4 ms·

Website is definitely underwhelming, but there's some documentation that appears reasonably detailed and plausible: Data sheet: http://nebula.wsimg.com/35f7566

Website is definitely underwhelming, but there's some documentation that appears reasonably detailed and plausible: Data sheet: http://nebula.wsimg.com/35f756676bf3290bebff9c9ac0f9e3ea?AccessKeyId=64577CB1C10F8DCEF8A3&disposition=0&alloworigin=1 http://nebula.wsimg.com/35f756676bf3290bebff9c9ac0f9e3ea?Acc... User manual: http://nebula.wsimg.com/6dba75009009af7a59036365876b3f66?AccessKeyId=64577CB1C10F8DCEF8A3&disposition=0&alloworigin=1 http://nebula.wsimg.com/6dba75009009af7a59036365876b3f66?Acc... Demo video: https://www.youtube.com/watch?v=0wQviPeaNpw https://www.youtube.com/watch?v=0wQviPeaNpw

theresistor 11y ago
After a bit more digging, it appears to be a silver chalcogenide memristor, as described here: http://scholarworks.boisestate.edu/electrical_facpubs/73/ http://scholarworks.boisestate.edu/electrical_facpubs/73/ This is very different materially from the TiO2 based ones used in HP-derived designs, though I'm not sure what differences that will lead to in practice.
- cbennett 11y ago
  TLDR: There is a major mismatch between the device physics of the academic group and what is being marketed/sold. If the actual device they were selling represented what they did in this technical work, I would be very, very excited. The I/V curve looks great. BUT, compare that figure to the I/V curve represented in the longer datasheet (linked to in parent comment). A one million fold increase in resistance in 5nS- That isn't analog at all. It could be fairly emulated as almost a flip-flop with just a few transistors!! Although a binary memristor at scale might be more interesting as a memory (eg ReRAM), at this scale meant for neuromorphic analog applications, it becomes utterly uninteresting. So to be honest I am baffled how they went from such a promising inspiration to a completely bland packaged device. Two possibiliies come to mind: 'they' are entirely different groups, or a major mis-marketing is occurring in the interest of gaining some funds for future R&D.

theresistor 11y ago

After a bit more digging, it appears to be a silver chalcogenide memristor, as described here: http://scholarworks.boisestate.edu/electrical_facpubs/73/ http://scholarworks.boisestate.edu/electrical_facpubs/73/ This is very different materially from the TiO2 based ones used in HP-derived designs, though I'm not sure what differences that will lead to in practice.

cbennett 11y ago
TLDR: There is a major mismatch between the device physics of the academic group and what is being marketed/sold. If the actual device they were selling represented what they did in this technical work, I would be very, very excited. The I/V curve looks great. BUT, compare that figure to the I/V curve represented in the longer datasheet (linked to in parent comment). A one million fold increase in resistance in 5nS- That isn't analog at all. It could be fairly emulated as almost a flip-flop with just a few transistors!! Although a binary memristor at scale might be more interesting as a memory (eg ReRAM), at this scale meant for neuromorphic analog applications, it becomes utterly uninteresting. So to be honest I am baffled how they went from such a promising inspiration to a completely bland packaged device. Two possibiliies come to mind: 'they' are entirely different groups, or a major mis-marketing is occurring in the interest of gaining some funds for future R&D.