Why weren't Vacuum Tubes designed for higher currents

[ZeroResistance]

I see that Vacuum tubes are typically rated for medium to high voltage (100's of Volts to few kV)  but currents are typically in mA upto a few Amps.
What factors limit their use only upto a few Amps?

TIA

[MagicSmoker]

Thyratrons, mercury arc rectifiers and quite a few other exotic tubes were (are) capable of 100s to 1000s of amps.

Hot cathode tubes - triodes, pentodes and the like - are limited in current mostly by the rate at which electrons can be emitted from the cathode (space charge limits, basically).

[Tom45]

I don't know the answer, but that isn't going to stop me from speculating.

In the linear region of operation a device dissipates E*I. For transistors, E is typically under 100 volts. For vacuum tubes E is more commonly in the 100s of volts. For a given power dissipation, a tube operating at a voltage 5 times that of a transistor, will max out at a current 1/5 of the transistor having a same power rating.

And the tube's vacuum makes it harder to get rid of the power it is dissipating.

There are water cooled tubes for high power operation. But these weren't seen in consumer electronics.

For example: http://www.tubecollectors.org/eimac/archives/8974.pdf

[ZeroResistance]

Thyratrons, mercury arc rectifiers and quite a few other exotic tubes were (are) capable of 100s to 1000s of amps.

Hot cathode tubes - triodes, pentodes and the like - are limited in current mostly by the rate at which electrons can be emitted from the cathode (space charge limits, basically).

Agreed! however I wanted to know regarding vacuum tubes.
I guess that Thyratrons and the other devices that you stated are "Gas Filled Tubes".

[ZeroResistance]

Hot cathode tubes - triodes, pentodes and the like - are limited in current mostly by the rate at which electrons can be emitted from the cathode (space charge limits, basically).

What are these space charge limits? I mean is there a known current (amps) limit beyond which vacuum tubes cannot be used?

[Gyro]

For a given cathode material / coating and temperature, there is a limit to the number of electrons that it can emit per unit area. Under non-biassed conditions, these electrons form a cloud (the space charge) around the cathode.

Once you bias the anode positive and the tube starts conducting, these electrons flow towards the anode. If you pull too much current, the space charge is exhausted and you start physically 'ripping' electrons from the Cathode surface, irreversibly damaging it (particularly if it is an oxide coated one for higher emission). The space charge also helps protect the cathode surface from positive ion bombardment.

Pure Tungsten or Thorated Tungsten directly heated (filament) cathodes are much more resistant to damage, they will effectively turn into constant current sources (again determined by surface area and temperature) when you try to pull more than the available density of electrons.

EDIT: The other physical limits are the maximum sustainable thermal dissipations of grids and anodes determined by voltage and current.

[IanB]

I understand there are "big" tubes used in broadcast equipment that can handle many kilowatts. I'm guessing they may also handle larger currents due to their physical size?

[Gyro]

Yes, large cathode area (often in a 'basket' form for maximum surface area) which need correspondingly high currents to heat them to emission temperature.

I believe large broadcast transmitter tubes were/are bright emitter Tungsten cathodes to withstand ion bombardment.

The other way they get so much power is to increase the anode voltage (a lot) of course.


P.S. Peter Millett's tubebooks.org site is an excellent resource for downloadable books on the subject... http://www.tubebooks.org/

[TimFox]

A normal vacuum tube has two limiting operating conditions:   emission-limited (“saturated”) and space-charge limited.  This is a simplified description of an ideal diode or triode:
If the anode-cathode voltage is high enough to attract all of the electrons emitted by the hot cathode, then (to first order) the diode current is independent of the voltage and is a strong function of the cathode temperature.  When the cathode current forms a space-charge cloud near the cathode, the negative charge cloud reduces the field at the cathode and controls the current.  In the limit of full space-charge control, the current behaves as the famous 3/2 power of the anode voltage, and does not depend on the cathode temperature, so long as it be hot enough to produce the current.
 When a control grid is added to make a triode, the 3/2 power law is applied to a linear combination of anode and grid voltages.  Here, I refer you to Wikipedia or any textbook.  This operating condition is very useful, since it doesn’t depend directly on the cathode temperature, so long as it is hot enough.
The vacuum tube structure is very easy to scale up for high power, but the space-charge equations and cathode current density limit this to relatively high voltage and low current, compared with a semiconductor device at similar power levels

[filssavi]

As other have already said there are plenty of limits, but I think the main reason high current tubes were not around is that simply there was no pressure from the market to develop them in a pre semiconductor world. Let’s keep in mind that the higher the current the higher the joule losses in the whole system...

