c) Now we are going to power down the channels we aren’t using. Do this by clicking the channel number buttons outside of the left side of the EEG DATA montage.
Are you using OpenBCI_GUI or some other program? For example with the GUI CSV file, it's easy to discard columns that are unused. Another example with neuromore or BrainBay, you just do not wire up the unused channels in the circuit diagram.
I need to increase the sample rate to 500 but with this sample rate I can't send the 31 bytes per sample in less than 2ms to recieve all the data because of the baud rate so I want to send only 4 channels to make this possible. Could this be done?
One other way, if you keep the RFduino link. Is to continue with the 250hz radio packet rate, but stuff 2 4 channel samples into each radio packet. That may be easier than the wired USB link. Another option you have is to reduce the sample size from 24 bits to say 16. All these firmware packet changes will require you to make the matching adaptation in the packet receiving parser.
If you check out the new firmware done by github.com/pushtheworldllc/, you will actually notice that you can increase the BAUD rate now. Could help with your sample rate problem.
New firmware baud setting does not get around the PIC -> RFduino bottleneck...
---------- Forwarded message ---------- From: Andrew Keller Date: Thu, Sep 1, 2016 at 12:22 PM Subject: Re: Firmware 2.0.0 Solution To: William Croft Cc: Joel Murphy , Conor Russomanno
Nope, the bottle neck you mentioned is a limiting factor in sample rate.
To my knowledge and testing baud rates above 115200 do not work and interfere with the transmission of data. However, you can still pull and send data faster from the serial port if you raise your baud rates.
3.0.0 will enable faster sample rates through the use of 1) External serialport on the board 2) Networking device (e.g. ESP8266) with SPI
AJ
From: William Croft Date: Thursday, September 1, 2016 at 3:14 PM To: Andrew Keller Cc: Joel Murphy , Conor Russomanno Subject: Re: Firmware 2.0.0 Solution
On your upgrade page you mention the new faster baud rates. Does that also enable higher sample rates or more data to be sent over the air between RFduinos? I recall there was a bottleneck on the serial port between the PIC and the Device RFduino; is that still the case?
Colin, yeah, AJ's new 'firmware' baud command is apparently just changing the baud rate of the FTDI chip (and RFduino) inside the dongle. So is not related to the baud rate between the mainboard PIC and RFduino (fixed at 115200), nor the speed of radio packets between the two RFduinos.
No, the external clock is for the PIC (chipKIT) processor. Cannot change that. To adjust the sample rate you're going to need to mod the firmware. I mentioned above 3 different ways that you can explore. (USB wired link, or send 2 * 4 channel 500hz samples per 250 hz packet, or saving to SDcard.)
What is your experimental setup that requires the 500 hz? Yet one other firmware trick you could try is to do what is called 'oversampling' at 500 hz say, then time series average those down to a lower rate.
Thanks wjcroft. It works with the USB wired link but now I have another question. If I use the USB wired connection, the peripheral bus clock to send the packages is 8MHz? If it is, I've found this guide: http://ww1.microchip.com/downloads/en/DeviceDoc/61107F.pdf. In the table of the page 15 for 8MHz of peripheral bus clock there isn't 115200 or 230400 of real baud rate so I don't know if I'm not going to have any problem with the data transmission.
AJ Keller's v2 firmware is now shipping on new OpenBCI orders. So it may be possible to skip some of the steps in Winslow's tutorial, as AJ added some new commands. For example to send out the serial data on Serial1 vs. Serial0. You'll have to check the v2 documentation.
But in any case you can choose which firmware you want to run. It may be simplest to just go with Winslow's, since that is known to work.
Comments
http://openbci.com/forum/index.php?p=/discussion/727/sd-only-version-for-increased-sample-rate-and-interrupts
Winslow's wired usb link incorporates the sample rate change ability.
From: Andrew Keller
Date: Thu, Sep 1, 2016 at 12:22 PM
Subject: Re: Firmware 2.0.0 Solution
To: William Croft
Cc: Joel Murphy , Conor Russomanno
Nope, the bottle neck you mentioned is a limiting factor in sample rate.
To my knowledge and testing baud rates above 115200 do not work and interfere with the transmission of data. However, you can still pull and send data faster from the serial port if you raise your baud rates.
3.0.0 will enable faster sample rates through the use of 1) External serialport on the board 2) Networking device (e.g. ESP8266) with SPI
AJ
From: William Croft
Date: Thursday, September 1, 2016 at 3:14 PM
To: Andrew Keller
Cc: Joel Murphy , Conor Russomanno
Subject: Re: Firmware 2.0.0 Solution
On your upgrade page you mention the new faster baud rates. Does that also enable higher sample rates or more data to be sent over the air between RFduinos? I recall there was a bottleneck on the serial port between the PIC and the Device RFduino; is that still the case?
Thanks,
What is your experimental setup that requires the 500 hz? Yet one other firmware trick you could try is to do what is called 'oversampling' at 500 hz say, then time series average those down to a lower rate.