Menu
▾
▴
Re: [Openeeg-list] OpenEEG Questions
|
From: Yonah E. <yo...@gm...> - 2020-09-23 19:43:28
|
Hello all, Thank you all so much for your great advice. I'm starting prototyping with one channel for now, to test my schematics and ICs, as well as my control mechanisms. If I make any headway with my changes, I'll make sure to reach out to you all and give you an update! On Mon, Sep 21, 2020 at 5:44 AM Blago Gmail <don...@gm...> wrote: > Hello Yonah, > > Your idea for developing upgraded hardware for OpenEEG is definitely good! > > I wonder how many users feel "performance constrained" with current > version? > > Here are you points: > >>> -Designing better noise cancelling to reduce commons noise, > >>> as well as noise between electrodes. > You can gain significant improvements over signal to noise ratio if you > "integrate" the signal processing chain down to the ADC - the so called > Analog Front-End (AFE). What I mean under integration is to choose from > available options multi-channel AFE implemented as integrated circuit > (ASIC) - Texas-instruments, Analog-devices, Linear.... Those AFE chips > are coming with TIA (first amplification, in Damian's terms) and PGA > (second amplification) amplifiers + different type (single-ended or > differential) and resolution, ADC inside. All of them requires some kind > of external digital control + reading out converted data samples (they > don't memory buffers or storage place inside) > > >>> -Integrating an FPGA into the design which acts as a > >>> hardware Fast Fourier Transform > I think you are underestimating the FPGAs. Despite the good use for > signal processing, FPGA are good for implementing timing engines (the > control signals needs to drive properly your AFE circuit). > > Imagine you are picking up an FPGA such as ZYNQ, who comes with dual > core ARM processor inside... you are pretty much done with the > configurable hardware you want. The firmware whiten for the ATMega > should be ported to ARM but this will come with reward to have much more > powerful processing unit. The FPGA fabric can be used for signal > processing and controlling the external AFE. As you will end-up with > good amount of data samples, better consider USB3/USB2 as interface with > the PC. > > >>> -Making the circuit modular enough that you can stack > >>> multiple together to have 4, 8, 16, etc. electrodes reading > >>> at the same time. > If you go the FPGA way, you will have ability to have reconfiguration on > PCB and IC levels. > > > >>> -Long term: modifying the design of the circuit to work > >>> without the AT processor, instead having hardware control > >>> from the GPIO of a Raspberry Pi, and all data being fed raw > >>> to either the GPIO or USB interface of the Pi. > Raspberry Pi is good platform but OpenEEG is too hardware specific that > I am afraid that platforms as Raspberry Pi or Begalbone or Arduino are > too generic for this project. > > Kind regards, > > Blagomir Donchev > > > > > > On 20/09/2020 23:46, Damián Filo wrote: > > I am not, really... :D > > Can you please define "leads"? Do you mean channels? > > And what challenges? > > > > OpenEEG is in default built as bipolar EEG. That means for each channel > > you need to connect two electrodes. The output is difference of voltage > > between those two electrodes. But even with two electrodes you won't > > need to make two amplfication circuits. The difference of voltages is > > processed in the instrumentation amplifier (first amplifier) and then > > the one difference signal is amplified further, so it can be reliably > > converted to digital data. So you will need an amplification circuit > > (cascade) for each channel (2 electrodes). > > OpenEEG with v2 protocol and baud rate 115200 can support up to 6 > > channels, although ATmega16 can support 8. If there is need for more > > channels, you would have to rewrite the firmware and change the baud > > rate. But with that you would lose compatibility with software > > supporting OpenEEG (like BrainBay). > > I simply chose not to face those challenges and I stuck to the > > specifications concerning OpenEEG compatibility. I use olimex v2 > > protocol firmware with ATmega16L. > > > > > > > > On Sun, Sep 20, 2020, 23:18 Steve Sterling <acu...