Fixed AND cathode bias?

[martin manning]

Greetings. I have a Marshall 4001 Studio 15, which has a number of interesting features in its circuit, but one that puzzles me is the 270 ohm cathode resistors and 22 uf bypass caps on the output tubes. The schematic is linked below (but the reverse polarity of the caps as shown is incorrect). Does anyone know what is going on here?

Martin Manning

http://www.drtube.com/schematics/marshall/4001m88.gif

[Enzo]

Bias is the voltage relation between the cathode and the grid. The tube has no idea how the bias voltage gets there, all it knows is that the grid is so much negative with respect to the cathode. usually we pick one way or the other, but nothing says you have to do it that way. You can have both methods contributing.

[martin manning]

So this is a tone thing, I guess to take some of the edge off by adding a little compression? I have usually seen one or the other bias method, or switchable from one to the other, but this is the first case where I have seen both used at once.

[pdf64]

Vox used the 2 methods with the ac50
http://www.voxamps.co.uk/downloads/circuits/ac50_1.jpg
Individually adjustable bias for each el34 as well, and a nice self balancing paraphase inverter rather than LTP - Peter

[martin manning]

Vox used the 2 methods with the ac50
http://www.voxamps.co.uk/downloads/circuits/ac50_1.jpg
Individually adjustable bias for each el34 as well, and a nice self balancing paraphase inverter rather than LTP - Peter

Interesting... thanks for pointing that one out. No bypass cap there, and a relatively small value resistor, so a different flavor.

MPM

[stokes]

Seen it on a Fender once as well.Had an early SF Bassman on the bench with,IIRC, a couple of 75ohms on the cathodes.Although I dont see it on any schem,it appeared to be stock.

[hasserl]

Yeah, that was one of the things people complained about with some of the SF Fenders. But I recall somebody on the old Ampage bbs experimenting with adding small cathode resistors to the power tubes of his Bassman head guinea pig amp to develope part of the bias voltage and adjusting the fixed bias supply to bring the bias voltage to what he needed to get the tubes idling properly. He liked the effect of the added compression. I think the cathode resistor was ~ 20 ohms, IIRC.

[martin manning]

Maybe some thread drift here, but I'll tack this query on:

I just swapped the output tubes in this little amp, installing a pair of JJ 6V6S's. While doing this I had to max out the bias trimmers and still got only 21 mA out of the plates, as calculated using the OT primary resistance and voltage drop.

Since there is a cathode resistor (here's the relevance to the original topic, I guess), there is 5.8V at the cathodes. There is -26V on the grids, and 381V on the plates. Relative to cathode voltage, the bias voltage is really -31.8, and Va is really 375. If I use these numbers to get on the published (JJ's web site) plate characteristic curves, things don't line up too well. Using Va and Vg, I should be seeing about 35 or 36 mA (not 21). Using Va and Ia, I should be at -35 or -36V Vg (not -31.8).

From this I have to conclude that this is one cold set of tubes. They do indeed match, but they are running about 15 mA low.

If I calculate Pa as Ia * Va, I get 0.021 * 375 = 7.9W. If Ia max is 14, then I am only running 56%. The amp doesn't sound bad, but I can't look at the output waveform right now ('scope, but no signal generator).

Questions:

Is my logic and math right?

The matching numbers on the boxes is "20," and they were stuck together with a "Mojo Precision Matched" sticker. Does anyone know what unit this is?

If I look at the published curves, the 14W Pa max line passes through 40 mA and 300V. Wouldn't that be 12W, not 14W? what's up with that?

MPM

[stokes]

12 watts would be about 90% max dissipation (14 watts) which would put you in class A territory.

[martin manning]

12 watts would be about 90% max dissipation (14 watts) which would put you in class A territory.

I guess I wasn't clear. There is a 14W line shown on the published curves, which passes through 40 mA and 300V. That seems to be 12W, not 14, no?

http://www.jj-electronic.com/pdf/6V6.pdf

MPM

[Old Tele man]

12 watts would be about 90% max dissipation (14 watts) which would put you in class A territory....idle dissipation wattage has no relationship to Class of operation (ie: Class-A vs. Class-AB, etc.), it's the relationship of plate idle current to peak plate current that determines whether it's Class-A (360° of conduction) or Class-AB (less-than 360° of conduction but greater-than 180° of conduction):

CA° = 2*ACOS[ - Ip.q / Ip.pk ]

where:
CA° = Plate conduction angle, degrees-of-conduction
Ip.q = Plate idle (quiescent) current (* both tubes value)
Ip.pk = Plate peak signal current (* single tube value)

...when Ip.q and Ip.pk are EQUAL, it's 360° conduction and thus Class-A. But, whenever Ip.pk is GREATER than Ip.q, it's less-than 360° conduction and (usually) Class-AB.



note (*) view 'push-pull' as a single composite tube representing the two separate opposedly-operating tubes.

