neuropixels
neuropixels copied to clipboard
cortex-lab
commutated tethers
Hi Neuropixies, During long free-moving chronic recording using passive probes alongside implanted optic fibres I use a concentric electro-optical rotary joint (Doric) to prevent tangled fibres/tethers. For the Neuropixel recordings, with the skinny coaxial tether to the probe alone for just the ephys it's manageable without, but add in a fibre with minimum permissible bend radius and the tether soon becomes a tangled mess.
My question is: Has anybody been brave (foolhardy?) enough to try cutting a tether to introduce a rotary joint? I've been waiting for a rat to bite a Neuropixel tether so I can repair it with the commutator inline to find this out, but until this happens, my main concern is threat to data integrity from electrical noise artefacts. Very rarely with passive probe recordings using commutated Intan SPI tethers, rapid head rotations can appear in the data as corrupted data frames (not sure if this is from the data lines or due to brownout), but the system recovers and the acquisition doesn't crash, I'm not sure if I can expect the same behaviour from the Neuropixel base-station.
Any insight is much appreciated... cheers, Aleks
IIRC there are 5 cores on that cable and all 5 are used. I can't remember exactly what the pinout is, but as far as I recall, two would be DC (GND/VCC) and at least two are for the LVDS pair running at 25MHz (?).
So as long as the HF signals were connected through a high-frequency slip ring connector (rotary joint) rated to at least 25MHz, you should be fine. I've seen (and used) HF slip ring connectors extending into the 10GHz+ range. You'd just need to find one capable of passing five signals - at least 2 of them to 25MHz.
(n.b. all these numbers are off the top of my head - you'd definitely be best double-checking off someone who actually has the specs, or better still checking with an oscilloscope...)
Thanks for the info - I'll investigate further. As I mentioned, the commutator in question can carry the signals from an Intan headstage running a SCLK signal at (off the top of my head) 21MHz. I don't know its max frequency rating but will enquire. It carries 12 channels so I should be well covered. I think I will pipe through some fast clock pulses under rotation before cutting the cord however.
I would definitely appreciate an update on this if anyone manages to pull it off!
Hi Rich, Ashley, Sorry for missing this reply earlier. I did get one tether/ commutator combo working through one part of the FPGA startup handshake procedure. It then threw errors before I could start acquisition. This was with some breakout wires on the chain, causing reflections due to improper signal termination. perhaps I'll have more luck soldering the connectors directly. Tangles are a real problem for me in my task and have led to lots of broken optical patch fibres too so am working on getting our opto-electric (Doric FORJ) working asap. I'm travelling for a week but will be back with a proper write up soon. Cheers, Aleks On 3 Feb 2018, at 18:25, Rich Gardner <[email protected]mailto:[email protected]> wrote:
Aleks, did you have any luck with this?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/cortex-lab/neuropixels/issues/6#issuecomment-362842493, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AJWM_eaEYCgtA5jxruwngGiafG1uruYtks5tRKSkgaJpZM4P104J.