject it to an amount that I
> > actually see brainwaves only when i connect myself to GND, can
> > somebody help here ?
>
> Do you mean ModularEEG ground or ground as in "earth" (eg a
> radiator....)?
>

I mean GND1 (from the power supply). I also tried analog-ground, that didn't
affect the scene much.

> > I also get this noise when I use battery-power. Could shielding of
> the
> > Electrode-cables help ?
>
> What kind of noise is it? Could you describe it more in detail?
>

a 50-hz - signal, ranging the full scale of the inputs (0 - 1024). even when
i try to measure stronger signals like EKG, i don't see anything unless i
touch GND1.I searched the archives of schematics once more and found the
'new-design-folder' of Jörg's rs232eeg, which shows a shield driver for the
electrode cables, which isn't implemented in the modular-EEG- designs. Why
isn't that one used - could it guide the signals ?

> Try connecting a large (>100uF), low-impedance capacitor across the
> battery (to lower its impedance) and see if that helps. Does anyone
> else on the list have any suggestions? Joerg, Jim-M?
>

I can try that. makes this also sense when i use a net-adaptor ? in
modularEEG v02-digital schematics I found the hint that the DCDC-converter
shouldn't be usedwhen driving in battery-mode. there is also another 7805S in
the voltage-section.
Is this necessary for battery-mode ? when i just replace the net-adaptor with
a battery, the signal looks pretty much the same to me...

> > I don't fully understand the concept of the right-leg-driver -
> > is it ok when i use the average-signal from the earlobes ?)
>
> Actually, the DRL is an output and replaces the regular passive
> ground lead. It is part of a regulator circuit that tries to suppress
> low-frequency common mode signals such as 50Hz hum about 100 times.
>
> The DRL should work anywhere (as far as I know). Try putting it
> farther away from the head (on the arm for example) and see if there
> is a difference in how well (or poorly) it works.
>
> Hum caused by electrode-impedance mismatch will remain unaffected,
> and may only be reduced by more careful cleaning and abrasion of the
> scalp.

ok, it couldn't do all the work. maybe my DLR doesn't work at all ?for
cleaning and mounting I currently use a granular-like, non-abrasive
electrode-gel, watered bridge-electrodes and a grid-like plastik
electrode-cap. that fits the needs, but I'm not sure concerning the electrode
cables/shielding. I currently make a 1-cablestrng-connection from the cap to
the beld, where the EEGamp is weared.

> However, if the hum does not saturate the amplifiers at any
> time, you can filter the signals digitally with a 50Hz notch filter.
> I can get you some source code for you to try, if you like.
>

does this mean to zero the 50Hz-bin in the fft and retransform to time-scale
? could such signal processing be done 'on-the-run' at the avr ?I'm heaviliy
interested in fft-resources (my fft has neither windowing, nor any signal
analysis like coherence yet ...)

> Another thing you can try, to reduce the noise:
>
> Put the amplifier board in a metal box and connect the box to virtual
> ground with a short wire, near the amplifier inputs. Drill holes just
> large enough for the electrode wires and ribbon cable to pass through.
>

I tried this already, I senses plausible to me. Unfortunately the shielded
box didn't give a better result in my case.

> Please let us know what you do, and what results you get, we'd really
> appreciate it!
>
> Btw, consider signing up to the new list at
> http://lists.sourceforge.net/lists/listinfo/openeeg-list so that we
> can continue the discussion over there... :o)
>

ok!

> Regards,
>
> Andreas
>

thank you for your fast reply and the information.

Best Regards,
Chris.



This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:44 BST