Class B, zero bias - satisfactory circuit arrangement?

[G8VOQ]

Hi. Modulator tube line-up is 6SJ7, 6SN7, 6L6. The 6L6 provides about 6W of power to the finals, which are a pair of KT88's.  These are wired as triode connection, class B,  zero bias.  Audio drive is applied (via an interstage transformer) to g2, and g2 is connected to g1 via a 15K resistor.

Is this circuit arrangement perfectly satisfactory?

I feel like changing the circuit and wiring the KT88's as tetrodes, having some vague notion that is might be better for audio quality.  I was even thinking maybe the line-up should exclude the 6V6 and be replaced by a splitter based say on another 6SN7.

Your comments much appreciated. Thanks.

[w4bfs]

single ended 6l6 transformer drive to p.p. kt88 .... I am guessing the generated audio is clean up to a point and then gets dirty before desired power level is made ....is this what is going on?

I spent several months this year in the decision process trying to decide how to best drive tetrode modders ... if you already have the grid bias and screen voltage and can get output power in class ab1, that's the easiest but perhaps not the most efficient way ... there is great appeal to the enhanced drive method in terms of no screen/bias supply but you have to pay the piper with audio drive power

after looking at several methods using transformers and analyzing the enhanced p.p. drive method one fact became obvious ... each tube at peak positive drive voltage is drawing current and the input impedance drops to a few kilo Ohms .... this places a large constraint on the driver if low distortion audio is the goal ... a proper analysis is called for here but most hams don't want to do this so here is a quick and dirty approximation to tie a distortion level to a driver required impedance .... if the load is several kilo ohms the the driver source impedance should be no more than 10% of that or sevral 100's of Ohms .... this is why most class b modder drivers use low mu triodes like 2A3 etc because their plate resistance is in the 700 ohm range .... this is why a 6080 makes a good replacement for 2A3 in this service if you have extra drive voltage

a cathode follower to directly drive the screen/control grid is called for here ....a discussion with Reed W2CQH pointed out the need for a high transconductance tube such as 6bq5/el34 to serve as a cathode follower which would drop the output Z below 100 ohms and further reduce distortion

most folks want to use negative fb to reduce distortion and it can work but can also give phase shift headaches as you deal with transformer imperfections ....I hope this helps ....there is much more to this and I hope others will share ....gl ...73 ... John

[G8VOQ]

The reason I post is a concern as to audio quality. I've never used the AM mode so far, but intend to. In other words I have not run the modulator and established audio quality.

In the 1950's and probably earlier, it was fairly common to use triode connection for a tetrode and wire for class B, zero bias.

I am rebuilding and I'm wondering whether today, for audio quality reasons, class B, zero bias would be discarded in favour of running tetrodes, as tetrodes, in class AB1 or AB2 employing fixed or automatic bias.

If audio quality is perfectly satisfactory with class B, zero bias, then perhaps I should keep the arrangement. But, I just don't know.

I'm a real beginner BTW in all this. :c)

[KC2ZFA]

another beginner here, who's been looking into the following

vol. 16, #2, and
vol. 10, #4 in http://n4trb.com/AmateurRadio/RCA_Ham_Tips/rca_ham_tips.htm

(I'm building the vol. 10, #4 modulator). There's also a bunch of other
issues on that web page that discuss other circuits and some design-related
theory.

[G8VOQ]

BTW, my basic starting position is that I have all the parts for wiring the modulator deck as it was.

I'm rebuilding, and the question is, should I rebuild exactly as was, or change from triode connection, to tetrode. In the interests of better audio quality.

[WBear2GCR]

The KT88, 6L6, EL34 etc tubes are NOT "zero bias" tubes.

Wired as triodes, with the grid driven they must have negative bias on the grids to bring the current down to the proper quiescent current for class B or more likely AB. Cathode bias (self bias) can work, but is not recommended usually for these tubes.

Typical tubes that are in fact "zero bias" tubes are 811, 805 and 838. Even these may need some cathode or grid bias depending on the plate voltage used, etc...

                _-_-bear

PS. assuming you have the bias properly set, there will not be a noticeable difference in audio quality when used in a modulator. The "pentode/tetrode" arrangement with proper screen bias will yield more power, which is desireable. I prefer class B2 or AB2 in addition...

[KM1H]

Id just drive the KT-88's in AB1 with a 6SN7 and be done with it as nobody will be able to hear any distortion as long as the preamp/driver is clean. Worry about FB later but odds are that it wont be needed on the real world bands. If you need more audio use 6550's or the new KT-120's.

