on careful rereading of the part about "swinging close to zero", i realize that there is a basic misunderstanding about AB2, the grid doesn't just swing to 0, it goes positive! the whole point is to go + on the grid! thats what makes it AB *2*. not AB1. and to really make it efficient, you want to drive the grid as positive as the grid (thermal) dissipation and cathode saturation characteristic will allow... that will be determined by plate voltage. the main reason most diy folks avoided 807's and 829's and all those wacky AB2 tetrodes (there are many of them for radio apps) was because you can't just drive them with a cap coupled driver stage. you need a transformer or direct coupled cathode follower. any short clipping transients tend to sink into the grids of those tubes (i wonder why?!), pulling the bias all over the place if fixed (and with the classically oversized grid resistors most people use... 220K in a fixed bias 6550 AB1 amp!, nuts! no wonder no one likes it) . most commercial tubes are not designed to go very positive on the grid... a few volts at most. but any tube made for class B operation is potentially useful for AB2 use as well. 807 is one such tube.
jc
jc
JC, a couple things got conflated. First, even in AB1, there's some grid current as one swings close to zero, not to mention what happens when the grid goes positive. Most of the comments about swinging to zero referred to the anode voltage, not grid. The only way to get the anode to swing close to zero for low perveance tubes is to pump the grid past zero.
Nanana: Thanks for the informative post!
By "swinging close to zero" I meant the plate voltage, not the grid voltage. Still, pushing the grid positive would allow higher current swing = more power (into the right load impedance)
Seems like transformer coupled AB2 is the way to go.
Since the goal here is maximum power, not maximum fidelity, this should be quite "doable".
By "swinging close to zero" I meant the plate voltage, not the grid voltage. Still, pushing the grid positive would allow higher current swing = more power (into the right load impedance)
Seems like transformer coupled AB2 is the way to go.
Since the goal here is maximum power, not maximum fidelity, this should be quite "doable".
+but the implication of that is just as preposterous!
it is plate resistance that determines the current through the tube... not plate voltage! it will be the current through the primary of the output transformer that will do the work in the secondary. in AB2, the plate resistance is lowered far below what can be done with AB1. that's nothing controversial at all... why would a pentode/tetrode have a disadvantage in this respect? actually, i see no ultimate difference in this respect to a triode or any other active device for that matter, other than the ultimate limit of plate resistance... that will depend upon just how positive a particular tube design will allow the driver stage to go before the grid evaporates.
a look at any grid curves in class B (above 0 volts) will show the saturation characteristics changing so don't tell me that its the "constant current" operation of tetrode/pentodes that makes this so... tetrodes and pentodes are NOT "constant current" above 0 volts bias.
jc
maybe there is some fundamental thing i've missed in all my years... it wouldn't be the first time! but i find it ridiculous to think that one could say "no point in AB2 and pentodes" seeing as hundreds of tetrodes and pentodes have been specifically designed to do just that better than any other job... like the 807 was.
fill me in...
jc
it is plate resistance that determines the current through the tube... not plate voltage! it will be the current through the primary of the output transformer that will do the work in the secondary. in AB2, the plate resistance is lowered far below what can be done with AB1. that's nothing controversial at all... why would a pentode/tetrode have a disadvantage in this respect? actually, i see no ultimate difference in this respect to a triode or any other active device for that matter, other than the ultimate limit of plate resistance... that will depend upon just how positive a particular tube design will allow the driver stage to go before the grid evaporates.
a look at any grid curves in class B (above 0 volts) will show the saturation characteristics changing so don't tell me that its the "constant current" operation of tetrode/pentodes that makes this so... tetrodes and pentodes are NOT "constant current" above 0 volts bias.
jc
maybe there is some fundamental thing i've missed in all my years... it wouldn't be the first time! but i find it ridiculous to think that one could say "no point in AB2 and pentodes" seeing as hundreds of tetrodes and pentodes have been specifically designed to do just that better than any other job... like the 807 was.
fill me in...
jc
Hi
All I´m saying is that I´ve read somewhere that pentodes does not benefit from AB2 operation.
I´m not saying that it is the ultimate truth, or even true at all.
Cheers!
All I´m saying is that I´ve read somewhere that pentodes does not benefit from AB2 operation.
I´m not saying that it is the ultimate truth, or even true at all.
Cheers!
nnn, what you're saying is equivalent to the plate voltage SWING being able to get closer to zero- if you can extend the load line end points, you can extend the voltage swing and hence the current through the OPT.
OK, now i understand the problem and the communication breakdown... my own misunderstanding. its just such an unusal way to describe the action of AB2... one thing needs to be clear: the swing in the terms you describe is not a straight line in AB2. its a bent curve with a distinct "knee", bent to accomodate the extended lessening of plate resistance in AB2 operation. and yes, the closer the plate swing gets to a short circuit, the more power you will develop in the load. if i get a minute, i will try to take a picture of this next week on a scope at work and i'll post it.
but that doesn't explain the original question... i still maintain that there is no inherent disadvantage of tetrode/pentode operation in AB2 operation! the idea is absurd. in fact, the only difference between triode and multigrid AB2 operation will be the amount of drive voltage necessary to develop the target current. using the 807 as an example, triode connection vs. tetrode connection, the amount of grid current necessary to get a particular plate resistance (as close to short circuit as possible) will be almost exactly the same! but the voltage swing into the grid will be greater in the triode connected 807 (less efficient i/o). the distortion will definitely be greater in the tetrode connected circuit (more efficient i/o). am i hallucinating or what?
more filling in please!