[stokes]

Okay,ease up OTM,if you see,I used the term "class A territory" so as to avoid the whole "what constitutes class A" can of worms.Very thorough mathematical explanation,I'll give you that,but it gives me a headache just reading it.For me,taking a SE amp like a Champ,lets say,targeting 90% dissip. at idle then tweaking it for what sounds best to me,and after all,"what sounds best" is the bottom line in guitar amps,is close enough to class A.
Martin,yes 40ma at 300v's would be 12 watts.I dont know what you mean by "published curves",but I suspect the "chart" you are looking at is stating the tube is good for 14watts max,but the curve it is showing is a "typical operation" example?After all you wouldnt want a tube that has a 14watt max rating idling at 14 watts,at least the manufacturer wouldnt.We all know that in guitar amps we push things to the edge and over,so some of us may want to idle the tube at the max.And all the math aside,I'll stand by my statement that a 14 watt 6V6 idling at 12 watts is "class A territory",maybe not "pure class A" but close.

[Old Tele man]

Okay, ease up OTM, if you see, I used the term "class A territory" so as to avoid the whole "what constitutes class A" can of worms....sorry, I just translated "class A territory" as meaning "...never ceasing conduction..." and not as meaning "...very close to..." which is what I believe you were actually conveying.

...the reason I "piped-in" is that the RCA 2 x 6L6GC 55W Class-AB circuit example has an idle dissipation percentage of 87% (58mA per tube at 450Vp) even though it's truly push-pull Class-AB, and not even close to Class-A, conduction-angle wise.

...granted, 87% isn't 90%, but it's close enough to both make and illustrate the point that "high" idle wattages approaching high-90% were/are feasible with Class-AB operation; and, thus, "high idle wattage" is not a valid differentiator of whether a circuit is Class-A or Class-AB operation.

[stokes]

The example you give for the 6L6GC is for class AB1,most guitar amps are AB2,no?Which,without going into the numbers,you have the manual, is about 70%max plate diss. as a general rule.And we could go back and forth for days.Look,I'll admit your explanation is probably a spot on,lab definition,for class A.But in guitar amps we tend to break a lot of rules and push tubes beyond their limits,but going by the manuals,and using guitar amps as the subject,AB2 would be described as about 70% max diss and class A is about 90%.I am not saying your definition is wrong by any stretch,and that for a technical definition there arent other parameters to be considered,but for guitar amps,the 70% class AB2 and the 90% class A as a general rule of thumb are close enough,in my experience anyway.

[Old Tele man]

...two points: (1) Leo Fenders original amps were biased and operated from the factory as basically Class-AB1 for "clean" sounds, hence the "cold" idle bias and typically "deep" negative fixed-bias voltages (ie: -51/-52Vdc, etc.)...but, (2) most people (especially today) don't leave their amps biased cold as they come from the factory, nor do the play them cleanly.

...when people cranked up Leo's factory bias, which was about 55% for 6L6s and 60% for 6V6s, and pushed ever greater guitar signals into the jacks, the result was indeed that most "original" -AB1 amps got pushed/played in "hot" -AB2 conditions...ie: the control grids went positive and began conducting, the very definition between suffix "-1" (never goes postive) and suffix "-2" (goes postive at some time).

...oh, what about that "deep" negative fixed-bias voltage (ie: -51/-52Vdc, etc.)? Well, it was there for a reason, because when the "numbers" are backsolved, a 40W output would only require a 35V.rms input drive signal amplitude, representing a 50V.pk (sinusoidal) value, which was about 1-2Vdc below the bias voltage...for Class-AB1 operation.

...as you said: it's all there in the tube manuals.

[martin manning]

Martin,yes 40ma at 300v's would be 12 watts.I dont know what you mean by "published curves",but I suspect the "chart" you are looking at is stating the tube is good for 14watts max,but the curve it is showing is a "typical operation" example?

Stokes, the chart I referred to is from JJ's web site, The link is in post #10, but it's here:
http://www.jj-electronic.com/pdf/6V6.pdf
This just looks like an error to me; a 12W line labeled 14, but then at the low Va-end it seems well below 12, so I'm still a bit puzzled.

I'm kinda new to this game, still collecting gear and books.

Bruce/Mission Amps' post over in the current "Bias is Bias" thread
http://music-electronics-forum.com/showthread.php?t=6259
gives some great insight as to how much variation one might expect to see in the same part number, even from the same manufacturer. Also interesting that 50-60% of max dissipation is acceptable in many cases.

