• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Filament temperature versus anode dissipation

tjerneld

Member
2020-11-06 12:24 pm
I think that the filament temperature of a valve is strongly dependent on the anode dissipation.
Does anybody know something about this ?
It is possible to determine the filament temperature by measurement of filament voltage and current.
Resistivity of tungsten: 300K 5.65, 1200K 30.98 1300K 34.08 1400K 37.19 1500K 40.36
Can anyone tell me the filament temperature for an 6DJ8, 12AU7, 6V6, EL84, EL 34, 300B and KT88 tube at zero anode dissipation and at maximum anode dissipation.
I think that tube lifetime and SPICE models are strongly dependent of filament temperature.
 
Last edited:

Elvee

Member
2006-09-08 2:04 pm
I am not an expert in this field, but I can say for sure that the temperature of the filament can be accurately determined based on its cold/hot resistance ratio.
Whether the filament's temperature depends on the anode dissipation is probably a more difficult subject: the hot anode has a radiative effect on the filament, but electrons emitted from the cathode take some of its heat away. How much? I leave the question to real specialists, I would have to struggle an hour or two with equations and constants to arrive at an estimation, which would probably be way out anyway
 

6A3sUMMER

Member
2016-06-07 6:50 am
You seem to be worried about tube life.
How accurate do you want the temperature measurement / calculation to be?

Are you worried about the total maximum tube dissipation?
An EL34 and KT88 dissipate heat in the filament, plate, And Screen.
Example, the EL34 pentode and KT88 Beam Power tubes both are rated for a maximum screen dissipation of 8 Watts.

Consider the tube materials and all the tube parts:

45, 2A3, 300B, and similar DHT tubes have coated filaments.

845, 212E, and 811, are also DHTs, but they have Thoriated Tungsten (Thorium infused in the Tungsten).

EL34, 6DJ8, EL84, KT88, etc. are Indirectly heated (Tungsten filament inside of a cathode).

All filaments radiate some light (or at least infrared light), and the different filament types, materials, and filament/cathode structures are not likely going to act exactly the same when it comes to heat 'from' and 'to' the filament.

Some Pentode and Beam Power tubes can not have both the maximum plate dissipation and maximum screen dissipation at the same time.
Example: KT88 max plate 42 Watts, max screen 8 Watts.
But the max plate plus screen is only 46 Watts . . . (not 42 + 8 = 50 Watts).

The nominal rating for a KT88 filament is 6.3Vrms @ 1.6A (10.08 Watts).
46 Watts plate and screen, and 10.08 Watts filament = 56.08 Watts total dissipation.

Some tubes filament voltage tolerance rating is + / - 10%.
The 300B filament voltage rating is 5V + / - 5%.
 
  • Like
Reactions: 1 user
There are some thermal investigations and discussions in the on-line 1962 RCA Electron Tube Design book, especially on a 6L6 - a long read even for the relevant parts of this thread, but it provides the best insight in to thermal design used in valves.

All the heat generated in a valve has to dissipate external to the valve - that is mostly by radiation from the internal structures to the glass and outside, with a minor part via conduction to the metal supports/leads. Radiative transfer works by temperature differential - so the heater/cathode assembly mainly radiates to the cooler anode due to geometry, but again minor transfer paths also exist. Radiative transfer is also a 4th power function to the temperature differential, so even if the anode increases temperature (within reasonable limits) the cathode temperature rise is significantly suppressed.

Apart from resistance change measurements (which may be tricky as such a measurement has to be taken very quickly after heater current is stopped), olden day non-invasive techniques include a Disappearing Filament Pyrometer (I have one of these but haven't practically used it) which needs line of sight to the filament.
 
The second law of thermodynamics says: Heat will always and exclusively be transferred from the hotter to the colder subject. As the cathode usually is the hottest part within a vacuum tube, plate temperature won't have an influence on it's temperature.

But you may investigate: Take a random tube, feed the heater with it's nominal voltage from a regulated supply (to exlude mains voltage variations) and measure the heater current. Then apply plate voltage and adjust plate current to the maximum dissipation rating. Measure heater current again.

Best regards!
 
There will always be an influence, albeit small due to the 4th power. If the heater/cathode power dissipation stays the same, then the only way for that heat to radiate away at the same heat flow rate is for the heater/cathode temp to increase if the anode temp increases by some other means - that is the basis of the (Ta^4 - Tb^4) equation. The complicating 2nd order effect is that a higher heater temperature may change the heater power dissipation level (ie. for a constant heater voltage due to a change in heater resistance).
 

Elvee

Member
2006-09-08 2:04 pm
No , the filament temperature doesn't change with anode dissipation ( current ) even for direct heated tubes . You should read some theory first ...
May I suggest you use your own advice:
There will always be an influence, albeit small due to the 4th power. If the heater/cathode power dissipation stays the same, then the only way for that heat to radiate away at the same heat flow rate is for the heater/cathode temp to increase if the anode temp increases by some other means - that is the basis of the (Ta^4 - Tb^4) equation. The complicating 2nd order effect is that a higher heater temperature may change the heater power dissipation level (ie. for a constant heater voltage due to a change in heater resistance).
A small, probably insignificant effect remains: the heat taken away by the electrons exiting the cathode.

