Re: bluetooth eeg

Performance of your design will depend on a lot of things. You will
probably benefit from moving to a home-etched PCB or one ordered through
OSH Park (https://oshpark.com/ - you could spent $20 on a 2 x 2 in
board, get 3 copies professionally done + shipping) due to internal
impedence and noise introduced on breadboards. Not to mention most
breadboards I've used don't make good contact, with high resistances in
odd places.

Selecting low-noise IC's and tighter tolerance passives in the right
spots will be big. Keep your traces short. The kind of gain needed for
an EEG is pretty high, so noise introduced before the amplification
stage will make the biggest impact.

If you go the PCB route, fill all the space without traces with a ground
plane. Or to get super fancy, you could use a 4-layer board and dedicate
an entire layer to ground.

Being an amateur radio operator I've messed with some RF designs, and
they are hellishly picky. I've built some that simply won't operate on
breadboards. My favorite weird thing was some time ago, I tried to
breadboard a Class D audio amp using mostly passives opamps, and FETs,
and when I touched certain parts of the circuit, internal resonance
would work to pick up the low end of the FM radio spectrum. Classical
music with no explicit audio input.

A good platform for getting into simple bluetooth serial can be
Sparkfun's BlueSMiRF units (https://www.sparkfun.com/products/12577).
With this you can build your entire design and have 115200 baud serial
coming from your PIC. Connecting this to a computer through an FTDI
USB-TTL interface (~$20) would work tethered, then you could swap the
FTDI out with the BlueSMiRF and with little modification have bluetooth
serial passing your EEG data back.

The ADS1299 is so expensive because it is being marketed for medical EEG
devices, and is touted as a "do-it-all" chip. It's nice, and I'm sure
24-bit resolution is nice too. However I have always seemed to find
tricks and workarounds to get "almost as good" performance out of
general purpose microcontrollers. For example, ATMEL published an
Application Note on a technique to improve the built-in 10-bit ADC's in
an AVR chip to higher bit resolutions, trading off time resolution.
Consistent noise is utilized to increase signal resolution - by taking
more frequent samples of the signal ("oversampling") and "decimating"
the result to produce a higher-resolution signal at the cost of 4x more
CPU cycles per bit of resolution and you still get some of the
non-linearity.

Check it out, though. It means you can throw a 10-bit micro into a
project and get 11 or 12-bit ADC if you have the cycles to perform the
samples and calculations. http://www.atmel.com/images/doc8003.pdf

On Thu, Nov 6, 2014, at 07:22 AM, wikitronic wrote:
> hi,
>
> The ads1299 looks like a great chip but the price is kind of steep. I
> could order a sample but, you know this is kind of a hobby project.
> Even for a professional price is an issue.
>
> Here is a great example of a hobby eeg:
>
> http://www.instructables.com/id/DIY-EEG-and-ECG-Circuit/
>
> Could anyone comment about what the performance of my design might be
> IF I get it to work?
>
> Thanks
> ------------------------------------------------------------------------------
> _________________________________________________
> Openeeg-list mailing list [email protected]
> https://lists.sourceforge.net/lists/listinfo/openeeg-list Go to the
> above address to change your subscription options, e.g unsubscribe.

------------------------------------------------------------------------------