Filament voltage vs saturation

[fdesalvo]

We know that at a certain anode voltage, no increase in current can be attained - saturation. This could not only be relative to the capacity of the anode to collect electrons and the cathode coating's capacity to emit them, but perhaps a limitation imposed by our standard of heater voltage, right?

When we say we are at saturation, does anyone here distinguish which element of the triode is saturating? If the cathode is at saturation due to the 6.3v limit, has anyone experimented with higher filament voltages to increase current output while staying below cathode emission/anode collection saturation? I have already considered ideal operating conditions are the best compromise between tube life and performance, but I'm curious and still learning.

Back in the day they used to use a bit of thorium to improve emission efficiency. Further research found that substituting a mix of strontium and barium oxides would improve electron emission - and at even lower temps (prob marked by the move down to 6.3v). Would also be interesting to see if there were other elements available that could prove to be even more efficient.

[eschertron]

My simple understanding is that the cathode's ability to produce electrons is the limiting factor in saturation.

What kind of tubes are you thinking about experimenting with? The common 12AX7 will melt down long before the cathode reaches saturation at normal operating conditions. That's the engineer's 'safety factor' in play to prevent old tubes from losing enough emission to allow saturation to become limiting.

Maybe real *low* heater voltages could allow you to reach saturation before the dissipation limit is reached. Or maybe use the heater as a control element for modulating the output! Bandwidth would be low, however.

[fdesalvo]

Any triode will do (with an indirectly heated cathode)!

I suppose that in any scenario presented by my questions, the grid would become useless or a hindrance.

[Old Tele man]

Lookup the definition and characteristics of the vacuum tube parameter PERVEANCE ( G or K).

[fdesalvo]

Will do - thanks, OTM.

Wait a second, are you referring to "space charge" - the bit of initial resistance caused by electron build up at/near the anode?

[R.G.]

Oxide cathodes are able to generate very much more electrons than normal operation requires. The electrons pop off the cathode and form a cloud of electrons just over the cathode. This is a good thing in most cases, as it shields the cathode from incoming gas ions which have had their outer electrons sucked away by bumping into the plate. Helps prevent cathode degradation by ion bombardment to some degree.

We know that at a certain anode voltage, no increase in current can be attained - saturation. This could not only be relative to the capacity of the anode to collect electrons and the cathode coating's capacity to emit them, but perhaps a limitation imposed by our standard of heater voltage, right?

I don't think so. It is true that higher temps cause more emission, but it also cuts cathode life, in some cases dramatically. And in any case, a decent-condition triode can never deplete the space charge cloud in normal operation.

When we say we are at saturation, does anyone here distinguish which element of the triode is saturating?

IIRC, what saturates is the ability of the voltage and mechanical spacings to accelerate electrons from cathode to plate, not a saturation of any single element's ability to handle electrons. It is possible to cause even more electrons to flow from cathode to plate by running the grid positive with respect to the cathode. The grid then conducts, but plate current still increases as grid voltage moves positive until either the space charge is depleted (which is the point at which the cathode emission limit is reached) or the grid melts, or, even more likely, the source impedance driving the grid can no longer eat the electron flow and hold the grid positive.

One of the little-known things about causing distortion with triodes is that you can make them clip more softly by driving the grid with a low impedance so the grid can be driven positive and still increase plate current as the grid goes a little positive.

If the cathode is at saturation due to the 6.3v limit,

In general, it isn't. While it's possible to concoct a scheme to do that, that's not how most amps run.

has anyone experimented with higher filament voltages to increase current output while staying below cathode emission/anode collection saturation?

Yes. All of the tube makers during the Golden Age messed with the compromises between cathode temperature, cathode chemistry, and cathode area to get the best-compromise of tube life and operation.

I have already considered ideal operating conditions are the best compromise between tube life and performance, but I'm curious and still learning.

There is some older literature I read that looked at the effect of heater voltage on tube life. IIRC, cathode/heater life goes down as the fourth power of the increase of heater voltage above the specified rating. It goes UP the same way until you cut emissions so much that the space charge is depleted in normal operation. You don't want to deplete the space charge cloud. Bad ju-ju for tubes.

Back in the day they used to use a bit of thorium to improve emission efficiency.

The important number here is the "work function", that being the thermal energy needed to get electrons to jump off a metal surface into a vacuum. Metals vary, sometimes a lot, in their work function. Thorium in the cathodes lowered the work function until a white-yellow glow temperature would boil enough electrons off. Before thorium, they used tungsten filaments run white-hot to get enough electrons. Thoriated tungsten cut that temperature back a fair amount.

Further research found that substituting a mix of strontium and barium oxides would improve electron emission - and at even lower temps (prob marked by the move down to 6.3v).

Coating cathodes in a mixture of barium and strontium oxides left the oxides as the emitting surface. These oxides were found to have an even lower function than thoriated tungsten, and that let the temperature to get that space charge cloud into place be cut a lot. The trade-off is that lower temperatures gets you (sometimes MUCH) longer heater and cathode life. Thoriated tungsten had to run so hot that the cathode itself had to be heated by internal current flow. The cathode WAS the heater. Oxide coated cathodes could be run at a lower temperature to give much longer life by emitting enough electrons at a lower temperature. The ideal, searched for by many, was the cold cathode - no external heating needed to raise it to emissions temperature.

As an interesting aside, semiconductor manufacturing did finally produce a cold-cathode tube with sufficient emission - the field-emission triode. Silicon micromachining could make arrays of atomically sharp spikes that would spray off significant amounts of electrons by the field intensity at the sharp tips. They actually made a few field emission triode arrays way back there. I haven't heard from the field in a long time. Probably worth looking up.

Temperature is an unfortunate necessity to get enough electrons. Any way to lower the cathode temperature prolongs its life.

