cannot get 500 or 1000 hz sample rate from Wifi Shield (or SD card)

profeellprofeell
edited June 2018 in Wifi Shield
[original thread title: "Do you release a standard configuration guide?"]

I work hard many months for getting the 1000Hz sample rate EEG-data, but I failed. I found the reason is that you didn't release a standard configuration for  user, for example:
1. system: windows*.*
2.update the cyton firmware, you need arduino version *.* and chipKIT version *.* cyton firmware version *.*
3. WiFishield firmware version *.*
4.Openbcihub v*.*.*
5..Openbcigui v*.*.*

if you use last configuration, it will be good, I spend many many time for getting the higher sample rate EEG-data and I am not succeed right now.

I hope you can give me a solution, I tried many time, it is too hard, too tired, can you help me?

«13

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Hi Profeell,

    Mentioning AJ Keller @pushtheworld. @openbci.

    On previous threads I suggested that you could first try recording at high sample rates to your SDcard. Was that successful?

    You have opened several new threads on your attempts to get higher sample rates. Unfortunately every time you open a new thread, (1) the staff members who have been contributing to that thread are not included in notifications.  (2) the context of what you have done and the issues are lost.

    I'm not sure what to suggest at this point. Does your Wifi Shield even work at the default sample rate? My sense is the answer is no. That is what you need to achieve first.

    Regards,

    William

  • profeellprofeell
    edited June 2018

    Thank you so much, you have been gave me many help, but really I can't get any
    succeed. I want to tell you some my work in recently several months.

    My goal is that I want to get higher sample rate EEG-date, I feel the 250 Hz
    sample rate is not enough for me, so I want get 1000Hz or 2000 Hz sample rate
    EEG-data. I have two way to get it by your guide, one is record the data to SD
    card and other way is record the data by WiFishield with cyton board. First
    I used the SD card to record the data, I used the Arduino(V1.8.5) to send the SDK code, the
    work state is cyton board with the openbci USB Dongle connected by Bluetooth. The
    step as follow:

    1.      1.  Send ‘ ~~  ‘to get sample rate currently, Arduino Serial
    Monitor returns ‘Sample rate is 250Hz$$$’;

    2.      2.  Send ‘~4’, to get sample rate 1000
    Hz , Arduino Serial Monitor
    returns ‘Sample
    rate is 1000Hz$$$’;

    3.      3.  Send ‘s’, to get SD card data logging for specified time
    s = 15 minutes;

    4.      4.  Set the serial baud rate 115200 (the Serial Monitor bottom option);

    5.       5. Send ‘b’, start to record the data. I found the Serial Monitor received many streaming
    data and show in windows, sometime it is so fast, sometime it turn into slowly.

    6.       6. I tried to send ‘s’ code after several minutes and I want to make it stop, but the
    Arduino Serial Monitor can’t stop to receive the data, I wait for long time,
    the streaming date stop;

    7.      7.  When I take off the SD card, I used MATLAB to read the EEG data, I found the data is no right, if it
    is 1000 Hz sample rate, I will get about 900,000 data point, but I only get 90,000
    more  data point, I don’t why?

    I tried many times to get the 1000 Hz sample rate data, but I am not succeed, can you
    help to check it out, which step is wrong, how can I get SD card 1000 Hz data?

     So get SD card data is no succeed, I decide to use WiFishield with cyton to get
    higher sample date. My cyton board and WiFishield both are new. I get them in April.
    So I directly make the WiFishield seat the cyton board and open Openbcihub and
    OpenbcGui , but the Opengui show: no cyton board attached the WiFishield. According
    to the guide, maybe I needn’t update the firmware and I did, but when I make
    the WiFishield seat the cyton board, I still can’t get succeed. My step as follow,
    please you check it out for me, which step is wrong.

    1.      1.  I make the WiFishield seat the cyton board;

    2.       2. I found the openbci 5717 WiFi name and connected it; my computer connected the
    Openbci 5717;

    3.      3.  I check WiFi and input 192.168.4.1 in the Internet browser, the page show the WiFi
    Manger menu, it is working;

    4.      4.  I open the Openbcihub the version is V1.4.2, the latest release;

    5.       5. I open the OpenbciGui and click the ‘live(from cyton)’ and click ‘wifi(wifishield)’,
    it shows ‘no cyton board attached the WiFishield, but I can get the WiFishied
    IP 192.168.4.1;

    6.      6.  I tried many times, all is fail. I read Openbci guide, maybe I need update the
    firmware, so I update my WiFishiel firmware from V2.0.4 to V2.0.5, it is
    succeed. Then I update my cyton board firmware, I used Arduino V1.8.5 and
    chipKIT V2.0.5 and get the latest OpenBCI_32bit_Library, OpenBCI_32bit_SD, WiFi
    Master Library release, my computer system is Windows 8.1, the update is
    succeed.

