EEG by USB6009

[kannajin]

harlow, to everyone!!

What is ur opinion is use DAQ USB6009 to read signal we get from the brain?

the resolution for USB 6009 is 14-bit. are this enough??

sample rate is 48 kS/s

 the min voltage is  (-1/+1) but our brainwave only around 10UuV, this can be slove if we gain it till 1V 

is around 1000 times.

will this spoil the wave n  more noise?

My design for amplifer is  Instrument Amplfier --> low pass -=-> high pass --->  amplifer (100times) --> notch flier--> non-inverting amplifer--->otocouple.

any idea for it?

[Coq_Rouge]

Hi Here is some eeg stuff http://openeeg.sourceforge.net/doc/modeeg/modeeg.html. You might find it useful, or not ;). I also think the basic idea pass the sanity check. A problem might be the last stage. The optocoupler. Almost all of the optocouplers on the marked are digital. But Agilent’s HCNR200/1 and HCPL-4562 constitute basic optical coupling building blocks. I also think Fairchild Semiconductor has something that can be used. Anyway Google the phrase "analog optocoupler" I will also recommend using battery as powers supply e.g. two 9 volts cells (giving +/- 9 volt). The reason for this is that batteries are a very low noise power source. I think noise is your worst enemy in this project. So you have to design the analog front-end with this in mind all the time. Battery is also good for the safety aspect. The problem is that battery gets worn out. But you may use an external power supply while experimenting and batteries then you are measuring on humans. good luck

[Steve.J]

The resolution will be enough - the dynamic range of EEG signals is not that great (unless you need to capture EEG during a seizure).  12-bits is plenty.  One possible issue is that the 6009 product is a multiplexed A/D so each channel is scanned/sampled in sequence, which could lead to loss of accurate phase information.  This may or may not be a concern for you depending on how the data will be collected and used.

There are still lots of analog optical isolation modules on the market - search for "isolation amplifiers".  Analog Devices has a good selection.  Battery power for the front-end makes a lot of sense, but you still need to be worried about DC fault currents even from the battery.  AC coupling the instrumentation amplifier solves this issue, as well as blocking the DC offsets from typical surface electrodes.

Steve

[kannajin]

hi, steve

what do u think if use RS232 pc port communicate with labview? 

so the RS 232 will sent data from the Microcontroller to Labview to show the wave??

do u have any example on it?

[Steve.J]

You should be able to do this fairly easily...you can use relatively high baud rates when using direct connections between microcontrollers and PCs.  You will need to determine how to bundle the data that you are sending from the microcontroller...RS-232 is usually used to send 7 or 8 bit data (usually in the form of ASCII characters) so you will have to devise a scheme to send two bytes (or characters) per data point for each channel, and the LV application will need to be able to take the two bytes and assemble them into a 16-bit "word", then apply a scaling factor to convert this 16-bit value to a floating point number that represents the actual amplitude of the EEG signal (in microvolts).  Usually some form of a header is sent (special character or string) that allows the LV application to identify the start of a sequence of channels/data and often the calibration or scaling factors are also sent as part of the header.  This allows the LV program to periodically reconnect and syncronize to the data stream if something goes wrong.  Look through the serial examples that ship with LV to get some ideas.