MikeDB: Yes, definitely interested in discussing more. FYI: I just came back from the big neuroscience conference (SFN) and there was a lot of talk about IMEC chips. Long-story short, they were paid from gov funds (Welcome Trust, NIH) to develop and completely open-source the API etc, but have been slow-rolling all the specs. However, it seems we're pretty close to having complete input:output documentation from what I was told (and Open-Ephys has reverse engineered them about 1 year ago anyways). The chips are here, https://www.neuropixels.org/specialty-products, and cost about $500/piece.
From what people tell me, the IMEC 384 channel chips draw about 10-20mA @ 1.8V for 30Khz sampling, whereas the Intan 32chan draws about 30mA @ 3.3V for 30Khz sampling. This means the IMEC is > 1 order of magnitude lower more efficient, and even overall gives 10x # of channels for even lower power. Bottom line, I'd love to try mounting some of these chips and coupling them with a Nordic MCU etc. (I'll get in touch shortly.)
I still think there's a lot of space for integrating neurotech into the RPI techstachs and with these desci movements it may something to consider (at some point...).
From what people tell me, the IMEC 384 channel chips draw about 10-20mA @ 1.8V for 30Khz sampling, whereas the Intan 32chan draws about 30mA @ 3.3V for 30Khz sampling. This means the IMEC is > 1 order of magnitude lower more efficient, and even overall gives 10x # of channels for even lower power. Bottom line, I'd love to try mounting some of these chips and coupling them with a Nordic MCU etc. (I'll get in touch shortly.)
I still think there's a lot of space for integrating neurotech into the RPI techstachs and with these desci movements it may something to consider (at some point...).
Statistics: Posted by catubc — Mon Oct 14, 2024 5:40 pm