@gm... > > <mailto:acu...@gm...>> wrote: > > > > You're really knowledgable about this Damian! How do you work > > around the challenges with multiple leads, will each lead require > > it's own circuit? > > > > Thanks, > > Steve > > > > On Sun, Sep 20, 2020, 1:17 PM Damián Filo <nog...@gm... > > <mailto:nog...@gm...>> wrote: > > > > I am sorry for the spam, but in my previous e-mail I said: "any > > of those changes you proposed are possible" - that is false. One > > change you mentioned was to feed raw data onto GPIO pins of RPi. > > That will not work, because what you need to do with the signal > > is to convert it from analog to digital data through ADC and RPi > > doesn't have an ADCs. You can avoid using ATmega > > microcontrollers, but you still would have to use some external > > chip with ADC functions. And then you would have to program your > > own firmware and protocol, so you can use the data in > > applications running on RPi. > > Though what you can do with RPi is to feed the data directly > > from ATmega through native UART interface on RPi without the > > need for USB-UART chip. But about that, I am not really sure > > about the signalling voltage mismatch 5V/3.3V. So it is still > > preferred to use a USB interface with UART bridge like CP2102. > > > > S pozdravom, > > Damián Filo > > > > > > so 5. 9. 2020 o 21:49 Yonah Elorza <yo...@gm... > > <mailto:yo...@gm...>> napísal(a): > > > > Hello! > > > > I'm a computer engineer interested in making some > > modifications to the OpenEEG circuit, and had a few > > questions, if anyone is available to answer them. I've read > > through all the documentation on the website, and have > > worked with the OpenEEG platform for a few years now, and > > wanted to see if the following was possible: > > > > -Designing better noise cancelling to reduce commons noise, > > as well as noise between electrodes. > > -Integrating an FPGA into the design which acts as a > > hardware Fast Fourier Transform > > -Making the circuit modular enough that you can stack > > multiple together to have 4, 8, 16, etc. electrodes reading > > at the same time. > > -Long term: modifying the design of the circuit to work > > without the AT processor, instead having hardware control > > from the GPIO of a Raspberry Pi, and all data being fed raw > > to either the GPIO or USB interface of the Pi. > > > > If anyone has reasons this wouldn't be possible, or if you > > have suggestions for any of these changes, please let me > > know. I absolutely love this project, and would love to be > > able to contribute my own work to it. > > > > -- > > Best Regards, > > Yonah Elorza > > Columbia School of Engineering | Class of 2018 > > Computer Engineering > > Tel: ( <tel:%28786%29%20205%20-%204304>305)-542-1735 > > _______________________________________________ > > Openeeg-list mailing list > > Ope...@li... > > <mailto:Ope...@li...> > > https://lists.sourceforge.net/lists/listinfo/openeeg-list > > Go to the above address to change your > > subscription options, e.g unsubscribe. > > > > _______________________________________________ > > Openeeg-list mailing list > > Ope...@li... > > <mailto:Ope...@li...> > > https://lists.sourceforge.net/lists/listinfo/openeeg-list > > Go to the above address to change your > > subscription options, e.g unsubscribe. > > > > _______________________________________________ > > Openeeg-list mailing list > > Ope...@li... > > <mailto:Ope...@li...> > > https://lists.sourceforge.net/lists/listinfo/openeeg-list > > Go to the above address to change your > > subscription options, e.g unsubscribe. > > > > > > > > _______________________________________________ > > Openeeg-list mailing list > > Ope...@li... > > https://lists.sourceforge.net/lists/listinfo/openeeg-list > > Go to the above address to change your > > subscription options, e.g unsubscribe. > > > > > _______________________________________________ > Openeeg-list mailing list > Ope...@li... > https://lists.sourceforge.net/lists/listinfo/openeeg-list > Go to the above address to change your > subscription options, e.g unsubscribe. > -- Best Regards, Yonah Elorza Columbia School of Engineering | Class of 2018 Computer Engineering Tel: ( <%28786%29%20205%20-%204304>305)-542-1735 |
