Open e-phys max channels?

[Tyler Benster]

I'm currently figuring out which headstage to get for use with an MEA. I see that Open Ephys supports 4 x 128 channel headstages (although the May newsletter makes it sound like only 64 channel headstages are supported at the moment?). Is it possible to stream 512 channels over USB2?

[Nikolas Karalis]

Hi Tyler,

from my (non exhaustive) tests, 128 channels are fine.
However when going for more. things start slowing down.

I can't recall at the moment what was the maximum number of channels for which I would get decent performance, but it was less than 256.

Thankfully there is an FPGA version with USB3, but unfortunately I think at the moment it is not supported.

Aaron, do you think it could just work with the newer FPGA or it would require updating the code?

Cheers,

Nikolas

On Sat, Jun 6, 2015 at 11:08 PM, Tyler Benster <tbe...@gmail.com> wrote:

I'm currently figuring out which headstage to get for use with an MEA. I see that Open Ephys supports 4 x 128 channel headstages (although the May newsletter makes it sound like only 64 channel headstages are supported at the moment?). Is it possible to stream 512 channels over USB2?

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+...@googlegroups.com.
To post to this group, send email to open-...@googlegroups.com.
Visit this group at http://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/82df16b7-269c-45d2-bace-9df928d36c52%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Nikolas Karalis - http://www.nikolaskaralis.gr

Neuroscience Graduate Student

Ludwig-Maximilians Universität München 

Centre for Integrative Neuroscience -  Universität Tübingen

Neuroscience Graduate
Université Bordeaux 2 - Charité Universitätsmedizin Berlin
Applied Mathematics and Physics Graduate 
National Technical University of Athens, Greece

[Jon Newman]

Some things to note concerning MEA interfacing:

1. There is currently not an mechanical interface for planar MEAs for slice preparations and dissociated cortical networks available. However,

 - There is a group of people at Stanford in John Huguenard's Lab, that is considering developing a planar array interface for Intan chips and targeting the OE acquisition system. I'm not sure if they are on this mailing list. You may need to contact them directly.

 - An MEA interface would be quite simple with the exception of the mechanical design for seating the planar array. This type of setup brings its own challenges that are unique compared to the in-vivo world which is what the OE headstages were designed for. I would highly recommend studying and learning from commercial designs (e.g. those of mulitchannel systems) before attempting to design your own. If you do not make the mechanical design solid, you will be fighting loose connections, dirty contacts, and broken 400 dollar a pop MEAs for 90% of the time and doing recordings 10%. Electrically, the simplest thing to do is to create an adapter PCB that passively routes the MEA contacts to 4 omnetics connectors where you could plug in 4 of the existing OE headstages. 

3. 512 channels is not possible on USB2. You will need to wait for USB3 support.

On Sat, Jun 6, 2015 at 5:08 PM, Tyler Benster <tbe...@gmail.com> wrote:

I'm currently figuring out which headstage to get for use with an MEA. I see that Open Ephys supports 4 x 128 channel headstages (although the May newsletter makes it sound like only 64 channel headstages are supported at the moment?). Is it possible to stream 512 channels over USB2?

--

You received this message because you are subscribed to the Google Groups "Open Ephys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+...@googlegroups.com.
To post to this group, send email to open-...@googlegroups.com.
Visit this group at http://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/82df16b7-269c-45d2-bace-9df928d36c52%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Jonathan Newman

Postdoctoral Fellow, MIT

http://www.mit.edu/~jpnewman/

[Aarón Cuevas]

Hi,

The absolute limit for USB2 is 256 channels (eight 32-channel streams). There are hard coded limits on this in both the firmware and the software so even if you plug more than that, only the first 8 streams (256ch) will be active.

USB3 support through the XEM6130 FPGA is something I'm working right now. It needs changes to both the Rhythm firmware (the bitfile) and the software, so it's not just a matter of removing the limits and changing the FPGA, but it will allow up to 512 channels, which is the maximum number of channels that can be plug into the board:128 ch (4x32ch streams) in each port. I expect to have it working this month.

For the record, Intan's 128ch headstage is basically two RHD2164 mounted into the same board with exactly the same wiring as two 64ch headstages plugged to a single port through a double headstage adapter (using both MISO lines on the board port), behaving exacly like that. Thus, the GUI should recognize a 128ch channel headstage as two 64ch ones. Since there is no way for the software to distinguish a single 128ch headstage from two 64ch ones (they are electrically the same) and only one 128ch hedstage is allowed per port (it uses both MISO lines, so it can't be used with the double adapter) we don't currently have plans to add any specific change for that, nor should it be needed. That said, we haven't tested any yet, so any test info on it would be appreciated.

