Merlin Blencowe's wepage on designing a Class AB power amp states that when the tube enters class B operation and sees only 1/4 of the primary winding, 'This load line may cross the max dissipation curve over a portion of its length- this is allowable.'
... I went and plotted a 50 watt Fender Hot Rod Deville ... It is REALLY through the max dissipation zone. ...
Look at the long distance moving from the Red Dot ("idle")
to the left where the loadline crosses the "Vg=0v" curve (at about 96v plate and 393mA of plate current). Think of this long distance as "a long
time."
Now look at the very short distance moving from the Red Dot
to the right where the loadline runs down to 0mA of plate current (at about 546v). Think of this short distance as "a very short time" until the tube shuts off.
----------------------------------
If the tube stays on all the time ("Class A") then Best Practice is to keep the loadline always stay below the curve of max dissipation.
One the tube's grid-input causes its plate to move along the loadline to the right and cut off (at 0mA plate current),
dissipation has fallen to zero. The plate's temperature is falling.
If the tube gets to cutoff very fast (short distance, "short time") when swinging in that direction, then
the plate cools off longer. That allows us to push it to higher peaks and further over the max-dissipation-curve when the plate is moving along the loadline to the left.
------------------------------------
The time (here, distance along the loadline) the tube will take to move from Idle to the furthest excursion to the left will be similar to the time the tube will take to move from Idle to the furthest excursion to the right.
Plate voltage dropped from 485v to ~96v, a reduction of 389v. Plate voltage will also rise from 485v to about (485v + 389v) = 874v. But all the time the plate is swinging from 546v up to 874v (328v worth), the plate current and hence the plate dissipation is zero.
------------------------------------------------
"Class B" is when we cut off the tube
exactly half the time. Theoretically, the tube is biased to the grid-voltage that "just shuts off the plate current."
What this allows is using a higher supply voltage, a smaller transformer primary impedance (which then results in), a higher peak plate current, which then takes the loadline further above the max-dissipation-curve. The additional plate-heat caused by straying above the max-dissipation curve is offset by the long time (exactly half the time) that plate current is cut off, plate dissipation is momentarily zero, and the plate's temperature is falling.
-----------------------------------------
The plate swing from its Idle voltage downward about 389v to peak plate current. If it also swung upward 389v (to 485v + 389v = 874v) and was idled at/near cutoff, we would call it "Class B."
Since the tube idles closer to cutoff (in terms of both "current" and "time") than to halfway between the extremes of plate current & plate voltage swing, we might call this "deep Class AB." As in, "deep into Class AB, and closer to Class B."
-----------------------------------------
What matters is the "Average plate dissiaption" over the entire signal cycle, and that can be painful to calculate by hand (and our handy website here doesn't give you most of the "in-between grid-voltages" you would need to calculate it properly).
The full method for calculating when using a loadline could be taken from RDH4 pages 563-565, 579-580 (especially the latter). You would be looking at the product of instantaneous plate current & plate voltage at each of the % of grid-voltage noted, but you would be taking the current of a single tube (not tubes from each side of the OT) so you will need to figure values over 180º of the signal, not the 90º used when only trying to figure out power output.
Topic: Class AB 6L6 design help (Read 3185 times)