    7.      7.  I make the WiFishield seat the cyton board again, then connected the Openbci 5717
    WiFi, open the Openbcihub, open the OpenbciGui, click the ‘live(from cyton)’
    and click ‘wifi(wifishield)’, it still shows ‘no cyton board attached the
    WiFishield, but I get the WiFishied IP 192.168.4.1;

    8.       8. I check some detail, the WiFishield ext pwr is on, I have taken off the USB
    Dongle, all is correct.

    9.       9. I tried many times, I can’t get succeed.

    10.   10. So I haven’t anyway, I
    tried to change the chipKIT version to renew update the cyton firmware, when I
    used the chipKIT version V1.3.0 or V1.3.1 to update the cyton board firmware, I
    can get the information, OpenbciGui shows: WiFishield connected the 8 channel
    cyton board, but I want to start system, the OpenbciGui is no work, no start.
    The hardware problem seem is solved, the software is no work. I still tried
    many many times, I can’t get it .

    So I have had not any way
    to get it, I post a new discussion, can you give me a standard configuration,
    so I can work to use the Openbci, for example:

    1.       system:
    windows*.*

    2.       update
    the cyton firmware, I need arduino version *.* and chipKIT version *.* cyton
    firmware version *.*

    3.       WiFishield
    firmware version *.*

    4.       Openbcihub
    v*.*.*

    5.       Openbcigui
    v*.*.*

    DIY
    to try those works, it really waste too more time, can I save some time and use in
    data analyze, can I get your better suggestion?  I must say: thank you wjcroft, you have been gave
    me many help, but now I hope you can take some time again and help me to solve
    my problem, really I like the BCI tech, I like Openbci, I hope I can get start
    my researching work. Thank you!


  • wjcroftwjcroft Mount Shasta, CA
    Regarding your SD card recording for 15 minutes, you need to send CAPITAL S, not lowercase s. Is that what you sent?

    Also, some Cyton command strings, that require multiple bytes, must all be sent within a short period of time, within 1 second.

    If you are not reading correctly from the SD card, it may be an issue with the formatting, or brand of the card. There are some specific suggestions on which brands / models work best.

  • profeellprofeell
    Oh, I got it. I tried different version OpenbciHub and OpenbciGui and I used other one computer run in windows 7, I got it final by OpenbciGui V3.2.0, my OpenbciGui can get 1000 Hz sample rate EEG-data. But I still want to get 2000 Hz or 4000 Hz sample rate EEG-data, how to use the WiFishield with cyton to get higher sample rate EEG-data? The OpenbciGui can't give over 1000Hz sample rate, so what software will be used? Arduino? But when I connected  the WIFIshield with my computer directly, the arduino serial monitor can't open because no serial connected, how do I send cyton SDK code, I want to set the new sample rate, how to get it ? I am a new one, I need you give me a solution. Thank you so much.



  • wjcroftwjcroft Mount Shasta, CA
    You can get 2000 or 4000 with the SD card, did you try that?


  • profeellprofeell
    edited June 2018

    Thank you , I have still several questions:

    1. The Openbci SDK give the information: IMPORTANT! The Cyton with USB Dongle cannot and will not stream data over
    250SPS. Plug in the     WiFi Shield to
    get speeds over 250SPS streaming. You may still write to an SD card though, the
    firmware will not send EEG data over the Bluetooth radios.  I want to know
    how to supply power for cyton board when I plug in the WiFi Shield, Power plug
    in cyton or WiFi Shield  (and put EXT PWR switch on Shield to ON)? Or no
    supply power for WiFi Shield, only plug in cyton board?

     

    2.When I plug in the WiFi Shield, do I need still connect USB
    Dongle for sending cyton SDK code?

     

    3.can you offer more detail introduction about getting higher sample rate
    EEG-data in tutorials?

     

    Thank you!!


     

  • wjcroftwjcroft Mount Shasta, CA
    edited June 2018
    1. The Wifi Shield with Cyton powering instructions are stated clearly here,


    Plug battery into Shield. Cyton will get power from Shield.