Best,

Aarón.

On Saturday, June 6, 2015 at 11:36:15 PM UTC+2, Nikolas Karalis wrote:

Hi Tyler,

from my (non exhaustive) tests, 128 channels are fine.
However when going for more. things start slowing down.

I can't recall at the moment what was the maximum number of channels for which I would get decent performance, but it was less than 256.

Thankfully there is an FPGA version with USB3, but unfortunately I think at the moment it is not supported.

Aaron, do you think it could just work with the newer FPGA or it would require updating the code?

Cheers,

Nikolas

On Sat, Jun 6, 2015 at 11:08 PM, Tyler Benster <tbe...@gmail.com> wrote:

I'm currently figuring out which headstage to get for use with an MEA. I see that Open Ephys supports 4 x 128 channel headstages (although the May newsletter makes it sound like only 64 channel headstages are supported at the moment?). Is it possible to stream 512 channels over USB2?

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.

To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+unsubscribe@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at http://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/82df16b7-269c-45d2-bace-9df928d36c52%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Tyler Benster]

Great info all around, thank you!

Nikolas—thanks for the info. We’ll start below 128 and experiment upward, will keep you posted.

Jon—appreciate the direction. Exploring options for connecting mechanical design now. I also dropped John Huguenard a cold email. Do you happen to know who at the lab is looking at this?

Aarón—thanks for clarifying the hardware stack. Great to hear USB3 support is on the way! Have you seen the slowdown that Nikolas mentioned over 128 channels? If so, is the solution to get a more powerful FPGA or is the cause something else?

Thanks!

Tyler

Tyler Benster | tylerbenster.com

Twitter: @tbenst | LinkedIn: tylerbenster

US mobile: +1 (206) 919-8004 | International: +1 (206) 629-8762

On Sat, Jun 6, 2015 at 9:54 PM, Aarón Cuevas <aaron.cue...@gmail.com> wrote:

Hi,

The absolute limit for USB2 is 256 channels (eight 32-channel streams). There are hard coded limits on this in both the firmware and the software so even if you plug more than that, only the first 8 streams (256ch) will be active.

USB3 support through the XEM6130 FPGA is something I'm working right now. It needs changes to both the Rhythm firmware (the bitfile) and the software, so it's not just a matter of removing the limits and changing the FPGA, but it will allow up to 512 channels, which is the maximum number of channels that can be plug into the board:128 ch (4x32ch streams) in each port. I expect to have it working this month.

For the record, Intan's 128ch headstage is basically two RHD2164 mounted into the same board with exactly the same wiring as two 64ch headstages plugged to a single port through a double headstage adapter (using both MISO lines on the board port), behaving exacly like that. Thus, the GUI should recognize a 128ch channel headstage as two 64ch ones. Since there is no way for the software to distinguish a single 128ch headstage from two 64ch ones (they are electrically the same) and only one 128ch hedstage is allowed per port (it uses both MISO lines, so it can't be used with the double adapter) we don't currently have plans to add any specific change for that, nor should it be needed. That said, we haven't tested any yet, so any test info on it would be appreciated.

Best,

Aarón.

On Saturday, June 6, 2015 at 11:36:15 PM UTC+2, Nikolas Karalis wrote:

Hi Tyler,

from my (non exhaustive) tests, 128 channels are fine.
However when going for more. things start slowing down.

I can't recall at the moment what was the maximum number of channels for which I would get decent performance, but it was less than 256.

Thankfully there is an FPGA version with USB3, but unfortunately I think at the moment it is not supported.

Aaron, do you think it could just work with the newer FPGA or it would require updating the code?

Cheers,

Nikolas

On Sat, Jun 6, 2015 at 11:08 PM, Tyler Benster <tbe...@gmail.com> wrote:

I'm currently figuring out which headstage to get for use with an MEA. I see that Open Ephys supports 4 x 128 channel headstages (although the May newsletter makes it sound like only 64 channel headstages are supported at the moment?). Is it possible to stream 512 channels over USB2?

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.

To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+...@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at http://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/82df16b7-269c-45d2-bace-9df928d36c52%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.