Dont get bogged down in the perfect audio scenario, thats strictly for the audiophools and most with them is myth anyway.

Carl

[WBear2GCR]

BFS,

I am not sure what impedance you are referring to? The input impedance of the grid of something like an 811? Since that tube will draw grid current, the driver only needs to be able to supply the requisite watt(s)... if you are driving the grid of the output tubes from a follower, almost anything will work fine. Like a 6SN7 for example...

Otoh, if ur driving the mod iron with a cathode, that is a different beast entirely, but somehow I don't think this is what ur referring to??

The 6080 and 6C33C are low gain Vreg tubes with super low plate Z, the plate Z on a 2A3 is substantially higher.

ZFA,

That is a nice classic modulator. Personally I wouldn't build one that way if I was going for a modern design. I'd get rid of the interstage transformer, unless I happened to have a high class one just laying about. I'd go with a form of cathode follower drive.

The only real issue with cathode follower drive is to make sure that there is enough swing available from the follower (therefore the tube that drives the follower) so with the gain of the output tube we get it to drive to clipping.

Imo the input stage of the old Ham Tips modulator is ok, but there are some better ways of doing the speech amp section that yield better phase splitting, but IF ur using an IT that really doesn't matter at all... but that RCA design will definitely work!

                   _-_-

[G8VOQ]

The KT88, 6L6, EL34 etc tubes are NOT "zero bias" tubes.

PS. assuming you have the bias properly set, there will not be a noticeable difference in audio quality when used in a modulator. The "pentode/tetrode" arrangement with proper screen bias will yield more power, which is desireable. I prefer class B2 or AB2 in addition...

Well, there is no bias to set. The circuit is very nearly the same as Fig. 9.35 on page 9.21 of the RSGB Communication handbook, 4th Edn.  Only difference is that cathodes were joined together then to ground via a meter marked "AUDIO WATTS".

P.S. I might be misunderstanding your quote here. :c)

[G8VOQ]

That is a nice classic modulator. Personally I wouldn't build one that way if I was going for a modern design. I'd get rid of the interstage transformer, unless I happened to have a high class one just laying about. I'd go with a form of cathode follower drive.

The actual interstage transformer is made by Chicago Transformer. It is stamped on the bottom: 7533-F.  Does anyone recognise this IT and can anyone please supply the specs?

I guess you are saying that class B, zero bias arrangement is not inherently a problem - audio quality wise. But, is the KT88 a suitable tube to be driven class B, zero bias? That is feeding g2 with a 15K resistor to g1?

Originally the modulator would have used 807's. For some reason, production went to KT88's.

[w4bfs]

BFS,

I am not sure what impedance you are referring to? The input impedance of the grid of something like an 811? Since that tube will draw grid current, the driver only needs to be able to supply the requisite watt(s)... if you are driving the grid of the output tubes from a follower, almost anything will work fine. Like a 6SN7 for example...

Otoh, if ur driving the mod iron with a cathode, that is a different beast entirely, but somehow I don't think this is what ur referring to??

The 6080 and 6C33C are low gain Vreg tubes with super low plate Z, the plate Z on a 2A3 is substantially higher.

ZFA,

That is a nice classic modulator. Personally I wouldn't build one that way if I was going for a modern design. I'd get rid of the interstage transformer, unless I happened to have a high class one just laying about. I'd go with a form of cathode follower drive.

The only real issue with cathode follower drive is to make sure that there is enough swing available from the follower (therefore the tube that drives the follower) so with the gain of the output tube we get it to drive to clipping.

Imo the input stage of the old Ham Tips modulator is ok, but there are some better ways of doing the speech amp section that yield better phase splitting, but IF ur using an IT that really doesn't matter at all... but that RCA design will definitely work!

                   _-_-

hi Bear ... the interesting thing about the enhanced drive method is that most of the tubes from 6v6 to 813 require about the same power level  :  3 to 5 watts  .... this tends to indicate roughly the the same drive requirements
and impedances ... any source impedance in a driver to a varying impedance load (such as this case) will distort the resultant waveform due to voltage drop

as to cathode followers not all tubes are as good as some .... from Radiotron hb the output impedance of a cathode follower is approximately the inverse of the transconductance (ex 6aq5 has 3300 micromhos which yields an output Z of 300 ohms - this does not count impedance of cathode path to ground ) the 6bq5 has a transconductance of 13000 micromhos and shows the preference of pentodes in this service

to other commentors ..... I do NOT agree that distortion reduction is a waste of time
 

[G8VOQ]

I guess you are saying that class B, zero bias arrangement is not inherently a problem - audio quality wise. But, is the KT88 a suitable tube to be driven class B, zero bias? That is feeding g2 with a 15K resistor to g1?

