• Please note that Audiokarma will be offline briefly for updates early on Monday morning, September 28th.

RCA Engineers Test Telefunken 12AX7 Tubes against Their 12AX7

This thread is fascinating in showing how the details and unintended consequences of something that seems simple is anything but in practice. The technically curious could use more of it.
 
Re Telefunken, etc. There's assumptions of the "proprietary" nature of various processes and compouunds used in tube manufacture and what has been lost with the ending of production and retirement of the principals. One wonders how much can be recovered with reverse engineering, and improvements in production and chemistry.
 
Pio1980 said:
One wonders how much can be recovered with reverse engineering, and improvements in production and chemistry.
Click to expand...

The magic ingredient here, is money. How many tubes can be sold, and at what price? How much R&D can this really fund? I think we will never be at the level of the mid 20th century, when the whole world ran on tubes.
 
wyn palmer said:
Actually, I don't believe that your explanations produce any increase in clarity, so perhaps it's time to end this exchange. Thanks again for an interesting time.
Click to expand...

I suggest reading the work for how noise effects arise in think/thick film if you disbelieve what I wrote. You can easily verify the truth of it.
 
Pio1980 said:
I have to go back and re-read some of the posts, but, thinking of a mentor that used to work R&D at Varian/Eimac (R.I.P, Lane) , I'm intrigued by the coating/conductor cathode interface behavior vs a pure hot tungsten emitter, such as used in long life industrial vacuum electron valves. It seems, without the coating and interface issues, the simple hot emitter would be much easier to model and analyze. Thoughts?
Click to expand...

The noise in a tungsten or thoriated tungsten cathode varies as 1/(f^2). Because the cathode is homogenous the effects of boundary layers are reduced. But the tungsten is not perfectly pure, so some boundary layers do exist. Anytime metal oxides exist, with any sort of grain, there's going to be some semiconductor effects.

The heating cycle can recrystalize the tungsten filament which shortens the lifespan and changes the emission characteristics. This is why the influx of aluminum and nickel into the filament can over time embrittle it. It's why, in an alkali-metal cathode with a heater, the direction of the heater-to-cathode potential difference matters, because one way drags atoms from the nickel sleeve and alumina insulation into the filament and the other way drags tungsten into the alumina and nickel sleeve.

The structure of the tungsten cathode temperature also changes. When heated above the recrystallization point (Larson-Miller) thorium is transported to the surface to protect against recrystallization, at least to some extent. If the cathode isn't hot enough, because the tube is not at operating temperature or the voltage is too low, the tube is damaged. This phenomenon is well known in RF power tubes. Transmitting tubes, where thoriated tungsten were most studied, were initially powered down as a standby procedure from a white heat to a black heat (about 40 to 80% of normal operating temperature) which just hot enough was enough to keep the filament hot for rapid operation but not so hot that it glowed (red heat). The cycling from black to red, however, turned out to damage the tube almost as much as going from cold to red, so the consensus was to instead maintain the tube at 80% of normal operational temperature and take it to 100% when moving out of standby to operation. (I remember the figure that each such power-cycle on a transmitting tube dropped its lifespan by about 60 hours.)

Anyway, it's not so much modeling the tube's behavior in Spice, as reducing noise in practice. A great deal of work went into removing unwanted gas from the tube: hydrogen, oxygen, chlorine, etc. Any water in the mica would dissociate, but all of the metal had absorbed/adsorbed gasses. It's also why overheating tubes shortens the lifespan. All the gas driven out creates reactions. While gas creates problems for operation, particularly when the ions impact the cathode or the grid, it also creates problems in storage,
 
Pio1980 said:
Re Telefunken, etc. There's assumptions of the "proprietary" nature of various processes and compouunds used in tube manufacture and what has been lost with the ending of production and retirement of the principals. One wonders how much can be recovered with reverse engineering, and improvements in production and chemistry.
Click to expand...

The tube manufacturers were very aggressive in filing patents which explain the ingredients and processes.

I don't think much magic existed, and the assembly processes were quite crude. It was more low-cost labor to assemble the tubes.

For example, many patents existed for coating heaters. The general idea was that an slurry of alumina would be placed in a trough, the tungsten heater would be dragged/rotated through it until coated, the heater would then be fired, then bent, then inserted into the sleeve. This fractured some of the insulation in the process. Not the fanciest assembly technique ever used.
 
maxhifi said:
The magic ingredient here, is money. How many tubes can be sold, and at what price? How much R&D can this really fund? I think we will never be at the level of the mid 20th century, when the whole world ran on tubes.
Click to expand...

That's true to some extent, but modern tubes are made, and can be made, with far better quality materials which would be lower noise for certain aspects. It would be possible to order better alkali metal mixes, for example, with more uniform particle size and where the quality control could be trivially verified. No way to do that in the 1950s. Analysis was laborious and cost too much. Inexpensive computing has done a lot for quality control using mass spectrometry or gas chromatography, as have better manufacturing techniques.

We really do live in amazing times.

But shot noise, Johnson-Nyquist noise, flicker noise, etc. will always be present.