[Gyro]

A search for 'broadcast transmitter valve' images brings up some very pretty pictures, eg.

[TimFox]

If you really need high current, there are low-mu triodes designed for series regulator service, starting with the 6AS7G.  The Soviet Union produced some big ones that have been used recently to make audio power amplifiers.  Look for the 6C33C (actually Cyrillic for 6S33), which can handle roughly 600 mA and 60 W dissipation.

[T3sl4co1l]

Cathode emission is relatively weak.  A typical sweep tube needs 15W of heat for about an ampere of emission, and that's supposed to be pulsed at that; a bigger transmitter tube requires hundreds or thousands of watts.

And yeah, between the space charge neutralization property of a plasma, and the much greater emission due to ion bombardment, the current density can be much higher.

We can go even further and consider a stationary matrix of ions; this further reduces the energy levels required (~eV to reach conduction band, versus ~10s eV to reach vacuum state), and greatly increases the current density.  Unfortunately, transit time is significantly impaired, because transport is via diffusion, at quite low velocities considering the energy levels and an electron's [free] rest mass).  On the upside, the breakdown voltage is also enhanced, so we can make very thin junctions, and this more than makes up for transport.

In other words, from a sufficiently high level, we can look at tubes and transistors as the same general thing: some exchange of charge carriers at the electrodes, with transportation inbetween, subject to applied fields.  For tubes, we use the ballistic transport of electrons, and space charges; for transistors, we use diffusive transport, and recombination.  Solving the equations gives a V^(3/2) characteristic for tubes, or a exp(V) characteristic for transistors.  (Depending on large/small signal regime.  Tubes, FETs and BJTs are all exponential at low current densities, driven by the tail of Maxwell-Boltzmann statistics.  At higher current densities, tubes go towards V^(3/2), FETs go towards V^2, and BJTs remain ~exponential.)

Of course if we put in a plasma, we get all kinds of weirdness, typically negative resistance and bistability.  This happens because emission is proportional to current flow times a multiplier (an ion impact begets multiple free electrons, on average).  The plasma is also not easily controlled by fields: due to its being conductive, it's an effective [electric] shield.  (The same is of course true of semiconductors, but we can harness very large surface areas in that case, making FETs practical.)

Tim

[ZeroResistance]

If you really need high current, there are low-mu triodes designed for series regulator service, starting with the 6AS7G.  The Soviet Union produced some big ones that have been used recently to make audio power amplifiers.  Look for the 6C33C (actually Cyrillic for 6S33), which can handle roughly 600 mA and 60 W dissipation.

The max I found on this page https://en.wikipedia.org/wiki/List_of_vacuum_tubes is 10kV, 20A switch.
I was wondering why they don't make tubes above 30amps+, looking at some of Gyro's posts above maybe I'm wrong and high current tubes (> 30Amps) do exist?
But I guess its difficult to find them>

[ejeffrey]

I was wondering about this and thought maybe you could make a vacuum tube that used electron multipliers dynodes like a PMT to increase the anode current beyond what a heated cathode could apply.  Yes: a "grid controlled electron multiplier" is a thing https://ieeexplore.ieee.org/document/1472778

[TimFox]

The series-regulator tubes were optimized for (relatively) high current at low plate voltage, but required high heater power.  For the 6C33, one could obtain 600 mA at only 50 V (conservative), with both cathodes and approximately 40 W of heater power.  The tube can make more current, but 600 mA is the continuous limit. With careful design, multiple devices can be connected in parallel for higher current, and to help heat your location during a Siberian winter.

[Gyro]

If you really need high current, there are low-mu triodes designed for series regulator service, starting with the 6AS7G.  The Soviet Union produced some big ones that have been used recently to make audio power amplifiers.  Look for the 6C33C (actually Cyrillic for 6S33), which can handle roughly 600 mA and 60 W dissipation.

Former Soviet military NOS tubes are a low cost goldmine for making audio equipment (for those of us who are into that sort of thing  ). I use GU50s for push-pull output tubes, very cheap and rugged (Pa 40W) with hot-swap handles on top - designed for battlefield RF transmitters (Hams like them too).

I believe the 6S33S was used in  low voltage series regulators in MIG fighters (discovered when a pilot defected irrc). If you look at the datasheets for some of the former Soviet tubes you'll find some very impressive specs for shock, vibration, altitude etc. Also EMP resistant of course.

