current and cathode bias SE

[kg]

Actually, if it were symmetrically biased, average current would stay the same and so would the voltage drop across Rk.

jon, that would be true, with an important caveat--a perfectly linear tube. basically you'd have to have a tube in which the spacing of plate curves against the load line did not "space out" at higher g1 voltage and "compress together" at lower g1 voltages.

i haven't seen one of those yet!

in every real incarnation of class A, the average plate current DOES rise... no cutoff necessary. i know it's all over the internet that a class A amp doesn't exhibit bias shift (even good old randall aiken states as much on his site), but my beat up old RDH4 says differently. there are literally plots showing two different load lines.. same slope--but the dynamic line is translated downward by the amount of ADDITIONAL negative bias voltage generated across the Rk.

i'll get the page/section numbers when i get home today.

this presents a problem to self/cathode biased output stages (at least for the hifi designer who wishes to reduce THD as much as possible), since that means the operating point under dynamic conditions will shift WRT the operating point under quiescent conditions. even worse, since it is dependent on the even order distortion of the output stage itself, to determine the extent of the effect requires an iterative process.

i will readily admit it doesn't really present a problem worth worrying about for MI use (thank god!).

ken

[Steve Conner]

Yes, the only significant problem for MI use is if the booze and groupies fail to show up.

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.

[lowell]

Ok I'm confused about this stuff. I get it however that with a fixed resistance, as voltage drops, so does current as well. Ohm's Law. So that's what is happening in my amp. However, is this because it's single ended? What about the Paul Ruby zener mod? I was under the assumption that this keeps the cathode voltage from rising during distortion.

[Wilder Amplification]

Ok I'm confused about this stuff. I get it however that with a fixed resistance, as voltage drops, so does current as well.

You've got it backwards. A decrease in current flow through a resistance causes the voltage drop across said resistance to decrease. The voltage drop across the resistor is not the source of the current flowing through the resistance. It is the current flowing through the resistance that is the source of the voltage drop across said resistance.

Ohm's Law. So that's what is happening in my amp. However, is this because it's single ended? What about the Paul Ruby zener mod? I was under the assumption that this keeps the cathode voltage from rising during distortion.

It keeps the cathode voltage constant no matter what. That's what diodes do...maintain a constant voltage drop regardless of current flowing through them. Every diode has a "reverse breakdown voltage". This is the voltage that must be applied to a diode to make current flow through it in reverse. For a normal diode, it is a very high voltage, but due to the current at that high voltage normal diodes cannot handle reverse current very well.

Zener diodes, on the other hand, are MADE to handle reverse current. They have much lower reverse breakdown voltages. As long as there is enough reverse current flowing through the Zener (the Zener current), it will maintain a constant voltage drop across it that is equal to its reverse breakdown voltage (some call this the "zener voltage") regardless of how much current is flowing through it just so long as the current flowing through it is greater than or equal to its rated zener current

So if you had a 15V zener diode from the cathode to ground, your bias at the grid would be -15V relative to the cathode. The cathode voltage would maintain a constant +15V whether the amp is at idle or under signal.

[lowell]

It keeps the cathode voltage constant no matter what. That's what diodes do...maintain a constant voltage drop regardless of current flowing through them. Every diode has a "reverse breakdown voltage". This is the voltage that must be applied to a diode to make current flow through it in reverse. For a normal diode, it is a very high voltage, but due to the current at that high voltage normal diodes cannot handle reverse current very well.

Zener diodes, on the other hand, are MADE to handle reverse current. They have much lower reverse breakdown voltages. As long as there is enough reverse current flowing through the Zener (the Zener current), it will maintain a constant voltage drop across it that is equal to its reverse breakdown voltage (some call this the "zener voltage") regardless of how much current is flowing through it just so long as the current flowing through it is greater than or equal to its rated zener current

So if you had a 15V zener diode from the cathode to ground, your bias at the grid would be -15V relative to the cathode. The cathode voltage would maintain a constant +15V whether the amp is at idle or under signal.

But in this case if the zener is 15v and the cathode voltage goes below that then the zener will have no effect right? Only if Vk is above the zener voltage will the zener conduct... ?

[Wilder Amplification]

But in this case if the zener is 15v and the cathode voltage goes below that then the zener will have no effect right? Only if Vk is above the zener voltage will the zener conduct... ?

And how could the cathode voltage go below that with a 15V zener diode biasing it at a constant 15V?

[lowell]

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?

[Wilder Amplification]

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?

OK you never mentioned that it would be in parallel with a cathode resistor. I was under the impression that the valve would have just a zener diode only.

And why would you want to do that anyway?

[lowell]

Oh, sorry. I would do that so the amp is cathode biased up until clipping... which I was under the impression that Vk rises during clipping. Guess I read something wrong or remember it wrong.??

[lowell]

Ok I found a post from Chuck that I had read. Chuck, so in your amp Vk rises during clipping?

You simply need a ZV that is above the bias voltage. In the case of a cathode biased amp you need to go generously above because the bias voltage shifts. Since the positive voltage on the cathode is the bias instead of the negative voltage on the grid, and rises, you need to go above the rise to insure no clipping of the waveform outside of the cutoff portion. Paul admits in the article that 7 ZV is too low but that's all he had on hand at the time. For a fixed bias amp you only need to exceed the fixed bias voltage. The way I do it with cathode biased EL84's is to measure the cathode voltage, then choose a THIRD zener (5W) that is somewhat above that voltage to parallel the cathode resistor. This effectively "fixes" the bias whoenever the voltage rises to that point. I choose the ZV for that zener by using the scope to determine when clipping starts. That way I have an amp that is cathode biased when playing clean, but fixed bias when over driving which, IMHO is more ideal. But that's JMO. Then you can choose your Paul Ruby Mod ZV based on the parallel zener ZV. If you chose, for example, a ZV of 13 for the parallel zener you could then go with 15 ZV for the Paul Ruby Mod and be sure the bias won't shift beyond that point.