As a general rule higher quality generally costs more, which is why the 300B clones remain expensive. It's not just supply and demand, but that's some of it, true. Making tubes has a high amount of labor content and a number of steps. It's not like stuffing a 300 mm wafers into a fab and letting the magic happen untouched by human hands. Same thing happened with the cost of LCD and OLED TVs. The more automated the production, the lower the number of defects and the lower the cost. The modern LDC panels are 2,200 mm wide, which is a shade over 7 feet. That's a big piece of glass. But it's all automated which has hammered the price into the ground.
 
I think one of the useful takeaways is that noise exists in components.

One of the biggest sources of noise in an amplifier is resistor noise and capacitor noise, plus their non-linear distortion. Another source is using high-voltage Zener diodes to regulate voltages without understand how to reduce that noise using a stack. LEDs don't have such noise, which is why cathode bias using stacks of LEDs is used.
 
Retrovert said:
That's true to some extent, but modern tubes are made, and can be made, with far better quality materials which would be lower noise for certain aspects. It would be possible to order better alkali metal mixes, for example, with more uniform particle size and where the quality control could be trivially verified. No way to do that in the 1950s. Analysis was laborious and cost too much. Inexpensive computing has done a lot for quality control using mass spectrometry or gas chromatography, as have better manufacturing techniques.

We really do live in amazing times.

But shot noise, Johnson-Nyquist noise, flicker noise, etc. will always be present.

As a general rule higher quality generally costs more, which is why the 300B clones remain expensive. It's not just supply and demand, but that's some of it, true. Making tubes has a high amount of labor content and a number of steps. It's not like stuffing a 300 mm wafers into a fab and letting the magic happen untouched by human hands. Same thing happened with the cost of LCD and OLED TVs. The more automated the production, the lower the number of defects and the lower the cost. The modern LDC panels are 2,200 mm wide, which is a shade over 7 feet. That's a big piece of glass. But it's all automated which has hammered the price into the ground.
Click to expand...

I'm absolutely certain, that if someone were to strategically hire very intelligent people, pay them well, and give them funding and the funding and resources to do what they needed to do, we'd have better than NOS Telefunken tubes within a few years, maybe even sooner.

The issue though, is that the tubes being made now, are "good enough" for the applications which still use them. I'm not seeing a lot of incentive for anyone to invest money into R&D, when the main market is to re-create the products of the past, rather than do basic research into improving them.

The other thing is the market for tubes could dry up on a whim.. as the baby boomer generation ages, the popularity of guitar based music is going to continue to wane, and that's the main tube market at this point. Similarly, the use of tubes in high fidelity applications may go away too, it hasn't been mainstream for two generations already.
 
Pio1980 said:
Dr. van der Bijl doesn't much mention in these discussions. There's an IEEE pdf online on this part of his career.
Click to expand...

van der Bijl did a bit of early work on tubes, and wrote the definitive text in the 1920s, but he wasn't much publishing about noise. Schottky was the first to identify shot noise in vacuum tubes. He did all the heavy lifting on that.
 
maxhifi said:
I'm absolutely certain, that if someone were to strategically hire very intelligent people, pay them well, and give them funding and the funding and resources to do what they needed to do, we'd have better than NOS Telefunken tubes within a few years, maybe even sooner.
Click to expand...

The Western Electric name, in fact, was purchased and a company was formed to sell modern replicas of 300B tubes:

Made in the USA, too.

But then there are the versions from China and Eastern Europe, manufactured to higher quality standards because so much money is sloshing around at the high end.
 
Retrovert said:
The Western Electric name, in fact, was purchased and a company was formed to sell modern replicas of 300B tubes:

Made in the USA, too.

But then there are the versions from China and Eastern Europe, manufactured to higher quality standards because so much money is sloshing around at the high end.
Click to expand...

I totally agree that there's a market for high quality tubes. But look how long western electric is taking to sell just one tube type? How long has it taken them to get their act together to make just this one tube> At this point it's a high end consumer product, which has almost zero relevance to high technology in general.

For a true advancement in the state of the art, we'd have to see a major company getting a hundred million dollar purchase order for meeting a performance spec. But that's about as likely as the US Navy ordering a fleet of wooden battleships. Tubes attracted massive research dollars when they were an integral part of the military industrial machine... those days are gone for good though, and I think the best we can hope for is small incremental improvements on what still exists of the ashes of the old industry.
 
maxhifi said:
I totally agree that there's a market for high quality tubes. But look how long western electric is taking to sell just one tube type? How long has it taken them to get their act together to make just this one tube> At this point it's a high end consumer product, which has almost zero relevance to high technology in general.
Click to expand...

Building the first one is often the hard part. That requires funding, the factory, the supply chain, etc. Work it all out on the high-end product, pay off the investors, then expand.

maxhifi said:
For a true advancement in the state of the art, we'd have to see a major company getting a hundred million dollar purchase order for meeting a performance spec. But that's about as likely as the US Navy ordering a fleet of wooden battleships.
Click to expand...

True that.

Contracts to build a slew of ENIACs might also do the trick. Never can have too many ballistics tables...
 
You must log in or register to reply here.
Back
Top Bottom