EDIT: https://en.wikipedia.org/wiki/Mikoyan-Gurevich_MiG-25

The majority of the on-board avionics were based on vacuum-tube technology, not solid-state electronics. Although they represented aging technology, vacuum tubes were more tolerant of temperature extremes, thereby removing the need for environmental controls in the avionics bays. With the use of vacuum tubes, the MiG-25P's original Smerch-A (Tornado, NATO reporting name "Foxfire") radar had enormous power – about 600 kilowatts. As with most Soviet aircraft, the MiG-25 was designed to be as robust as possible. The use of vacuum tubes also made the aircraft's systems resistant to an electromagnetic pulse, for example after a nuclear blast

The max I found on this page https://en.wikipedia.org/wiki/List_of_vacuum_tubes is 10kV, 20A switch.
I was wondering why they don't make tubes above 30amps+, looking at some of Gyro's posts above maybe I'm wrong and high current tubes (> 30Amps) do exist?
But I guess its difficult to find them>

Once you get into pulsed operation or gas filled tubes, ratings get very high. For instance, multi-phase mercury vapour (liquid cathode) rectifiers were used to rectify the 600V traction current for railways! I'm sure there are some still in service somewhere. Also Thyratrons for RF heating and welding equipment.

[Gyro]

I was wondering about this and thought maybe you could make a vacuum tube that used electron multipliers dynodes like a PMT to increase the anode current beyond what a heated cathode could apply.  Yes: a "grid controlled electron multiplier" is a thing https://ieeexplore.ieee.org/document/1472778

That's impressive! I was aware of such structures in photomultiplier tubes of course, but not for high power, high current applications.

[hfleming]

TWT’s are high power. 180kW, 40kV, 16A are not uncommon.
the big boys are high-power magnetrons, like 5MW peak, 50kV anode voltage, and 250A.
Then there are deuterium-filled thyratrons, 35kV, 10000A.

Just browse through the datasheets of a manufacurer, like Teledyne, and you will see some really interesting and high-power vacuum tubes.

[ZeroResistance]

TWT’s are high power. 180kW, 40kV, 16A are not uncommon.
the big boys are high-power magnetrons, like 5MW peak, 50kV anode voltage, and 250A.
Then there are deuterium-filled thyratrons, 35kV, 10000A.

Just browse through the datasheets of a manufacurer, like Teledyne, and you will see some really interesting and high-power vacuum tubes.

Thanks for the reply, I appreciate it.
I understand that gas filled tubes  like Hydrogen filled Thyratrons are capable of higher currents due to the "gas ionization" causing a electron multiplier effect.
I would like to know how high a current can a vacuum tube handle. Would it be safe to say that a Vacuum tube > 50A was never built till date and the only tubes that can handle this current and above is a gas filled tube?

Magnetrons though would be vacuum tubes correct? and you iindicated a current of 250A ? Was that the main anode cathode current?

[ZeroResistance]

Once you bias the anode positive and the tube starts conducting, these electrons flow towards the anode. If you pull too much current, the space charge is exhausted and you start physically 'ripping' electrons from the Cathode surface, irreversibly damaging it (particularly if it is an oxide coated one for higher emission). The space charge also helps protect the cathode surface from positive ion bombardment.

Wouldn't the electrons be returned back to the Cathode after hitting the Anode and therefore replenishing it?
In that case why would the electrons in the Cathode get exhausted?

I didn't get the "positive ion bombardment" part too...

[hfleming]

Here are 2 datasheets, one is for a 5MW magnetron, the other is for a 66kW broadcast tube

[Gyro]

Once you bias the anode positive and the tube starts conducting, these electrons flow towards the anode. If you pull too much current, the space charge is exhausted and you start physically 'ripping' electrons from the Cathode surface, irreversibly damaging it (particularly if it is an oxide coated one for higher emission). The space charge also helps protect the cathode surface from positive ion bombardment.

Wouldn't the electrons be returned back to the Cathode after hitting the Anode and therefore replenishing it?
In that case why would the electrons in the Cathode get exhausted?

I didn't get the "positive ion bombardment" part too...

Don't forget there's an external circuit - electrons reaching the anode will pass into the external bias 'supply' ie. circuit. In fact some electrons can be knocked off the anode (secondary emission) which is responsible for the characteristic kink in a Tetrode's characteristics. This is resolved in the Pentode by adding a coarse negatively charged Suppressor grid between the anode and other electrodes to repel secondary emission electrons back to the anode.

The secondary emission kink is also suppressed in the Beam Tetrode (aka Kinkless Tetrode) which creates a 'virtual' screen grid in the voltage gradient to the anode (it's complicated). It's worth googling Tetrodes and Pentodes at this stage [Edit: and taking a browse through some of the books in Peter Millett's archive].

