Pentode question.

[Ptron]

This is from the Valvewizard http://www.valvewizard.co.uk/ article on preamp pentodes but it appears aimed at explaining why pentodes are usually used in power amps. "Pentodes have very high internal impedances and can be considered to be constant current source devices. This makes them excellent for use as power valves."

I thought a power amp was all (well mostly) about matching impedances and maximum transfer of power.
a) I don't see how a current source gets you toward these goals.
b)I don't see how a pentode is really a high impedances current source any more after you've (somewhat) matched to the speaker via the OT.
Can anyone explain what he might mean by that statement?

[Old Tele man]

..."constant CURRENT" is visualized by the almost horizontal plate current graphs.

...in other words, plate CURRENT doesn't change much when the plate VOLTAGE is increased or decreased.

...and, "Yes," the low OT impedance (from reflected speaker load) literally "swamps" the typically high plate impedance. Remember, most power pentodes (beam and "true" 5-element pentodes) have rp-values in 20K-80K ranges, while power triodes are MUCH lower, typically less than 5K ohms, but (through use of OT) work with and drive output loads of 3K-10K.

...for example: (A) 6L6 has rp = 22.5K (at Vp = 250V) and is optimum at 3.8K, but "works" with anything between about 3K and 6K. (B) 6V6 has rp = 50K (at Vp = 250V) and is optimum at 6K, but "works" with anything between 6K and 10K, etc.

[Steve Conner]

Well, I think Valvewizard is true in his statements but wrong in his reasoning. What makes pentodes excellent for power amps is not their impedance, but their ability to generate large output voltage and current swings with only a little grid drive.

The high output impedance is actually a disadvantage, in audiophile circles at least, because it hardly damps the speaker. You actually have to use negative feedback to bring the amp's output impedance down for good speaker damping, whereas triodes can damp it fine without NFB.

Also, the "output impedance" measured from the slope of a pentode's plate characteristic is a small-signal value. It's not the value to be matched for maximum power transfer.

A rough estimate of this large-signal impedance can be got from the ratio of V to I at the "knee" point in the top-left corner of the tube characteristic, because that's where you want to aim your load line for maximum output. If you aim too low (on the chart, ie too high an impedance) screen current skyrockets and you melt the screen grid. Aim too high and you melt the plate: power gets burnt in the tube instead of transferring to the load.

(Multiply by 4 to get plate-plate impedance in a Class-B push-pull amp.)