    2.


    Dongle is not used when connecting GUI via Wifi. SDK codes for Cyton are primarily used for dongle serial port connection.

    3. Were you able to get the SD card to save your higher sample rate test? You don't need to use the Shield for this.
  • wjcroftwjcroft Mount Shasta, CA
    edited June 2018
    To repeat, I suggest you try the higher sample rate recording with the Cyton and Dongle alone, forget the Shield for the moment.

    2. To clarify the response to question #2. When using the Shield, SDK commands can still be sent, but they must go over TCP / HTTP protocol, using the POST command "/command", with the argument being the Cyton SDK string. Documented here,


    This requires a lot of programming knowledge and is not recommended (in your case). Use the GUI instead to send your commands or settings adjustments.

  • profeellprofeell
    Thank you so much, let me try.
  • profeellprofeell
    edited June 2018

    Thanks for your recommended,  I am a new one, but I have a little bit programming knowledge. So I used the TCP/HTTP protocol and got the 4000Hz EEG data. But I only got 2000Hz and 4000Hz data, when I want to get 8000Hz or 16000Hz data, I found the data sitll is 4000Hz data, I don't know why? I want to know whether I need change some Parameters? For example the latency or others? I still need your some suggest. Thank you. 


  • wjcroftwjcroft Mount Shasta, CA
    You likely won't get the top possible speeds through the Wifi Shield. Are you using Wifi Direct? There are limitations on how fast the TCP stream can go. Mentioning AJ @pushtheworld.

    Did you get the SD card recording to work? That should work at the higher speeds, providing you have a decent brand sdcard.

  • profeellprofeell
    edited June 2018

    I used the WiFishield and SD card and get 4000Hz data, but I still can't get 8000HZ and 16000Hz data.

    Next is my steps:

    1. I make the WiFishield seat the cyton
    board and insert SD card ( my SD card is kington);

    2. Insert the USB Dongle.

    3.Open the WiFishield powerful.

    4.I use arduino to set SD card data logging
    for specified time    , I send 'L' and get return, the block count is  about 1.6G

    5. I use arduino to set the sample rate is
    8000Hz, send '~1', and get return.

    6. I open the Google chrome and input
    '192.168.4.1/stream/start and get return 'ok'.

    7. After 1 minutes, I input
    '192.168.4.1/stream/stop and get return 'ok'.

    8. I turn off the WiFishield power and take
    off SD card.

    9. I used matlab import the Openbci
    EEG-data.     

    I hope you can help me to check it, do I
    need add other step and get higher sample rate data?

    Thank you!


  • wjcroftwjcroft Mount Shasta, CA
    I'm not clear WHY you are using the Wifi Shield to initiate your SD card recording session. You can do that just with the serial connection / dongle. The serial connection will not see the faster sample rate data, but that is ok -- it is getting recorded to the card.

    I don't know if the lab ever tested 8000 or 16000 hz sd card recording. It may or may not work.

    I also don't know if your Wifi Shield usage might be a factor in limiting the sd card recording speed. Which is why I am suggesting eliminating that extra potential for interference. And instead just use the serial / dongle to do your recording.

    Mentioning AJ @pushtheworld.
  • BillhBillh
    profeel,

    What is the application for which you need the >4000 Hz sampling rate? I ask because the fastest scalp EEG recording rates actually needed for a scalp EEG interface would be in the 800 to 1000 range, for gamma, and in actual practice 200 Hz is actually enough. Is this for single cell recordings?


  • wjcroftwjcroft Mount Shasta, CA
    Bill, hi.

    Profeel's project is quite unusual. I've never heard of any EEG applications like this. I've expressed my reservations to Profeel on these other threads below, but he is marching ahead. 

    From what he has communicated, he is wanting to capture EEG at a high sample rate. 1K or 2K should be completely adequate, as you say. He says he is already getting 4K, so that is overkill. Then after he has captured the EEG, he wants to play it back in realtime via a modified sound card output, that is capable of very low frequencies in the EEG range. This can be done by removing the output coupling capacitors that impose a high pass on the audio output. He then is going to apply that output waveform to some coils, possibly amplified, and the coils are in contact with his head or the subjects head. For "transfer" of information somehow?? EEG does not work like this, as anyone with experience in the field knows.

    Sounds very far out there. I think this is considerably "over the top" in my opinion, but I'm just trying to be helpful here on the forum.