I have used the filament resistance to measure the overall heating/cooling of a tube in different conditions:
https://www.diyaudio.com/community/...rging-a-e-pcl82-amplifier.351667/post-6134275

Luminaries like DF96 are no more with us unfortunately, and it would take me ages to figure out what quantity of heat is transported by electrons emitted from the cathode, but if someone can give an easy, first-order, simple answer, it would help a number of people, including myself
 
  • Like
Reactions: 1 users
Can anyone tell me the filament temperature for an 6DJ8, 12AU7, 6V6, EL84, EL 34, 300B and KT88 tube at zero anode dissipation and at maximum anode dissipation.
In bright emitters the normal operating temperature is around 2500K. For thoriated-tungsten filaments it is around 2000K (2A3, 300B, 845 etc). Indirectly heated cathodes operate at close to 1050K and the heater only needs to run a little hotter –about 1500K. I doubt anode dissipation has much effect on filament temp, given the huge differential.
 
1962 RCA p.261 discusses a thermal assessment on 6L6G where Tk=1060degK, and Tp=695degK for 18W plate dissipation, with the heater wire itself up around 1500-1550degK for this indirectly heated cathode example. Dull cherry-red plate is likely up to 400deg higher.

For constant heater power, if the plate was to somehow rise 100deg, then the heater wire would simplistically rise 11deg due to the 4th power relationship.
 

Elvee

Member
2006-09-08 2:04 pm
OK, intuitively I would have thought that the ambient temperature of the medium (~the plate in this case ) acted as an offset. If absolute temperatures rather than the delta has an effect, it means that the thermal resistivity of vacuum varies with temperature. Interesting and challenging
 

6A3sUMMER

Member
2016-06-07 6:50 am
Suppose the filament is powered, and there is no other dissipation in the tube.
Then . . . the plate temperature will rise.

As was mentioned earlier in this thread, some of the heat goes out through the lead wires and pins.

In many tubes, the plate "hides" almost all of the filament and/or filament plus the cathode.
All that heat has to go somewhere.
I suspect the plate warms up, and the warm plate heats the glass.

Example, a KT88 6.3V x 1.6A = 10.08 Watts dissipation.
The heat does not just "sit" there in a vacuum, more and more heat energy is constantly being produced.

I was testing a filament power supply for a 5894. 6.3V x 1.8A = 11.34 Watts.
The two plate pins out the top of the envelope got very hot to the touch.

Just my opinions.
 
Elvee, heat flow is modelled like an electrical equivalent circuit, although the heat flow by radiation between two nodes such as cathode to plate is a function of (Tk^4 -Tp^4). The plate then thermally conducts that incoming heat flow to its outer surface and radiates it to the glass inner surface as well as through the glass to external ambient surfaces by the same function, although transmission through the glass is frequency dependent due to the glass composition and thickness and is typically a minor contributor to heat flow.

Knowing the hotter 'sides' of the plate structure can be used to better orient the valve sockets for output stage valves in new builds.
 
Heat does not dissipate to zero in a vacuum.
It has to go somewhere, no matter how much it may spread out before it finds something to heat up.

In very old Physics theories, it was believed that there was a medium in a vacuum . . .
. . . It was called Ether.
The Michelson and Morley experiment disproved the theory that Ether existed.

Your Physics May Vary.

Just my opinions.
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
thermal resistivity of vacuum varies with temperature
I dunno about that. But the electrical resistivity of a heater varies with temperature, a lot. If you surround it with a hot object, resistance rises, power reduces, temperature does not go up as much as a simple "offset" says.

And FWIW, tube life is hardly affected by heater life. Most heater/cathode tubes die of something else long before heater failure. Not counting early series-string TV sets, I've only seen one failed heater: it had been thrown in a dumpster from at least 10 feet, and it still worked when cold (for a few seconds).

Yes, raw filaments are different' but post #1 mentions several tubes, all heater/cathode types.
 
Ok, here's what I'm wondering. As I said, I've been playing with the 327A, which as near as I can tell is very nearly a repackaged 100TH with a 110W filament instead of the 30W filament. That's an additional 80W of filament power! Most of that energy is going to radiate into the inner surface of the plate, since it covers it. Eimac never made a datasheet for the 327A that I have been able to find, but it obviously has the same plate as the 100TH, which is a "100W" plate. I've seen some online sources stating that the tube is good up to 100W and I have been running them at 90W or so. It seems like the ridiculous amount of filament power on this tube should be adding significant heat to the plate.

But....here's the thing. These tantalum plate 100T variants start to show a bit of color at fairly low plate dissipation, maybe 40 to 50 Watts or so. This subjectively seems to go for the 100TH or 327A. If I had to go by plate color, I'd say that the extra filament power doesn't seem to add any plate color at all, or in other words, color starts to show at about the same plate dissipation.

Granted, I'm not running the two tubes side-by-side at the same time since I only have one test setup but the massive filament of the 327A doesn't seem to be heating the plate as much as I'd expect to see with my admittedly uncalibrated eyes and brain.