MPM

[Old Tele man]

This just looks like an error to me; a 12W line labeled 14, but then at the low Va-end it seems well below 12, so I'm still a bit puzzled....you've got sharp eyes, I never noticed that! you're correct, it looks to be mis-labelled, for example, the horizontal 60mA line crosses the vertical 200V line right at the 14W Pa label, but would actually only be 12W Pa.

[stokes]

Although it doesnt state it,I am still going with the notion that the tube is a 14 watt max tube and the chart is showing a "typical opertion" condition.Or it is a misprint as you guys think.
OTM,now you are really confusing me.In your first reference to the 6L6GC example you seemed to be inferring that the AB1 numbers you were qouting (87% max diss) were hot.Now you are saying AB1 was considerd "cold".I'm probably missing something there,and I really dont want to get into arguing math,I am sure you've done your homework on that front,so I'll concede,90% idle diss has nothing to do with class A,it works for me,even if the numbers disprove it,I'm too old to go back to math class,cant teach an old dog......etc,etc..you win.

[Old Tele man]

OTM, now you are really confusing me. In your first reference to the 6L6GC example you seemed to be inferring that the AB1 numbers you were quoting (87% max diss) were hot. Now you are saying AB1 was considerd "cold". I'm probably missing something there, and I really dont want to get into arguing math, I am sure you've done your homework on that front, so I'll concede, 90% idle diss has nothing to do with class A, it works for me, even if the numbers disprove it, I'm too old to go back to math class, cant teach an old dog......etc,etc..you win....I apologize for not being more concise; both postings refer to 2 x 6L6GC Class-AB1 circuits, but the RCA-example has an 87% idle wattage (58mA at 450Vp), while in the second posting I describe Leo's later brown/black tolex-era mid-range 40-45W amps (Pro, Concert, Super, Bandmaster, Vibroverb, etc.), which, by current standards and tastes, were idled rather cold at about 55% idle wattage (about 35mA at 450Vp).

FWIW -- the real differentiators between the RCA and Leo amps are (a) the OT Zpri values: 5.6K-ohms (RCA) and 4.0-4.2K-ohms (Leo), and (b) their output distortion levels: 1.8% (RCA) and about 5% (Leo, at rated power into resistive load).

...and, surprisingly, the PI output drive signal levels to the 6L6CG's for both amps are identical: 35Vg.rms (RCA for 55W) and 35Vg.rms (Leo for 40-45W).

[stokes]

Okay,so if I got this right the RCA example,you cite is right out of the book,yes?And of course Leo's is obvious.So it kind of confirms part of my point,we are obviously getting at it two different ways,and that point is, that in guitar amp world there are a lot of liberties taken and the parameters are stretched at times.And as you point out,the book almost always goes for the lowest distortion levels,while we dont.Choice of OT's makes a world of difference as to the parameters the tubes are working within,blah,blah,blah these are all things we agree on,so,dont laugh yet,using the mathematical examples as you do gets us to the book values,which would seem to be useless in guitar amps,(okay,not totally,but stick with me here),okay,now you can laugh,I am going to say that in the case of a SE guitar amp idling the power tube at 90% dissipation will get you in class A "territory".Sorry,but I am a thick headed old Irishman.And this all started how?I gotta go back and read the first post over again.

[loudthud]

The plate curves on the JJ 6V6 are for triode operation so they aren't much use for predicting bias with a pentode connection. The spec sheet lists the max plate dissipation in triode mode as 10W so why would they draw a 14W line? I'm happy to see the published some sort of spec sheet though.

(side note) O'Connor writes in one or more of the TUT books that you should still put a resistor in series with the screen grid when connecting tetrodes as triodes. This is probably a good idea in guitar amps to limit screen current since screen current spikes up when the tube is driven hard.

[Old Tele man]

(side note) O'Connor writes in one or more of the TUT books that you should still put a resistor in series with the screen grid when connecting tetrodes as triodes. This is probably a good idea in guitar amps to limit screen current since screen current spikes up when the tube is driven hard....+1, also, GEC always recommended use of at least a 100-ohm screen grid to plate resistor when operating any of their "KT-" tubes in triode mode...something that was seldom specified by USA tube manufacturers.

[martin manning]

The plate curves on the JJ 6V6 are for triode operation so they aren't much use for predicting bias with a pentode connection. The spec sheet lists the max plate dissipation in triode mode as 10W so why would they draw a 14W line? I'm happy to see the published some sort of spec sheet though.

Yes, now that you mention it those are triode-shaped characteristics, aren't they... duh.

Anyway, I went ahead and e-mailed the folks at JJ. The response came back that their 6V6-S is a 14W max Pa tube, and the line on the chart is "approximate." So there you go. I guess.

MPM