current and cathode bias SE

[kg]

I feel like I'm taking crazy pills here. If a zener diode is placed in parallel with a cathode resistor, the zener will only keep the Vk from rising ABOVE the zener voltage, not below. Thus, if it's a 15v zener, and the Vk drops to 14v, then it will do just that, because the zener is basically not in the circuit until the Vk rises above 15v. Am I taking crazy pills?

lowell, this situation, given that the zener is sized a couple of volts higher in value than quiescent cathode voltage, is a hybrid self/fixed bias setup: self at lower power outputs, and as the output power rises, gradually becomes more of a fixed bias amp.

tbh the only advantage i can see with doing it this way is that the amp will be more resistant to thermal runaway (a hallmark of cathode bias vs. fixed, and an effect that leads to the larger max Rgk ratings on the datasheets). then again, that resilience would only be present at small/no signal, which may not be enough to save it.

[lowell]

RE,
That makes sense now. My amp is class A AND I'm using a 5k resisting load for my 6v6. This must be why the parallel zener trick won't work. Hmmm, interesting.

[lowell]

RE,
That makes sense now. My amp is class A AND I'm using a 5k resisting load for my 6v6. This must be why the parallel zener trick won't work. Hmmm, interesting.

[TDS]

RE,
That makes sense now. My amp is class A AND I'm using a 5k resisting load for my 6v6. This must be why the parallel zener trick won't work. Hmmm, interesting.

I saw a similar thing last night with my newly completed a 3W S/E amp - SLO preamp with a class A 6V6 with a 5k transformer load without NFB, 250V B+.

With a 220 Ohm cathode resistor (47uF bypass cap) I get 12.4 volts quiescent and up to the point of clipping. When overdriven, the voltage drops to 11.2 volts with a square wave output on extreme overdrive. I had expected it to increase and had two series 6.2V zeners across the cathode resistor, but I've now removed them as they clearly have no effect.

[lowell]

TDS thanks for confirming this. How does your amp sound? Seems very gainy, but cool!

[Chuck H]

tbh the only advantage i can see with doing it this way is that the amp will be more resistant to thermal runaway (a hallmark of cathode bias vs. fixed, and an effect that leads to the larger max Rgk ratings on the datasheets). then again, that resilience would only be present at small/no signal, which may not be enough to save it.

There are other advantages. I use it on one of my AB designs to "fix" the bias as the tubes start to distort. I still get a little envelope when playing clean but the amp tightens up when overdriven. Also, on my scope there is a significant reduction in crossover distortion since the bias no longer shifts cold. I only use it with EL84's that are prone to sounding ugly with too much crossover distortion instead of a little swirl like you get with the bigger bottles. EL84's don't so much "swirl" as BeWIZZZerrrzz.

P.S. This info is only applicable to AB designs. Class A amps biased correctly don't have a voltage rise at the cathode. I only key'd up this thread because I saw a ref to an earlier thread I was on. Not hijacking, much.

[imaradiostar]

I've only noticed significant decrease in current once you start overdriving the output stage. I think the duty cycle drops below 50% because of grid current, so the average current drops down to less than half of the peak.

Thank goodness you said that- I was about to!

It also sounds like some people may be confusing the ruby zener mod (which eats up the excess charge on the output tube grid coupling caps to prevent grid blocking distortion) with using a zener diode to prevent the cathode from rising too much as more current is drawn.

It would be interesting to build a fixed bias SE output section with adjustable plate, screen and bias voltages and twiddle the values to see if you can find a place where the current draw stays more or less constant from idle to substantially distorted.

jamie

[Steve Conner]

Thank goodness you said that- I was about to!
It also sounds like some people may be confusing the ruby zener mod (which eats up the excess charge on the output tube grid coupling caps to prevent grid blocking distortion) with using a zener diode to prevent the cathode from rising too much as more current is drawn.

Well, they're kind of equivalent, they both attack the same problem from opposite ends. Both mods have a similar effect on the tone, and if you do both of them, you get twice as much of it.

It would be interesting to build a fixed bias SE output section with adjustable plate, screen and bias voltages and twiddle the values to see if you can find a place where the current draw stays more or less constant from idle to substantially distorted.

