Multiple Open Ephys acquisition boards to one PC

[Adam Harris]

Hi all,

Sorry if this is a pretty beginner question: is it possible to connect and use two acquisition boards simultaneously with one PC? 

The total number of channels I will be recording from at any one time is 512, and these will always be across just four inputs, so could if absolutely necessary they could be fed into one board. 

Ideally though I'd have a total of 6 input channels (four coming from animal-specific sleep boxes, and two coming from behaviour apparatus) from which I can use any subset of 4 at any one time without having to move around too many cables.

Many thanks! 

[aaron.cue...@gmail.com]

Hello, 

Yes, it is possible to use multiple acquisition boards with a single pc. Just add two Rhythm nodes to the Open awohus GUI and each will use a different board. There a some caveats, though:

First, board order is somewhat arbitrary, as it depends on some usb descriptors. When you instantiate a Rhythm node it will use one of the boards, and the second the other, but loading the signal chain does not guarantee the nodes will be connected to the same boards. That is something you should manually check prior to acquisition. 

Second, the two boards will act as independent, asynchronous devices. Their internal clocks are not synchronized and the small delays on start and stop commands might cause misalignments. I would recommend an external ttl trigger routed to both boards to be able to synchronize them. 

Finally, due to the asynchronous nature and the block processing the GUI uses, trying to do operations mixing inputs from the two boards (eg.: using a channel from one board as a reference to channels from another) might cause issues. 

Apart from those details, everything else will work. You could record data from both boards as normal. Data will be stored in different files, though. But that is done automatically by the software. 

Let us know if you have any more doubts about how multiple boards could behave. 

Best, 

Aarón 

[Jon Newman]

Additionally,  if you want to use Bonsai, you can use the Bonsai.Ephys package along with some asynchronous-capabple (e.g. Zip) operators to get block-level synchronization between the two boards in software. As Aaron mentioned, you will still need some way to physically synchronously trigger acquisition starts. Also, board-to-board clock drift will always be present and may cause excessive buffering in the Zip node and resultant degradation of real time performance for one device over long time periods.

- Jon

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/d76b4fe8-2006-4db9-9bb0-eb62e0ab19adn%40googlegroups.com.

--

Jonathan Newman

Postdoctoral Fellow, MIT

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

[aaron.cue...@gmail.com]

Unfortunately, I believe the Bonsai.Ephys package seems to support only one board. I will probably issue a pull request in a couple of days trying to address this, as Bonsai is indeed a excellent alternative for handling asynchronous streams.

[Adam Harris]

Thanks so much for your responses! I will be getting this working later in the week. Aaron, regarding data being saved in separate files, is it possible to do that anyway (e.g. one file per animal) using multiple recording nodes?

Cheers 

-Adam

[aaron.cue...@gmail.com]

Yes, with multiple record nodes you can save in as many places and with any channel configuration you need. The only limitation is mixing channels from two sources in a single record node, in which the node will create two parallel datasets (two sub folders in binary format), but you can manually select completely different locations with multiple nodes with no issue.

Best,
Aarón