You can run them simultaneously with two dongles. (Provided the RFduinos don't interfere with each other, different RF channels, etc.) However the sample clocks will not be synchronized. Hence one group of 8 channels could be offset from the other by up to ~ 4 milliseconds. That may be ok in your application. (But generally is considered undesirable, since measures such as coherence / phase will be thrown off. If you're just measuring band amplitudes, it could work.) You'd obviously have to hack together some code that combines two COM port streams.
re: "daisy chain them": You'd have to do major surgery on a main board to convert it into something like the small daisy board. So unlikely. And since the daisy board is always used with a 32 bit OpenBCI, your 8 bit board would not meet that requirement either.
William's comments are all correct, daisy chaining the hardware to keep them sample-accurate synchronized will require some hardware surgery...and even if you were successeful, the 8-bit microcontroller might not be fast enough to keep up. Running the two boards in parallel (via their own USB dongles) is the only decent way to do it.
If you use two boards, each with their own dongle, you would then run two GUIs to receive data from each board. While this is not ideal, it'll probably work. I've never tried it, though.
If you want to use the same GUI to receive data from both boards simultaneously, I got most of the way through modifying the Processing GUI to receive data from two boards. It might work for you, if you're interested. You can download it from here:
Comments
re: "daisy chain them": You'd have to do major surgery on a main board to convert it into something like the small daisy board. So unlikely. And since the daisy board is always used with a 32 bit OpenBCI, your 8 bit board would not meet that requirement either.