Minimum distance between two EL34

[Barnee345]

Hello Community

What should be the minimum distance between two EL34 amplifier tubes (similar to KT99)? The planned installation position is horizontal. In a GEC document it was once written of 4 inches, regardless of whether a vertical or horizontal mounting position is used. With a vertical installation position, too large an installation area can have negative effects, as natural convection is extremely impeded. The situation is different with a horizontal mounting position, when a free air flow is enabled.

[Gyro]

There are two factors at play. The first is glass envelope temperature, this is affected by airflow, how much energy it absorbs, cleanliness of the glass, and airflow. This relies on convection, it may be affected by mounting orientation (it may actually get better airflow with the chimney effect of vertical mounting vs the turbulence of a horizontal cylinder - I don't know). This is the minor factor anyway, as long as the pin seals don't overheat. The max pin seal temperature is normally around 150-200'C.

The major factor is thermal (IR) radiation. Becaues of their high operating temperature and the internal vacuum, the anodes dissipate the vast majority of their heat by IR radiation, directly heating bodies that they can 'see'. The manufactur spacing advice is based on this. If two tubes are spaced too close together, their anodes will mutually heat each other, raising both of their temperatures. This effect is independent of mounting orientation.

If you de-rate the anode dissipation of the tubes, you can mount them closer.

[TimFox]

Morgan Jones gives a rule of thumb: center-center spacing > 2 x envelope diameter.
That gives one diameter of free space between envelopes.
He also reminds the reader not to put the power tubes in the middle of the chassis, which can restrict convection.
Another issue is with beam tubes (e.g., 6L6, 6550, KT88):  be careful how they are oriented with respect to nearby transformers.
See Morgan Jones, Building Valve Amplifiers, 2nd Ed, Newnes 2014, pp 16-21.
The EL34 is a power pentode, so that is not so important as for a KT88.
Is the KT99 a beam tube or pentode?

[Gyro]

The KT designation indicates that it is a beam power tube (Kinkless Tetrode according to the original GEC designation). Athough not an original GEC design, it does appear to be of the same type.

Further indication... https://en.wikipedia.org/wiki/KT90


P.S. Both of the Morgan Jones books (Valve Amplifiers and Building Valve Amplifiers) are very well worth reading.

[Barnee345]

The major factor is thermal (IR) radiation....

You should not overestimate the value of infrared radiation. It only becomes important at higher temperatures. In order to be able to assess this better, you should get to grips with the Stefan-Boltzmann law. You need to know that the energy transport depends on the 4th power of the temperature difference. The colder the environment, the greater the temperature difference, and this can only be achieved by convective cooling.

[Barnee345]

Morgan Jones gives a rule of thumb: center-center spacing > 2 x envelope diameter.
That gives one diameter of free space between envelopes.

Thank you very much, I think I'm on the right way. I have already bought Morgan Jones and read the relevant chapter. But I need to study it more carefully.

[Kleinstein]

The major factor is thermal (IR) radiation....

You should not overestimate the value of infrared radiation. It only becomes important at higher temperatures. In order to be able to assess this better, you should get to grips with the Stefan-Boltzmann law. You need to know that the energy transport depends on the 4th power of the temperature difference. The colder the environment, the greater the temperature difference, and this can only be achieved by convective cooling.

It is not that 4 th power of the temperature difference that matter, but the 4th power of the absolute temperature (e.g. Kelvin scale). For the heat transfer it is the difference in 4 th power of the temperature and this can already start of small temperature differences to work. Another complication is that the glass is only partially tranparent and the metal parts can have low emission factor in some ranges.
IR radiation can still be relevant, as the anodes can rund relatively hot (e.g. 150 C and more) so something like 4 x the IR radiation as at room temperature. 

[Roehrenonkel]

Hi Barnee345,

Thank you very much, I think I'm on the right way. I have already bought Morgan Jones and read the relevant chapter. But I need to study it more carefully.

well, get John Linsley Hood and Allen Wright (SuperCable CookBook and Preamp Cookbook) and maybe Malcolm J. Hawksford (Riaa, Crossovers)
and you're covered. :-)

Best regards

[Gyro]

A quick look a Morgan Jones says around 50% convection and 50% radiation from power tubes (so yes, I was a bit off on the relative proportions). That's presumably for normal soft glass, my output tubes (GU50) are hard glass - I've no idea how much thermal IR transparency differs.

...
well, get John Linsley Hood and Allen Wright (SuperCable CookBook and Preamp Cookbook) and maybe Malcolm J. Hawksford (Riaa, Crossovers)
and you're covered. :-)

Allen was a really nice, helpful guy, it's sad that he was taken from us so early.

[Barnee345]

It is not that 4 th power of the temperature difference that matter, but the 4th power of the absolute temperature (e.g. Kelvin scale).

Hello,

I did not want to discuss the calculation path of heat transfer now    I have been dealing with this topic for a long time in my professional life. I am also still very familiar with the influence of the emissivity of sheet metal at Epsilon of 0.11. I only brought up the subject because elsewhere (also in Morgan Jones) the ratio of heat transfer from radiation and convection was very jovially put at 50:50. After all, within the tube, one assumes a predominant transfer by radiation in a vacuum, doesn't one? That should not really be in doubt. That most of it comes from the anode should also be pretty clear. I am not a specialist in tubes, but I could imagine that the manufacturers deliberately do not use glass that transmits IR radiation preferentially. Or maybe they do? If I am correct in my view, then surely it is absolutely right to take the necessary measures to increase convection. When studying old designs (e.g. BRAUN CSV60), I have seen that the manufacturers sometimes installed the tubes very tightly.

[Gyro]

Respected manufacturers have certainly seriously broken the rules in the past, the original Quad II (KT66) being a particularly blatant example...  https://duckduckgo.com/?q=quad+II&atb=v328-1&iax=images&ia=images

[floobydust]

This reminds me of the old mainstream construction technique of placing the aluminum can electrolytic filter cap right next to the 5U4. This caused countless capacitor failures due to the heat. Their cans are brushed/polished aluminum, you would think the radiation would reflect a lot of the heat away, but that is wrong for some reason.
With power amps I leave a good amount of room to allow fat tubes i.e. KT88 (OD 51mm/2.06") to fit in place of other slim envelope tubes like EL34 (OD 33.3mm).
There is also capacitive coupling between plates but not an issue with push-pull operation.