    Regards,

    William

  • wjcroftwjcroft Mount Shasta, CA
    I've conveyed over and over to him, that scalp EMG contaminates EEG at 40 hz and above, so these high sample rates are pointless. See the second link above for references I gave him.

    I guess I'm just trying to be helpful. Profeel did finally get his SD card recording to work, as well as his Wifi Shield. So that's progress.
  • profeellprofeell

    Well, I tell you about my researching. I am
    researching one sixth feeling phenomenon. I often feel I can send the sixth
    feeling information to other people. The sixth feeling information include taste,
    smelling or other feeling. Please don't ask me how to know I have the sixth
    feeling and how I get know I can send the sixth feeling to other people. I only
    want to let you know I can send the six feeling to other people and I am
    researching it. Maybe you don't believe the phenomenon, I believe it. I can
    send the sixthing feeling to arcoss the woodwall ,glass windows,air and into
    other people' brain. Maybe you think that it is no true. No people have the phenomenon.
    Nothing, I believe, I research. Now we have a presupposition, the sixth feeling
    phenomenon is true, so I have a problem, when the phenomenon happen, what
    information from my brain out and into other people, I think that it is cortex
    wave, it is different with general brain wave, though I still don't know its
    frequency, and I want to get it. I need highest sample rate data and research
    it. I haven't more money, so I only use openbci to get more information. If I
    get the sixth feeling information, I want to reverse to input brain and test
    the information correct or not.  

    I don't know java, I know more hardware
    knowledge, if I know java, I can use WiFishield to control the cyton to get the
    higher sample rate information  So I use
    two way to get may data, cyton have cyton SDK, I can use cyton SDK to set the
    SD card block count and sample rate, then I use WiFIshield to get stream data(
    I can't find the WiFishield SDK to set sample rate code). So maybe you can't know
    my means, it is no profession, but I can get the 4000Hz sample rate data. I
    want to get more precision data because the brain information is very intricate,
    so I want to get 8000Hz or 16000Hz data. Please forgive me, I ask you many
    times, my researching is very special. Maybe I am only one to have sixthing feeling
    phenomenon and want to be BCI tech.

    I am new one, I am learning openbci, so I
    have many question. 

  • wjcroftwjcroft Mount Shasta, CA
    Profeel, hi again.

    No I did not intend to disparage your line of research. I am familiar with some very competent researchers in the field of parapsychology and consciousness studies. If you are not familiar with their work, check out Dean Radin and Arno Delorme. Arno is the inventor of the EEGLAB suite used by many research labs.





    ----

    EEG data of high frequencies does not really pass through the skull. It can only be recorded with surgical procedures that implant electrodes directly into cortex. So the 1000 is plenty fast enough for you, there is no more 'resolution' available at the skull surface at higher sample rates.

    Another aspect you may want to consider is that much research is ongoing in the parapsychology field using very sophisticated equipment. See some of Arno's papers. A lot of this uses high channel count equipment that can do source localization, and in these situations higher sample rates ~ 1000 hz or 500 hz are usually plenty. In the high frequency gamma paper I linked to earlier (your 100hz thread), I mentioned that some of these studies can get higher frequency gamma (~ 100 hz), by using very sophisticated signal processing that eliminates the scalp EMG noise. Unless you have high channel counts (more than 16) and complex DSP, you are going to get swamped with noise much above 40 hz.

    Lastly, many parapsychology / EEG researchers are speculating, that 'consciousness' / information is a primary substrate of the universe. More basic than even physical matter or energy / fields. So while telepathy and information transfer is indeed real, be very cautious in expecting that it must be conveyed by physical transmission methods. You might find this paper helpful, written by MD, Dr. Larry Dossey.


    Regards,

    William

  • profeellprofeell
    Thank you. And right, let me try 4000Hz sample rate. I need take some time to look these articles that you give me. No way I meet the sixth feeling phenomenon, I want to know why it is, it is science, I must research going. Thank you so much. 

  • BillhBillh
    edited July 2018
    Profeel,

    Is the played-back recording going to be audible? Human hearing is going to be about 20 Hz to 20,000 Hz, so you can just barely get the higher end of beta and gamma on EEG played as the bass tones with even a 400 Hz rate. And, as wjcroft has said, you are just going to be recording muscle and other artifact with anything you record that plays back as a frequency over 100 Hz, not scalp EEG.