The Positive Ions are formed by collisions with trace gas atoms remaining in the tube. Being positively charged, they are strongly attracted to the cathode, and being relatively massive, they can severely damage the emissive coating on the cathode. This is why Tungsten is used for very high voltage tubes, even though the work function is lower, the surface is immune from high energy ion bombardment.

[chris_leyson]

Beam tetrodes were designed to get around the Mullard pentode patent and eliminated the tetrode kink hence valves like the KT66 or KT88, KT standing for kinkless tetrode, allegly.

One thing you don't see with vacuum tubes is the non-linear voltage controlled junction capacitance that you get with semiconductor devices. Less harmonic distortion to some extent.

[ArthurDent]

Here is a photo and close-up of a Tung-Sol CTL-6336A dual triode with an octal base. It was typically used as the series pass tube in higher current power supplies than the 6AS7 dual triode mentioned earlier and could handle about 4 times the current or about 450 ma per triode section..

The problem with vacuum is that it sucks as a conductor, both for electricity and for temperature transfer. Conduct may not be the technically correct explanation but that is the effect and easier to visualize. The 6336A didn’t use steel plates like the 6AS7 and other lower current (and less expensive) tubes, but had machined graphite plates that wouldn’t warp if they got hot enough to glow red. The heater was 6.3 volt @ 5 amps or over 30 watt and the plate dissipation was 30 watts for each of the triode sections. The electrons had to be boiled out of the cathode and were literally fired at the plate and their flow was controlled by the potential on the control grid. The spacing between elements of this triode was quite small to increase the efficiency by decreasing the distance the electrons had to travel through the vacuum.

To get higher power, some transmitting tubes didn’t have their plates inside a glass envelope but made the plate the outside of the tube so the heat could be more easily dissipated. This also meant the outside of the tube would have a high lethal voltage on it. The plates could be water cooled and here is the datasheet on a high power water cooled tube. The filament current of this tube is 640 amps and the anode current is 125 amp max. If you’re wasting tens of kilowatts just to boil electrons out of the filament, you can see why much higher efficiency solid state devices are so popular for transmitters today.

http://www.tubecollectors.org/eimac/archives/8974.pdf

[T3sl4co1l]

I was wondering about this and thought maybe you could make a vacuum tube that used electron multipliers dynodes like a PMT to increase the anode current beyond what a heated cathode could apply.  Yes: a "grid controlled electron multiplier" is a thing https://ieeexplore.ieee.org/document/1472778

Neat, I wonder what it looked like.

Secondary-emission tubes have a long pedigree:
http://www.r-type.org/exhib/aai0191.htm
https://www.radiomuseum.org/tubes/tube_1630.html

They were generally unpopular due to the higher operating voltage, current consumption (secondary emission counts twice!), and noise (in addition to pentode partition noise, there's multiplier noise too), as far as I know.

I remember something else about high power vacuum devices, but can't recall exactly what they were calling them.  Offhand I see keywords like "triggered vacuum switch", which appears to be a vacuum ignitron as it were (so, operation will be limited by sputtering of the electrodes providing the plasma).  These are often discussed in context of mains distribution handling (surge arresting, arc extinguishing, etc.; but probably not carrying load current, or turning off, for which a recloser is usually needed, AFAIK).  So, on the order of 50kV and about as many amperes peak.

Tim

[ZeroResistance]

What kind of heat losses is one looking at in vacuum tubes. If a tube is rated for 100kW, how much % of that would be wasted in heat?

[tszaboo]

If you place a conductor (wire) in vacum, the only way it can cool down is through conduction and emission. So your wires in the tube would get very warm if the current is high. Also, I dont have the number to back this, but high voltage low current is probably more efficient as high current low voltage. Then if you have a load witch requires high current, for example as 8 Ohm speaker, then you just run the current through a transformer. Which was the standard way of making these amplifiers.

[ZeroResistance]

I read that

"vacuum tube current is limited by space charge"

I understand that an electron cloud forms on the cathode but is space charge the effect of the electron cloud that is near the cathode repelling back the electrons emitted from the cathode?
How then do high current vacuum tubes mitigate this effect?

[bsdphk]

Some of the best articles about this can be found in Bell Systems Technical Journal (can be found on archive.org), in particular there are some good ones in the years after second world war and up to the transistor taking over.

The answer is that there are a large number of limiting factors, including the already mentioned cathode material issues, but many other more or less obscure effects.