Just to answer my own quote: In the books that I have, the authors quite happily show class B, zero bias circuits using 807 or KT88.

There is no sense of a problem with audio quality, as long as there is sufficient drive power from a low impedance source.

But here we are several decades later and I wonder is there any real advantage to discard the class B, zero bias arrangement using the triode connection? Possibly not from an audio quality perspective if the drive arrangement is sufficient.  Which I presume would be the case in my professionally designed circuit. Not sure about any other perspective.

Would you use a pair of 807's today for the finals?

[WBear2GCR]

Sorry, afaik there is no way to NOT provide a bias on the grid of the 807, KT88, or other tubes of this type. Cathode bias puts a negative bias on the grid by "raising" the cathode above ground via the voltage drop across a resistor, still negative bias.

Please scan and post the schematic you are referring to? Afraid I don't have an RSGB book handy! (a pdf of one would be nice, though!)


BFS, sure in theory, but in practice there is no appreciable distortion added assuming the tube driving has sufficient watts available to drive the grid fully... that implies that the Z of the driver is low enough right there.

One of the problems with adding loop feedback to flatten response and/or "reduce distortion" comes when you clip the circuit. Feedback has a nasty habit of making clipping look very nasty, whereas ZFB (zero loop feedback) circuits tend to clip much more calmly and sweetly, with few if any nasties on the waveform... Ymmv, of course.

                _-_-bear

[KC2ZFA]

Imo the input stage of the old Ham Tips modulator is ok, but there are some better ways of doing the speech amp section that yield better phase splitting, but IF ur using an IT that really doesn't matter at all... but that RCA design will definitely work!

thanks for the info. Yes, I have an IT (UTC S-something which I now forget)...some time ago I bought
a modulator discussed in AMFONE from Bob w4rfm and have been experimenting with it...and when I
saw the ham tips article I decided to rebuild it according to that schematic (I am going for an antique
design with plenty of anode caps !).

[k4kyv]

The triode-connected configuration will give more output than AB1, and probably even more than AB2.  The difficulty with the zero-bias triode connection is the aforementioned driver problem.  A screen grid tube like the 6L6 makes a very poor class B driver unless negative feedback is used.  With feedback, the 6L6 makes almost as good a driver as a 2A3.  The feedback should go round the 6L6 only, and be accomplished with R-C networks, so that there is no transformer in the feedback loop other than possibly the driver transformer. That way you avoid the phase shift problems inherent to transformers.  The  6L6's driver transformer is not really in the feedback loop, since the secondary winding is effectively in series  with the source of feedback voltage, which is not "transformed" between the primary and secondary windings of the driver transformer. The driver secondary acts more as a passive component in series with the circuit.

The real weak link in the class-B modulator chain is the class-B driver transformer, the one that goes between the driver stage and the class-B modulator stage. There should be extremely tight  coupling (very low leakage inductance) between windings.  It is difficult to construct such a nearly-ideal transformer, so they tend to be expensive and hard to find, especially for single-ended circuits.  The cathode follower route may be a better way to go if a good driver transformer cannot be had.

I wouldn't classify a UTC S (Special) series transformer in that category.  You need something more like the LS (Linear Standard) series.

[KE6DF]

Here is a link to a 6as7 (6080) speech amp described in RCA ham tips.

http://n4trb.com/AmateurRadio/RCA_Ham_Tips/issues/rcahamtips0704.pdf

I does use an interstage transformer before the 6as7. Probably this could be eliminated by using a tube phase splitter.

The 6as7 is used in place of a pair of 6b4s (or 2a3s). Lots cheaper these days, but in the 1947 article they talk about the advantages of the 6as7 over the 6b4 -- and back then the cost difference wasn't an issue.

[G8VOQ]

The triode-connected configuration will give more output than AB1, and probably even more than AB2.  The difficulty with the zero-bias triode connection is the aforementioned driver problem.  A screen grid tube like the 6L6 makes a very poor class B driver unless negative feedback is used.

I wouldn't classify a UTC S (Special) series transformer in that category.  You need something more like the LS (Linear Standard) series.

