[EEG] Open EEG board / next steps

Jonathan rrmutt at gmail.com
Tue Mar 31 10:36:23 PDT 2009


Hello campers,


Just thought I'd summarize a little of what we found looking at
Tracy's OpenEEG board. I also had a chance to look at the schematic a
bit.  So I put a scope on the analog output of one channel of the
amplifier board. I saw mostly a 60-hz square wave, in other words
clipped line noise (hum). There was one or more loose connections in
the active electrodes, so we unplugged them. I tried to get a heart
pulse (EKG) signal by touching the differential inputs with different
hands, but did not see much beyond noise. None of this is surprising;
this is pretty much what I would expect from a high-gain amplifier
flopping around without a shielded enclosure or dedicated electrodes.

So the first thing I would do is put it inside a metal box. We have a
design decision to make on the electrode connectors. Tracy has some
nice mini-Cannon plugs (like small microphone plugs) that have three
conductors. Those would require mating panel-mount jacks (Tracy, do
you have those?) and three-conductor wires. Advantage: you can run two
signal lines with one cable, with the disadvantage that you need a Y
junction to split them off near the ends. An alternative is to use BNC
jacks which are cheaper and readily available, but would need two for
each amplifier channel, and two corresponding wires. And we should
give some thought to finding light and flexible electrode cables.

(Here's something I don't completely understand: the amplifier
channels are differential, meaning they have two inputs and amplify
the difference between them. Though this is very good for rejecting
common-mode noise like hum, I don't know how you set that up in terms
of single-ended scalp electrodes. Is there a standard practice for
that?)

Looking at the board design, it looks pretty good. They could have
done better analog filtering (I would have used a Chebyshev with a
steeper cutoff) but looks otherwise reasonable. The interface board
supports three amplifier boards (we currently have one) for six
channels, that is 12 differential-pair electrodes. I need to look at
the serial output format; I'm pretty sure that should be
straightforward. My first impulse is to write a little python script
to sanity-check the serial output and capture the raw data but if
other folks want to play with the brainlab software we can do that
too.

Safetywise, they did all the right things: the serial port is
optically isolated and the input power goes through an isolating dc-dc
converter. I think this is as safe as you can make it and wireless is
overkill (though we could put the serial data through a Zigbee
wireless module for portability should we need it). I think we should
use a good switching power supply for the OpenEEG board: this will
save the hassle of  low battery and current draw issues.  But if folks
are paran^H^H^H safety-conscious we could use a battery pack.

So we need a box and a design decision regarding electrode connectors.
I could likely source panel-mount mini-Cannons, but those may be a
little expensive (I'm guessing $5+ per differential channel). And we
should probably design the enclosure to house two sets of OpenEEG
boards, that is: two interface boards and six amplifier boards, for 12
differential channels for 24 electrodes. It would be really nice to
stack boards using mating connectors to reduce the flat cable
sphagetti and noise pickup; I can source M/F mating .1 inch connectors
from digi-key if we build more amplifier boards.  So we need a metal
box from somewhere that we can drill and mill for attaching boards
inside and connectors through. We could be as fancy/expensive as
making a custom front plate with CAD and having it fabricated, to
cheap and cheerful as repurposing an old hard drive enclosure.

Just to be clear that there is going to be a fair amount of  work
involved; just building the enclosure is going to be a little fiddly,
and it's going to be a while before we see results. But I'm willing to
put a little time into chipping away at it.

I can't make Thursday for a meeting, but I'd be willing to look at the
serial output of Tracy's board sometime. I'm not particularly keen on
building the enclosure myself: does someone else want to take the
lead? I'm happy to give advice and find parts as necessary.  We also
might want to get some more OpenEEG amplifier boards for testing the
serial data throughput.

Cheers all,

-Jon



On Fri, Mar 27, 2009 at 5:09 PM, Tracy Jacobs <kinetical at comcast.net> wrote:
> Hi,
>
> Thanks to everyone who came to the meeting last night, I'm really excited
> about the group.  After everyone went home I was talking to Michael Kan, who
> had some interesting ideas about engineering the EEG.  I encouraged him to
> join us next time.  He suggested we make the connection between the EEG and
> the computer wireless, using a bluetooth device, to reduce noise and also to
> ensure safety by having the circuits completely separate from one another.
>  I don't have any bluetooth device, but if any of you do, bring it along
> next time.  I will bring a box  for the circuit boards.
>
> Tracy
> _______________________________________________
> EEG mailing list
> EEG at lists.noisebridge.net
> https://www.noisebridge.net/mailman/listinfo/eeg
>



More information about the EEG mailing list