For a little while vacuum tubes and semiconductors were fighting a heated battle for dominance, with bipolar junction transistors and 1959’s RCA Nuvistor both allowing you to build a compact circuit with relatively low power usage and no high voltages. Although we now know that semiconductor technology won out overwhelmingly, that doesn’t mean that you cannot build a brand new Nuvistor board in 2026, as [Eric Schlaepfer] AKA [TubeTimeUS] recently did.
Nuvistors saw their most use in small-signal radio frequency applications, like VHF and UHF, with excellent low-noise characteristics that saw them used until the early 1970s in television sets, radios and oscilloscopes, as well as in space probes like the 1960s US Ranger Moon missions, so by that metric they had a good run.
Nothing so exciting is built in this video, sadly, but alongside a breakdown on how nuvistors work, we do see a discrete 555-style timer built using a gaggle of tetrode nuvistors, giving a pretty good idea of what using them in a project is like. Being a vacuum tube at its core, nuvistors still have the heater element, which is what gives vacuum tubes their reputation for being slow to start working and large current draw.
Despite their drawbacks, nuvistors still have a range of benefits compared to modern-day transistors, including being practically immune to electrostatic discharge (ESD) and electromagnetic interference (EMI) all the way up an EMP that will destroy most semiconductor electronics.
That said, the somewhat limited 8-nuvistor implementation of the 555 had to get a few extra pins for the heater supply, which burns up about 7.5 Watt just to allow the circuit to function. Terming it the ‘hollow-state 555 timer’, it works effectively just like any semiconductor 555, just with that extra power cost and of course no significant prospect of making it smaller, barring a semiconductor evolution as with the pixels-sized CRTs in the SED and FED type displays.
We covered the nuvistor before, including a great reference on this device, and its history that was much longer than people often assume today, as well as the vacuum tubes we use every day in for example our microwaves.
That’s a very cute giant DIP package-style board. Yes I know that “DIP package” unpacks to “Dual In-line Package Package” kind of like saying “ATM machine,” leave me alone okay
DIP can also be used an abbreviation for “Dual Inline Plastic”, in contrast ti dual inline ceramic.
Cer-dip was what we called it where I worked. Plastic was p-dip.
That would have to be a DIC.
I rescued some nuvistor filled gear, and would like to build something useful with them. A high impedance input buffer might be a good thing. An e-field probe or lightning detector might be another.
As a 72 yr old engineer who cut his Twoteeth on tubes:
KUDOS TO BIG TIME!
IM GUESSIN 4 TUBES WERE USED AS 2 DIFFERENTIAL MasPS TO DO 1/3 AND 2/3 Vcc COMPARATORS.
Then 2 more uses as the SR flip flop.
One more as pin 7 the capacitor discharge pin?
Designing a 100 MHz RF amplifier in 1975, I looked at nuvistors and found them to have noise characteristics no better than the best MOSFETS. The poor transconductance of vacuum devices make low noise voltage difficult. Semiconductor noise performance and bandwidth has continued to improve over the last 50 years.
Hm. The way it reads, it also could be interpreted the way that the old nuvistors were still able to compete with the best MOSFETs of the time.
That’s not bad, I think.
I often wonder how nuvistors would look today if the US industry (and the far east) hadn’t basically killed the tube/nuvistor through mass-production of low quality transistors.
Other places on earth still continued to use the tube technology for a little longer,
but international trade caused an end to this.
There were no “untouched”, isolated regions anymore where tube technology could have evolved while being undisturbed.
No far island or isolated mountain region etc.
This is very depressing, I think.
Because tube technology was more elegant than transistor technology, more ingeniuous, more human.
It was an improvement over the existing solid-state/semiconductor technology (detector crystal, glass diode).
It also allowed for a higher maximum frequency, was electrically robust.
The FET is closest to the thermionic valve, I think.
It’s basically an emulation of the real thing.
Maybe world’s first “emulator”. :)
Funny how all it takes is a change of perspective to go from “it kinda sucked” to “wow, it was pretty damned good”! 😊
“I often wonder how nuvistors would look today if the US industry (and the far east) hadn’t basically killed the tube/nuvistor through mass-production of low quality transistors.”
It wouldn’t have changed much and the far east production had a negligible impact. Transistors were also mass produced in Europe, but most of them ended up in equipment where the general public knew not much about the inner workings, like telecomms and computing. (You could build thousands of radios from the transistors in just one Telefunken or Bull mainframe.)
Nuvistors aren’t all that great, they are expensive to manufacture and not easy to design with. When they arrived on the market, they had to compete with cheap mass produced frame-grid UHF tubes for TV tuners (PC86, PC88, etc.) so they never made it into consumer appliances outside the USA. And not much later, the industry already started moving to integrated circuits…
The fact remains that the Nuvistor’s undesired characteristics outweigh even a 1970s era MOSFET. Worst is the heater which wastes 0.95 Watt doing nothing more than making the cathode emissive. The Beam power 7587 Nuvistor requires a plate supply of 250 Volts while only managing g a transconductance of only 10.6 ma/V. That yields a gain of 58 in the reference design. The plate impedance is an unfriendly 200K Ohms. The plate dissipation of up to 2.2 Watts is probably the most useful statistic on the data sheet.