[T3sl4co1l]

I read that

"vacuum tube current is limited by space charge"

I understand that an electron cloud forms on the cathode but is space charge the effect of the electron cloud that is near the cathode repelling back the electrons emitted from the cathode?
How then do high current vacuum tubes mitigate this effect?

Big fucking cathodes.

Have you read the numbers being thrown about in this thread?  Cathodes requiring 10s of kW have been mentioned.  That's a lot of white-hot tungsten!

What kind of heat losses is one looking at in vacuum tubes. If a tube is rated for 100kW, how much % of that would be wasted in heat?

Plate rating is plate rating.  Total input and output are whatever.  They can be higher (efficiency > 50% say for class C amps).

I've figured that a typical 30W sweep tube in class D can deliver 100-200W output, per tube.  Downside is, class D is very rough on the screen grid, which is typically rated for 3-5W for these types, but this service often dissipates >5W (depending on how carefully you've designed the inverter and its drive circuitry).

Still a far sight from the 90%+ of a MOSFET class D amp, but not bad considering.

They're still surprisingly competitive in terms of overall ratings -- it used to be a sweep tube (or the various other types: radar modulators, thyratrons, etc.) was about the only single device that could switch 5kV+ and nearly an ampere.  Now that 4.5kV 2.5A+ MOSFETs are cheap* and readily available, they really aren't good for much at all.

*I mean, they're like $20-30 each, but NOS sweep tubes are fairly in demand, easily running $60 each.  So, relatively speaking... not to mention all the savings from no screen or heater supply, and the lower drive voltages, of course.

Tim

[TimFox]

I read that

"vacuum tube current is limited by space charge"

I understand that an electron cloud forms on the cathode but is space charge the effect of the electron cloud that is near the cathode repelling back the electrons emitted from the cathode?
How then do high current vacuum tubes mitigate this effect?

Higher cathode current increases the space charge density until an equilibrium is reached with zero net field at the cathode and a finite current value less than the possible cathode emission at cathode temperature.  Increasing the applied field (from anode and grid voltages) increases that equilibrium cathode current in a nicely repeatable way.  The space charge controls the current, but does not kill it.

[ZeroResistance]

What is the significance of higher Anode Cathode Voltage, I believe that vacuum tubes were available in several volatges, starting at a few hundred volts to tens of kV's.
A high voltage would accelerate the electrons faster to the anode, wouldn't it.
So how does this high energy come into play? Because a higher energy electron slamming into the anode would manifest as heat once its velocity is put to a dead stop by the anode.
Is high volatge dependent on the anode cathode gap? I mean you need that amount of electric field to attract electrons towards you?

[Gyro]

What is the significance of higher Anode Cathode Voltage, I believe that vacuum tubes were available in several volatges, starting at a few hundred volts to tens of kV's.
A high voltage would accelerate the electrons faster to the anode, wouldn't it.
So how does this high energy come into play? Because a higher energy electron slamming into the anode would manifest as heat once its velocity is put to a dead stop by the anode.

It's like any component, P=V*A. For a given anode current, if you increase the voltage, the power dissipation (and available voltage swing) increases. Yes, the kinetic energy of the electrons hitting the anode is higher so the heat dissipation increases, just as you would expect.

Is high volatge dependent on the anode cathode gap? I mean you need that amount of electric field to attract electrons towards you?

No. For smaller gaps, the field gradient is higher and the electrons accelerate faster. For larger gaps, the field gradient is smaller but the electrons have a longer distance to accelerate in. The velocity with which they hit the anode is the same - you are accelerating them with the same amount of energy. Larger gaps simply make flashover between electrodes less likely.

[OldEE]

One of the problems with water cooling vacuum tubes is high voltage of the anode.  The first UHF TV transmitter that I babysat summers while in college had final amplifiers which ran at 6kV.  To minimize corrosion problems there was a 2ft diameter ceramic coil of purified water between the anode and the grounded plumbing.  How that thing was cast and fired always amazed me.  Once a day we check the conductivity of the water and it was usually 10-15 megOhms per cm.

Same job later had an RCA UHF TV transmitter which used 2 55kW klystrons for visual amplifiers.  These Varian klystrons were driven with 1 watt for 55kW out (47db gain) and consumed about 7A at 23kV continuous - not pulsed.  They were only 35% efficient so most of the power turned cooling water into steam which was condensed and recirculated.  The whole thing consumed just under a half a megaWatt.  The power company loved us.  Klystrons of this size were (are?) used on particle accelerators.

Larry