    You might be able to play back the EEG as a frequency shifted recording, say at 100X speed, which puts the mid EEG frequencies up to mid human hearing frequencies. Would this still allow your intended application? 

    I wonder about the mechanism of transmission of "taste and smell" information transfer you have seen, since most distinctive smell and taste information CAN be transferred via the air, as an odor. I am sure my dog can just about taste what he smells me eating sometimes. I can see why you would want to time-shift the playback, for that reason.


  • wjcroftwjcroft Mount Shasta, CA
    Bill, hi.

    re: playback bandwidth, 20 hz limitation.

    Actually not, checkout the links in the other thread (below). This 20 hz highpass filter is caused by capacitors in the sound card output section. It is straightforward to remove those and get the "full" scalp EEG band from say .5 hz to 100 hz. There are some how-to links in that thread.


    William

  • profeellprofeell
    Thank you so much, My idea is that I want to use sound card to output analog signal, then change electric signal. The good sound card has 24 bit, it can more match the 32 bit EEG data,
  • wjcroftwjcroft Mount Shasta, CA
    Profeell, just to clarify your statement "The good sound card has 24 bit, it can more match the 32 bit EEG data,"

    The OpenBCI samples are 24 bits.

    The sound card will only reproduce low frequencies from say .1 to 100 hz, after you modify the output capacitors. Without that mod, it will only output say 20 hz to 100 hz EEG. The fidelity and accuracy of reproduction is more related to the sound card capacitors, than the number of bits resolution.

    The low order 5 or so bits of the Cyton EEG are primarily noise, since the low order bit on the Cyton represents about .02 microvolts.
  • profeellprofeell
    Thank you so much, let me try it, My English is no good, I am very happy that you know my means. Thank you for you take watching football world cup time for me and give me many suggest. 
  • BillhBillh
    William,

    I meant the limitations of what the ear can hear in terms of low bass notes.


  • wjcroftwjcroft Mount Shasta, CA
    edited July 2018
    Bill, I'm not sure how he is applying / transmitting the amplified EEG signal. I assumed it was with some type of electromagnetic transducer, since he mentioned it goes through walls or windows. There are neurofeedback practitioners using low field transducers on the head, in the range of microteslas.


    That seems considerably overpriced, considering that they are just small coils, not very different from these,


    The latter coils have been used by experimenters using PEMF, pulsed EMF fields.

  • BillhBillh
    Well, I wonder if the coils have any chance of working as expected anyway. Generating EMF is one thing, receiving it is another. You can make snow on the screen of old TV monitors with an un-shielded electric motor, but if you play the video on screen back next to the motor, of course, nothing much happens to the motor. Higher powered cortical stimulation is an entirely different thing of course.


  • wjcroftwjcroft Mount Shasta, CA
    Low level PEMF, as generated by those telephone coils, definitely has an effect. It might depend on the perceptions of the receiver. I've done qigong for decades, and microtesla / milligauss PEMF affects chi flow.

    This website extensively documents the effects of low level EMF fields. With many research papers in one section.



    Here's a site on do it yourself PEMF using audio recordings, an mp3 player, and small coils.


    Regards,

    William



  • BillhBillh
    Thanks WIlliam! Note that the document at 
    almost entirely shows a lack of effects of controlled studies and only very nonspecific, chronic effects from the uncontrolled studies.

    Here's the issue: If, for example, we send someone a coded signal.of a book or a photo using a microwave antenna, and they report  "that made my skin feel hot" does this count as a transmission of the book or photo received? I do not think so. So showing that people get nonspecific chronic changes from being near EMF does not show that we can be communicate information beyond just the presence of signal itself (and controlled studies don't even show that as possible with low intensity).
    .


  • wjcroftwjcroft Mount Shasta, CA
    Well, in the first place I'm not suggesting that 'information' can be transferred via low level EMF. But the effects on biology of low level EMF, even as low as microwatts per square cm, are well documented in various in vivo / in vitro studies.


    Remember this whole idea of sending EEG by some physical field transmission, is Profeell's hypothesis.

    On the other hand, I have followed parapsychology studies for many years by Dean Radin and other respected researchers. And they indeed have shown inter-subject effects. The evidence for what is called "presentiment" is undeniable.


    Taken as a whole, the well conducted parapsychology research seems to show that consciousness has non-local, even retro-causal effects. This has profound implications for how we view the world and our place in it.

    Best regards,

    William

Sign In or Register to comment.

© Vanilla Keystone Theme 2026