Would also be interesting to see if there were other elements available that could prove to be even more efficient.

That's an old, well-plowed field. There was a lot of effort by Very Smart Guys back in the Golden Age to make better tubes, the same way lots of Very Smart Guys now search semiconductor physics for ways to make better, smaller transistors - and for the same reasons. Back in the 40s, 50s, and 60s, metallurgists and materials physicists thrashed through every metal (there are only 92 naturally-occurring slots in the periodic table) and compound they could think of to try to get better emissions.

[loudthud]

The first widely available curve tracer setup for tubes was the Tektronix 570. On the front panel was a big rotary switch to set the heater Voltage and there was a Variac and meter that let you dial in the Voltage in exactly. If you adjusted the heater Voltage while curves were being displayed, there was a corrisponding movement in the curves. In most cases a 10% increase in heater Voltage increased plate current by roughly 10%.

[teemuk]

Filament voltages were simply chosen to be compatible with common battery voltages. I would imagine tube manufacturers then researched to find the most optimal filament material characteristics / filament construction within that given limit.

And yes, filament voltage/current certainly affects overall tube operation (e.g. gain, point of clipping..) There's even a Music Man amplifier patent where they address and exploit that particular characteristics as early "power scaling" setup. ...But since -reliable- filament voltage has very narrow margin (for sake of reliability / tube life you don't want it to be too high or too low) I would only encourage to -experiment- with the phenomenon but not to build or design anything that permanently relies on that specific characteristic. It will eat your tubes faster, and given wide variance of tube characteristics, how much -consistency- could one practically expect from this feature? Could a different filament construction already affect this characteristic? I have a hunch this is largely undocumented territory.

[teemuk]

Back in the 40s, 50s, and 60s, metallurgists and materials physicists thrashed through every metal (there are only 92 naturally-occurring slots in the periodic table) and compound they could think of to try to get better emissions.

This. Plenty of research and development also went to optimizing tube constructions to most effectively dissipate all that waste heat. Look inside a tube and you probaly see a "heatsink" or two with a surface and texture optimized for efficient heat radiation. The envelope is another heatsink. Even when its glass, it will sink most of the heat radiation (infrared) and thus heat can also be dissipated by ordinary convection. Other than cathode needing to be in approximately certain temperature tubes were made to run COOL.

[g1]

There's even a Music Man amplifier patent where they address and exploit that particular characteristics as early "power scaling" setup.

This is interesting. Could you expand on this or provide any links to further info?

[J M Fahey]

Filament voltages were simply chosen to be compatible with common battery voltages.

^^^^^^^^^^^^^^^^^ that .
1.5V is a single carbon zinc cell.

2V is a single lead acid cell

3V is 2 x 1.5V cells

6.3V is standard "old car" 3 cell lead acid battery

12.6V is modern car battery, 6 lead acid cells.

Now higher ones, say 35V or 50V or odd values were rather current matched (say, 150 mA) so they could be wired in series for direct 110V mains . AC or DC .

In any case voltage by itself is not important, but cathode temperature 9which requires a certain power) .

And as I see it, the main current limiting factor is available electric field (voltage and distance dependent) to pull those electrons from the cathode.
Positive screen helps and is normally used that way.

Positive grid too, but not popular because is hard to drive.

If the limitation were only in cathode emission, not even positive grid would help.

[Merlinb]

I once designed an overdrive pedal that used 6AL5s as clipping diodes like a regular Tubescreamer. I had to run the heaters low (around 4.5V) to reduce the perveyance and get the clipping threshold where I needed it.

[teemuk]

This is interesting. Could you expand on this or provide any links to further info?

Certainly. The patent of concern is U.S. Patent # 3,978,421

"Electrical musical instrument amplifier having improved tremolo circuit, improved reverberation control, and power reduction circuit for distortion mode operation"

"The power amplifier can be operated in a distortion mode to achieve desirable harmonics without overloading the voice coils in an attached speaker system through use of a power supply which applies reduced voltages to the power amplifier whenever the gain of the power amplifier is adjusted to a high setting while at the same time regulating the voltage applied to the first and second preamplifiers to keep such voltage constant."

"When the power reduction switch 44 is positioned at the high terminal 156, the plate voltage to the triode tubes 132 and 134 is reduced together with the plate, grid and screen voltages to the pentode tubes 138 and 140. At the same time the cathode heater voltage provided by the secondary winding 160 is reduced. The reduction in the cathode heater voltage plays an important role in the proper operation of the various tubes within the power amplifier 36 in the distortion mode. The heater voltage is reduced to a point where cathode emission of the various tubes occurs in an on-off manner. This facilitates operation of the tubes in the desired squarewave fashion rather than in the much more linear fashion present during normal operation."


Google patents link:

https://patents.google.com/patent/US3978421A/en

[Old Tele man]

Will do - thanks, OTM.

Wait a second, are you referring to "space charge" - the bit of initial resistance caused by electron build up at/near the anode?

No, I was referring to Perveance (G or K or P):

[J M Fahey]

I once designed an overdrive pedal that used 6AL5s as clipping diodes like a regular Tubescreamer. I had to run the heaters low (around 4.5V) to reduce the perveyance and get the clipping threshold where I needed it.
[ATTACH=CONFIG]35950[/ATTACH]

Hi, THANKS for posting this very interesting experiment.
I have experimented with tube dual diodes myself but dropped because they are complicated to use (compared to a couple plain SS diodes) and impart no "tubeyness" to sound, clip like any diode.

But that's not the point: I would never have experimented with filament voltage as a method to change clipping threshold.
Very original

As`a side note: their performance as rectifiers with very low level signals is astonishing, and explains why apparently crude VTVM behaved better than expected from such simple circuits.