Tube Dissipation

[Structo]

There was a thread about the identity of some tubes over at TGP.

That raised a question in my own mind.

What makes one tube a 25 watt tube (EL34) and another tube a 30 watt
tube (6L6GC)?

Is it the physical plate area where a larger plate can dissipate more power?

Or take the 6550 tube that can sometimes output 40 watts.

I think the plate voltage can also affect it as per Ohm's law.

[Firestorm]

Interesting question. I'm sure the ratings are based on tube geometry, but it must be more than just the area of the plate; also its shape, the shape and placement of suppressor elements (grids or "beam forming plates"), plus the distances between cathode, screen, plate, etc.

The power dissipated in heating the plate is "wasted" from an audio perspective, so the engineers were clearly trying to maximize power transferred to the load while minimizing waste heat. With plates, screens and signal voltages set just right, a 6550 (in a push-pull pair) can transfer more than 75 watts to the load even while "wasting" 35W (or 42W) as heat.

[tubeswell]

What Firestorm said - its a mixture of everything to do with thermal dissipation capability, including surface:mass ratio of the metal (and the whole assembly and envelope), the type of metal, the spacing of the electrodes, etc.

Also, tube design evolved over several decades (and then was refined over a few more). So I suspect the dissipation rating isn't so much a result of perfect upfront design, as much as the testing of 'new types' of tubes under various conditions to see how long they would last, and then giving them a power-handling rating associated with so-many-thousand hours running time (I guess). You can run a tube hotter than the manufacturer's recommendations, but just don't expect it to last as long.

[Firestorm]

You can run a tube hotter than the manufacturer's recommendations, but just don't expect it to last as long.

The other interesting twist in this is that this is where plate-plate impedance really comes into play. Unless you have a crappy transformer (one where big voltage spikes can punch through the winding insulation), the tubes do not much care what the impedance is (sonic considerations aside) until you get close to the max ratings. Then, if the impedance is too low you can wind up with too much current flowing and cook the plates (among other things).

[Structo]

Thanks guys, for dumbing it down a bit for me.

I knew it was probably a complex question but what you say makes sense.

And of course the output transformer plays a huge role in how the power tubes behave.

[rdjones]

Also remember that the current flowing through the plate is DC not AC, so the measured (or calculated) power is the average of idle bias current plus the signal, and plus or minus any operating point shift that occurs at maximum signal.

rd