Yes, it should be possible to find such a place. AFAIK, it's the manufacturer's recommended operating point, or not far off it.

[Chuck H]

AFAIK, it's the manufacturer's recommended operating point, or not far off it.

Right. Now if we could just get specs accurate to the tubes being sold and not just recycled info from older tube construction we'd be in.

[TDS]

I saw a similar thing last night with my newly completed a 3W S/E amp - SLO preamp with a class A 6V6 with a 5k transformer load without NFB, 250V B+.

With a 220 Ohm cathode resistor (47uF bypass cap) I get 12.4 volts quiescent and up to the point of clipping. When overdriven, the voltage drops to 11.2 volts with a square wave output on extreme overdrive. I had expected it to increase and had two series 6.2V zeners across the cathode resistor, but I've now removed them as they clearly have no effect.

I have experimented some more today by changing the value of the cathode resistor upwards from 220 Ohms. It seems there may be a point close to the Class A application parameters given in the tube datasheets where the cathode current might remain steady, because when I increased from 220 to 240 ohms the cathode voltage drop began to INCREASE on extreme overdrive, going from 11.4 Volts quiescent to around 13 Volts when fully overdriven. The opposite to what I saw at 220 Ohms. The B+ and all other factors were unchanged, the power at the point of clipping had dropped from 3 Watts to 2.4 Watts. When I tried an even higher resistance (330 Ohms) the voltage rise was greater still and the power output dropped even further.

The B+ had risen a little (around 5 Volts) but I expect that if I had a higher B+ available and moved away from my original Class A conditions towards the typical "Champ" settings (360 Volts B+, 7000 Ohm load, 470 Ohm Cathode Resistor) I would get more power along with a even greater increase in Cathode current as power increased.

Could it be correct to think that the "Champ" PA is actually a form of S/E Class AB?

[Chuck H]

Could it be correct to think that the "Champ" PA is actually a form of S/E Class AB?

At todays AC wall voltage, yes. At least when the amp starts to clip. I can't say about how it would be when AC from the wall was 115V. The last VC I worked on (my wall AC averages 125V) I resd 428 Vp.

[redelephant]

Classes of operation go out the window when overdriving. A single tube in an SE amp does the entire cycle, all 360 degrees of it and even tho it my turn more one way than the other when clipping, it is regarded class-A.
Your load impedance together with the available voltage swing, determines what is the optimum quiescent current (bias). If you have 300volts of swing, a 5k ohm plate load, the idle current should be 300/5k, which is 60mA. If you bias it lower (idle current) it will be able to swing more down than up(the plate), and vica versa. This gives u the clipping behaviour and of course max clean power. Some like the warmth in unsymmetrical clipping while others prefer the fuzzier clipping of a symmetric ditto.
I'd say if your cathode voltage changes less than 10% during clipping you have it pretty much nailed in terms of optimum bias for max clean. Doesnt mean that gives you the optimum tone for you tho. Thats what makes it fun, guitaramp are allowed to do fun stuff ulike HiFi amp that need to be optimized and play along with the rules.

[mooreamps]

Classes of operation go out the window when overdriving.

and I'll just never go along with that... You bias a power tube for the text book definition of Class A, over-driven or not, it's still Class A.

-g

[redelephant]

and I'll just never go along with that... You bias a power tube for the text book definition of Class A, over-driven or not, it's still Class A.

-g

I thought thats what I said. Class of op is defined in the linear region of op. So if its class-A PP clean, it's still defined as class-A when overdriven even tho a tube's cycle is more off than on (or the opposite) during clip.

[Steve Conner]

I disagree. It helps me to think about these things if I consider that the stage "goes into Class-C" when overdriven. Or Class-D or E even. The average SE EL84 amp will overdrive so hard that it's more or less a square wave chopper.

Or I could have a push-pull amp that's "Class-A up to 1 watt output", say. Above 1W, one of the output devices cuts off at some point in the cycle: it goes into its Class-AB region.

I'm aware that this is not a standard use of the terms, but I'm not the only one who does it, I have seen it in writing elsewhere, such as Douglas Self's books on hi-fi amps.