From: Joel Eriksson (jen_at_ettnet.se)
Date: 2002-01-18 00:09:24
On Thu, Jan 17, 2002 at 03:00:12PM -0500, Jim Meissner wrote:
> Dear Joel:
>
> > It would be very nice to build the EEG as small and cheap as
> > possible......
>
> It is my personal opinion that this project has not progressed
> because of the above requirement. Joerg has spent a considerable
> amount of time attempting to make everything as cheap as possible.
> He has come up with some extremely clever designs.
>
> My view is just the opposite. Get a design that works FIRST.
> Build it, and then look at how to make it cheaper.
I agree, a working design is of highest priority of course.
> > and make a separate multiplexer that can be used to
> > stack multiple EEG-modules to allow for as many channels as
> > one wishes.
>
> The design that Andreas is working on right now has a built in
> multiplexer for 8 channels. All you need to do is hook up more boards.
Ok, nice. :-) Since I'm a software-guy I have not realized that.
> > other interfaces (USB, Firewire, IrDA, whatever), which might
> > be needed when increasing the number of
> > channels and discovering that RS232 is not fast enough
>
> The device that I built ran at 9600 Baud and transmitted 5
> bytes at a 128 Hz rate. That worked as well or better than
> most machines out there. So what is this USB, Firewire, IrDA for?
> I thought you wanted a CHEAP brain monitor.
I would prefer a SCALABLE brain monitor. The basic modules
should preferably be inexpensive of course.. USB, Firewire,
etc are just examples of interfaces that may be needed /
wanted for some applications. For EEG "brain mapping" you
need quite a few channels (like 32 or even up to 128).
> So as not to keep this argument going for another few months, the
> present goal is to transmit RS232 data at 112.2 K Baud over
> fiber optic isolated transmitter receiver pairs. We will be able
> to satisfy the safety isolation ( 5,000 volts ) requirement with
> a 2 inch length of fiber. The transmitter and receiver will be
> on separate boards. Andreas has picked out the chips that should
> just "bolt" together, and we all agree that this is not the lowest
> cost solution, but should work without fiddling.
>
> Joerg is in search of a "safe" 5,000 volt isolation DC/DC converter
> that will eliminate the need for batteries.
A working EEG is the current primary goal of course..
> > simple configurationfile could define what kind of modules
> > are on each channel and applications just register themselves
> > for the kind of data they can handle and doesn't have to
> > know about anything else.
>
> You want "cheap as possible" and Plug and Play windows environment?
Actually I'm talking about software-design, from the programmers
perspective. The application developers should not have to worry
about hardware-specific issues since that has already been taken
care of in the base libraries. Modular software design as well as
modular hardware design.. As for the "Plug and Play windows
environment" I don't even use Windows and avoid GUI's whenever
I have a choice. :-) Unix and textbased programs is my thing.. :-)
> My view is: lets get something working FIRST before we get so fancy.
> At this point Andreas has provided 8 channels that can be addressed
> by the "downloaded" firmware. Later we could have a 2 way comm.
> link to the micro.
Again, I agree.
> > Of course, all modules should be able to connect directly to
> > a PC too if multiple channels are not needed.
>
> No, even for one channel you still have to go through
> the "save floating isolated" A/D converter!
I thought the A/D would be on the basic EEG-module, I think I understand
the design better after Andreas last post, where he writes:
The idea is to use these as plugin-boards, similar to the PCI-cards
in a PC. The ADC module would be the motherboard. Still working on that.
This seems rather like what I had in mind with the exception that the
multiplexing device is a requirement, even though I probably did not
express myself very clear. :-)
Let's see if I have understood this correctly now; The plugin-boards
sends analog signals to the ADC module on the motherboard and the
digitized samples are multiplexed and sent through RS232 to the PC.
Then we have already achieved the kind of modularity I wish for, only
the motherboard will have to be modified for using a different interface.
Let me know if I have misunderstood something.
> Juergen P. (Jim) Meissner
-- Joel Eriksson
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:36 BST