In my situation, when the designer chose the circuit (G8VB) he has focused on efficiency and power output. So, he chose class B, zero bias, triode connection for the 807's/KT88's. I get that much.

Audio quality must have been "acceptable" at that time

OTOH, the 6L6 driver appears to have no feedback arrangement. The driver transformer in my case is a Chicago Transformer 7533-F. I don't know how good this transformer is.

So, it seems that if class B, zero bias is kept, there needs to be some additional feedback circuitry around the 6L6. That is my present understanding.

[k4kyv]

So, it seems that if class B, zero bias is kept, there needs to be some additional feedback circuitry around the 6L6. That is my present understanding.

Agreed.


Interesting article from RCA Ham Tips.  The RCA book says the maximum plate voltage is 250v.  Tung-Sol says 275.  They won't quite take the +300 plate voltage of a 2A3, but will generate more plate swing voltage.

I once ran my quad of 2A3s at +400 volts @ 30 MA per tube, instead of the recommended +300 volts @ 40 MA per tube.  I could get more grid swing voltage at the same plate dissipation, but the tubes didn't last long.  I later got a driver transformer  with a better turns ratio and was able to lower the voltage to +300 and run them at specs.  I have had that same quad now for over a decade, and last time I tested them they were still perfectly good.  You can't always get away with running a tube beyond its published ratings.

[WA1GFZ]

FQA11N90 1 ohm RDS on

[G8VOQ]

So, it seems that if class B, zero bias is kept, there needs to be some additional feedback circuitry around the 6L6. That is my present understanding.

Agreed.

The way this is going, the answer to the question, is class B, zero bias arrangement satisfactory, is yes. But, it would be an advantage, or an improvement could be forthcoming if the 6L6 was wired for some feedback.

I myself cannot work out the feedback stuff. Well, not at the present time.

I'll put up a copy on the WWW of the modulator schematic and post a link to it. And folks can say how the driver stage to the KT88's can be improved.

[WBear2GCR]

Ok, let me be the first to admit to extreme brain density...

Here is a data sheet for a 6L6G by GE:

http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6L6GC.pdf

Note the class A p-p suggested values?
In Class A, not B, there is suggested 250 B+ and ~-18v negative bias voltage.
Similarly they show a single tube in Class A triode connected at also 250vdc and a similar bias voltage...

So, how does one get less quiescent plate current with similar plate voltages and NO BIAS on the grid?? Seems to me that gets us more plate current, not less??

What am I missing here?

                        _-_-bear

[WD5JKO]

So, how does one get less quiescent plate current with similar plate voltages and NO BIAS on the grid?? Seems to me that gets us more plate current, not less??
What am I missing here?
_-_-bear

 Bear, you are spot on, but reread the original post. The modulators (KT88) were to be run class B driving each tube screen, and then tie G1 to screen through a resistor. So without any audio drive, G1 volts = 0, G2 volts = 0, and the static plate current is just barely above zero. This is similar to what Gonset does on the G76 with 6DQ6 modulators. Also recall that handbook modulator with 12AU7 cathode followers driving a pair of 7C5's? No bias supply, or screen supply needed; great for AM mobile where battery draw was a big concern.

Jim
WD5JKO

[WBear2GCR]

Oh - driving BOTH the screen and grid tied together?

Triode connection is usually the screen tied to the Plate via a resistor.

If you tie the screen and grid together it is more or less screen drive... but ok.
I doubt that is terribly linear, as screen driven tubes are very non-linear... but it would be interesting to read about it in more detail.

Edit: wonder what the gain looks like with this connection??

              _-_-bear

[KE6DF]

Oh - driving BOTH the screen and grid tied together?

If you tie the screen and grid together it is more or less screen drive... but ok.
I doubt that is terribly linear, as screen driven tubes are very non-linear... but it would be interesting to read about it in more detail.

http://www.pmillett.com/tubebooks/Books/orr_radio.pdf

I believe that that is commonly done. See page 662 of Orr's 1959 Radio Handbook (see the above link)

He shows 813 and 803 modulators that are driven with the screen and g1 tied together.

Also see the previous page where he shows the g1 resistor recommended with the lower power beam power tubes (e.g., 6l6, 807)

[Gito]

Hi

here's a schematic From Altec Amplifier ,using 811 triode as Finals,maybe it can be changed with 807 or KT 88 triode connected.
Used it with a modulation transformer with a feed back winding.

Gito