Firestorm wrote:Lower grid circuit resistance can have a subtle impact on sound and output volume in a push-pull amp because you have essentially reduced the resistance isolating two out-of-phase signals. But to get a big change (like 42 watts going to 80 watts), it's likely that the amp shifted into AB2 operation and began drawing grid current (if the source can provide the current, the amp will be louder).
So you're saying that Bob-I's amp likely got audibly louder because the supply was able to provide more current than mine does?
Firestorm wrote:But this also means that ions are building up on the grid which destabilizes the bias. Gassy tubes can destroy themselves very quickly this way, but even tubes with a solid vacuum can runaway.
Some of the published grid circuit resistances are overly conservative; I think the 6V6 is limited to 100K, but Fender ran them at 220K with no ill effects. The internal geometry of 6550s seems to be less forgiving. I have killed several of them with too much grid resistance.
Good to know...
Firestorm wrote:The other spec that gets ignored too often is the requirement that each tube have it's own bias supply (or it's own cathode resistor). If you do that, you can probably exceed the maximum grid circuit resistance with less danger.
Am I correct in assuming that this is because you can properly bias each tube, or is there more to the story? In other words, if I have a pair of tubes that biases within a few milliamps of one another, I would ordinarily consider that "good enough" even with a single bias supply. But are 6550s and KT88s and their ilk more inclined to go off reservation and become unstable or something?
The only thing I can think of to explain the significant jump in power that Bob-I got is a transition to Class AB2 and a PI that was able to deliver the additional power the grids required to run in that class. What happens in Class AB2 is the grid is driven more positive than the cathode. Since the grid is in between the cathode and the screen (or the plate in a triode), some of the electrons flowing from the cathode are attracted to the grid instead of to the screen and the plate. The electrons striking the grid need a return path to the cathode (ground) so current flows in the grid circuit. (That's why the grid resistance is limited -- with current flowing in the grid circuit, too large a resistance will develop a significant voltage across it and alter the tube's bias).
The data for 6550s and all the KTs include the "requirement" of separate bias for each tube, but I've never found a good explanation for why this is so. It may be because if one tube drifts significantly it can throw the other one through a shared grid circuit or cathode resistor. Or it may be related to the grid current issue, since these tubes also all have low maximums for grid circuit resistance. There might be conditions where the most attractive return path is the other tube rather than the intended DC ground node.
I've killed tubes this way too -- but only in really loud bass amps.
I think that we could see the differance if zero and max signal current figures
were provided for the plate and screen. That would instantly tells us if a
condition was present that exceeds the screens dissipation in watts, and would
provide enough data to find the conduction angles for each value of grid resistor
to judge "class"
It would mean the installation of a 1 ohm res. in the screen circuit and
require a max power test, but really is just a few circuit measurements
after that.
Yeah, I agree ! I,ve been trying to get a handle on the function and effects of the suppressor and screen grids in a pentode. This thread has been very helpful ! Art