jc
but that doesn't explain the original question... i still maintain that there is no inherent disadvantage of tetrode/pentode operation in AB2 operation! the idea is absurd. in fact, the only difference between triode and multigrid AB2 operation will be the amount of drive voltage necessary to develop the target current. using the 807 as an example, triode connection vs. tetrode connection, the amount of grid current necessary to get a particular plate resistance (as close to short circuit as possible) will be almost exactly the same! but the voltage swing into the grid will be greater in the triode connected 807 (less efficient i/o). the distortion will definitely be greater in the tetrode connected circuit (more efficient i/o). am i hallucinating or what?
more filling in please!
jc
one additional note: the plate curves allow a straight line to be drawn across the characteristics... but this can be deceiving because if you were to make an actual voltage transfer characteristic graph - input volts vs. output volts - the line would be anything but straight! if it were a good choice of driver stage and output transformer, it would be an "S" curve. this is because the current vs. volts relationship varies with plate resistance, which is not linear when you pass through the diode line (0 volts bias) nor as you approach cut-off. you can see this in the plate curves clearly but the straight line can lull one into thinking that this is a linear transformation. it is not.
almost done now...
whew!
jc
almost done now...
whew!
jc
EC8010 said:
SY said:
Repeat after me:
CLASS 2 INCREASE PLATE CURRENT.
In pentodes it actually increases the saturation voltage (as Fuling's graph illustrates), but the gain in current offsets that for a very long while (you can keep increasing current, reducing impedance and gaining more power, until the grid melts!).
In most pentodes, screen voltage is indeed increased, making for a good easy-to-drive situation. Sometimes it doesn't have enough perveance even at max. screen voltage, transmitting pentodes for instance, so you may still need grid current to reach full ratings.
Or you can completely blow off all ratings and try getting 500W out of a 6L6 (as Gregggggeek has musingly suggested ... which reminds me, I was going to draw a loadline doing that ...).
Tim
I´ve spent an hour or two this evening to run some tests on that old OPT I mentioned and the results are quite encouraging!
I injected 70V rms (50Hz) across the full primary and got 12-0-12V out (secondarys loaded with 1,8k per side), so the turns ratio seems to be something like 3:1
Just about right to get a bit of driving power to the 807 grids!
The transformer has two quite interesting features:
1: A feedback winding. A bit of local feedback around the driver stage might help to get the driver output impedance down.
OTOH it might interfere horribly with the global feedback loop...
2: UL taps at 50% of the primary windings. Well, at least it would be quite fun...😀
I injected 70V rms (50Hz) across the full primary and got 12-0-12V out (secondarys loaded with 1,8k per side), so the turns ratio seems to be something like 3:1
Just about right to get a bit of driving power to the 807 grids!
The transformer has two quite interesting features:
1: A feedback winding. A bit of local feedback around the driver stage might help to get the driver output impedance down.
OTOH it might interfere horribly with the global feedback loop...
2: UL taps at 50% of the primary windings. Well, at least it would be quite fun...😀
Agreed, AB2 allows you to increase anode current in a pentode or tetrode, allowing a steeper load line. I have no problem with that. I haven't investigated as to whether it's especially useful, because...
At the transition from AB1 to AB2 the grid impedance drops like a stone. Imposing an undistorted grid voltage across that abruptly changing impedance isn't trivial. My opinion is that it isn't worth it, but I'd be interested to hear from anyone who has made it work for them.
At the transition from AB1 to AB2 the grid impedance drops like a stone. Imposing an undistorted grid voltage across that abruptly changing impedance isn't trivial. My opinion is that it isn't worth it, but I'd be interested to hear from anyone who has made it work for them.
Ah, someone at the back of the class has a question?
(Have to be quick, I'm off to bed soon to gird my loins for all the feasting.)
(Have to be quick, I'm off to bed soon to gird my loins for all the feasting.)
Sir, I was quite delighted running EL37s in AB2 mode with 5000 ohm anode-anode.... And, sir, if one can drive screen grids with a simple source follower, one should be able to run some poofy little control grids with the same driver... Yes, sir, I did say "source follower..."
OUCH!
Thank you sir, may I have another?
OUCH!
Thank you sir, may I have another?
OUCH!
Thank you sir, may I have another?
OUCH!
Thank you sir, may I have another?
Are you sure you are at the correct forum? Perhaps you should be at a "correction and discipline" site...
Hmmm, a source follower to drive the control grid? But surely the problem is that although the screen grid is far less poofy (I'm happy with technical words), it is at least (almost) constant impedance, whereas the control grid goes from being near infinite impedance to only a few thousand Ohms partway through the cycle. Were you thinking of a power FET?
Hmmm, a source follower to drive the control grid? But surely the problem is that although the screen grid is far less poofy (I'm happy with technical words), it is at least (almost) constant impedance, whereas the control grid goes from being near infinite impedance to only a few thousand Ohms partway through the cycle. Were you thinking of a power FET?
Hehe, if I were into correction and discipline I´d visit sites way different from diyaudio.com😀
I guess the sudden jump from "almost nothing" to "lots of current" can be smoothed out by loading the driver stage down with resistors.
BTW, forget them Mosfets. What about a 12B4A PP driver stage with 300V B+ (fixed bias) feeding those 807´s through a 3:1 stepdown IT?
I guess the sudden jump from "almost nothing" to "lots of current" can be smoothed out by loading the driver stage down with resistors.
BTW, forget them Mosfets. What about a 12B4A PP driver stage with 300V B+ (fixed bias) feeding those 807´s through a 3:1 stepdown IT?
- Status
- Not open for further replies.