Asianometry always does a great job with this stuff, although I find myself grimacing when he speaks positively about Sarnoff. Nevertheless the tube history is good (although he fails to mention that De Forest never figured out how his tubes worked).
I have four Nuvistors! They are in a 2 meter converter from Parks Electronics of BEAVERTON, OREGON. I may have this wrong, but I think the video said the Nuvistors were very expensive. This can't be right. Gemini AI says the 6CW4s in my Parks device sold for $1.50 to $2.50 in 1961.
Three cheers for the Nuvistor! I may try to see if they still work.
Another great piece by Jon Y! Nice pick, Bill.
ReplyDeleteHaving grown up in NNJ, one of my Amateur Radio Elmers was formerly on the RCA Nuvistor development team. He had told me (around '68) that the Nuvistor was two years too late. That makes sense as the planar process hit around 1959, and transistors were already being designed exponentially into products.
The Nuvistor was just one case of many where where RCA flopped. Another was LCD's. Who invented the first ones? RCA Princeton. But then they gave it all away. A few years later, a chemist that worked on the LCD team at Princeton told me "they are more interested renting cars and making crappy mainframes!" (--that was a shot at the "Spectra 70"). In other areas RCA hit home-runs- think the CD-series CMOS family, the Overlay Transistor, and the CA-linear integrated circuits.
Nuvistors are an example of hanging on to a technology that one knows well and is comfortable with, but cannot make the big leap, in this case to the next requirement of the 1960's high speed computing.
Likewise, the AI and GPU people today are at that "1959 point"- they are hanging onto silicon, driving Moore's Law as far as they can. Maybe they will do well heading towards 1 nanometer with CFET 3-D processes, but are there a disruptive technologies being developed? Could quantum computing (IBM,etc) or analog processing (MIT) or something else be the key to energy-efficient neural networks? When one technology runs out of steam, look to another solution.
Today, we are dependent upon vacuum tubes where nothing else is practical. Look at the weather radar (WSR-88D). Those images come courtesy of a whopping big 750 KW peak Klystron!