I still have a few Nuvistors in their original RCA red cardboard and styrofoam boxes in my shop. They’ve been there since at least the 1980s. I have never found a project where is Nuvistors would be more suitable than an easier to work with solid state component.
Correction the recommended plate voltage is 125 Volts.
That’s exactly why I wondered about the future of the Nuvistor that never was! :)
Personally, I had quite a few situations in which battery tubes such as EF98 were better than modern transistor.
They were very good for sound card based SDR reveivers, for example.
Because unlike transistors, they can go softly into saturation.
In the 1960s, UHF IV/V tuners with germanium mesa transistors like AF139 had better noise figures than their predecessors with frame-grid tubes. These transistors finally killed tubes in small signal RF stages.
But was it because of transistors in particular
or because of a different assembly style, such as use of PCBs or chamber technique (sections in a metal can that are RF isolated)?
The lack of a power supply in many tube based circuits might been another source of noise.
– Back then, an external separation transformer could help to compensate for fluctuations in the AC mains.
The idea of directly driving things from AC was a deficiency of its time.
I assume the 1960s people were either ignorant or foolish, not taking the dangers of AC seriously.
They still used two-wire AC plugs widely without safety earthing..
Shamefully, radio amateurs are guilty of this, too.
I’ve seen homebrew devices that had cables/plugs with two leads, despite them using a metal chassis.
The worst was them using Schuko plugs with two leads connected (no protective earth).
Some devices were built that way as late as the 1980s. Yuck.
Such careless people are no role models to the next generation! :(
But let’s go back on topic..
By about the 1970s, TV sets and similar devices got power supplies and ICs as a standard, which was due.
The change also was better for tube circuits, I’d say,
because tubes and transistors can work hand ind hand very well.
Stable, regulated power also made tubes live longer.
If memory serves, one has high impedance output (tube; input also high impedance) while the other has a high impedance input (transistor; output low impedance).
So the tube could work as a pre-amp to a transistor circuit.
In a radio receiver, for example.
But backt to PSUs.. Portable b/w TVs popular in the 1970s/1980s used PSUs and often had an 12v input, too, because everything was upconverted from 12v DC anyway.
The then-new semiconductors and ICs helped to make things more stable.
Especially the IC supported flyback circuits were more reliable.
About tube era TVs and power supplies..
I forgot to mention, the lack of PSUs also caused house fires.
The older TVs without PSUs had big ceramic resistors underneath, which easily got hot and eventually would catch fire.
These series resistors were there because of the heating, mainly.
But also to lower voltage and power of other things, incl. early transistor circuits. Resistors, resistors everywhere.
In retrospective, this setup was barbaric, really.
This could have been done better with the existing technology of the 1940s to 1960s, with rectifiying tubes and and stabilizing tubes and what not.
But it was considered too sophisticated for mass-produced domestic appliances, I guess.
Studio equipment probably was built correctly at large, I assume.
Anyway, the advent of both linear power supplies and switching power supplies were an improvement here.
Including stabis, such as 78xx series.
Tubes are much noisier than transistors because of simple physics. Higher operating temperatures and higher internal impedances mean higher thermal noise. And then there’s emission noise in the cathode, secondary emission noise, ionization noise and what not. These things were already optimized with tubes available in 1960, the frame-grid PC88 has F=10 dB @ 800 MHz, which is not bad for a cheap TV tube. Germanium Mesa UHF transistors could do the same nearly from the start (AF139: F=8 dB @ 800 MHz) and somewhat later the AF239 brought the noise figure down to <5dB. Add to this the lack of heater-induced hum modulation and less thermal drift, and it becomes obvious that tubes were no longer an option for UHF tuners.
Regarding hot chassis appliances: These were not inherently unsafe, the are just class-II-devices, most Western European sets were properly built to VDE and similar standards. Dropper resistors were usually located a the back of the chassis, not underneath, and they were of the fused type, they would go o/c on overheating. House fires were not caused by hot chassis TVs, but by badly maintained electrical installations, bridged fuses and general negligence and stupidity of their owners. A flowerpot on a doily on top of the set…
I designed a 6 meter down converter using a 6CW4 Nuvistor in 1960. I have a Fisher Stereo Console with 3 6CW4s run conservatively at 40 Volts. Still operating. At that time the only other low noise choice was bipolar transistors. Those are exponential trans conductance with high orders of harmonic and intermodulation. The tubes are 2/3rds power and the output swing small. It took years for MOSFETS to catch up.
Warm friendly mechanical electrical naturally oscillating terra herts switching angels . Vacuum tubes showed me how to love electronics.