HTH

Chuck

[lowell]

Ok I found a post from Chuck that I had read. Chuck, so in your amp Vk rises during clipping?

You simply need a ZV that is above the bias voltage. In the case of a cathode biased amp you need to go generously above because the bias voltage shifts. Since the positive voltage on the cathode is the bias instead of the negative voltage on the grid, and rises, you need to go above the rise to insure no clipping of the waveform outside of the cutoff portion. Paul admits in the article that 7 ZV is too low but that's all he had on hand at the time. For a fixed bias amp you only need to exceed the fixed bias voltage. The way I do it with cathode biased EL84's is to measure the cathode voltage, then choose a THIRD zener (5W) that is somewhat above that voltage to parallel the cathode resistor. This effectively "fixes" the bias whoenever the voltage rises to that point. I choose the ZV for that zener by using the scope to determine when clipping starts. That way I have an amp that is cathode biased when playing clean, but fixed bias when over driving which, IMHO is more ideal. But that's JMO. Then you can choose your Paul Ruby Mod ZV based on the parallel zener ZV. If you chose, for example, a ZV of 13 for the parallel zener you could then go with 15 ZV for the Paul Ruby Mod and be sure the bias won't shift beyond that point.

HTH

Chuck

[Wilder Amplification]

Zener diode or not, it's cathode bias no matter what. The proper definition of cathode bias is when the bias is derived by referencing the grid to ground via a leak resistor and applying a positive voltage to the cathode, whether it be from an external supply or it be from deriving a voltage drop across a cathode resistor to elevate it positive, which makes the grid more negative than the cathode (or it can be said the cathode is made more positive than the grid). Whether or not the cathode voltage is fluctuating, it is still considered cathode bias if the cathode is elevated above ground potential while the grid leak is tied to ground potential.

Fixed bias is called "fixed bias" because the grid instead is tied to an external fixed voltage reference that is negative relative to ground while the cathodes are at ground potential.

For the record, the cathode bypass cap across the cathode resistor is supposed to be large enough in value to hold the cathode voltage constant at all audio frequencies.

[redelephant]

Traditionally cathode bias is achieved by the resistor from cathode to gnd and the cathode current giving the required voltage between grid and cathode for proper quiescent current. Another term used is auto-bias b/c the current thru the tube automatically sets the voltage...this we all know... But when the cathode is placed at a voltage with a low impedance, so the voltage doesn't change when DC current varies, it no longer is auto bias. Can we call it fixed-cathode bias?
Those who build for long tube life favour cathode bias b/c a thermal runaway is prevented by any eventual current increase producing a larger grid-cathode voltage in turn hindering current from increasing further. For instance in some books where longevity is the missino, fixed bias is not recommended. The fixed voltage, either from a zener or any voltage source, will not prevent thermal runaway and since the cathodes are 'locked' to a fixed voltage, the tube behaves more as if fixed bias is used.
I've been so used to building class-A amps that I was surprised when I discovered that cathode voltage went up from 30V to 48V in my latest class-AB2 build, (when overdriving). It had no effect if the cathode cap was 10uF or 1000uF, the latter only made recovery longer. I should've known, but hadn't thought about it. As soon as the amp starts conducting more current than it's bias is set for, the current is no longer average but rising. In a class-A build it will be pretty constant, and the cap shunts all signal swing to gnd, but as soon as it starts to clip even a class-A amp will likely either see the cathode voltage go up or down. If the plate load is a relatively small impedance, the plate can swing more current and the cathode voltage rises, if the plate load is relatively high impedance, the plate will swing more down than up (using less current) , and the cathode voltage sinks.

[Chuck H]

Zener diode or not, it's cathode bias no matter what. The proper definition of cathode bias is when the bias is derived by referencing the grid to ground via a leak resistor and applying a positive voltage to the cathode, whether it be from an external supply or it be from deriving a voltage drop across a cathode resistor to elevate it positive, which makes the grid more negative than the cathode (or it can be said the cathode is made more positive than the grid). Whether or not the cathode voltage is fluctuating, it is still considered cathode bias if the cathode is elevated above ground potential while the grid leak is tied to ground potential.

Fixed bias is called "fixed bias" because the grid instead is tied to an external fixed voltage reference that is negative relative to ground while the cathodes are at ground potential.

For the record, the cathode bypass cap across the cathode resistor is supposed to be large enough in value to hold the cathode voltage constant at all audio frequencies.

Yup. Cathode bias means the cathode is the source for the 0V differential used to bias the amp. Perfect. But I'm not sure we can define "fixed" bias as being a -V applied to the grid. You could say "grid biased" I suppose. If, for example, you had a standing voltage of 15 volts on a cathode resistor and you parallel a 15V zener across it you have "fixed" or secured the bias to that point. So I suppose this would be fixed cathode bias. Semantics and all aside.

IMHE a bypass cap won't stabilize the voltage at the cathode resistor. There is a rise in cathode voltage as current increases.

[kg]

Zener diode or not, it's cathode bias no matter what. The proper definition of cathode bias is when the bias is derived by referencing the grid to ground via a leak resistor and applying a positive voltage to the cathode, whether it be from an external supply or it be from deriving a voltage drop across a cathode resistor to elevate it positive, which makes the grid more negative than the cathode (or it can be said the cathode is made more positive than the grid). Whether or not the cathode voltage is fluctuating, it is still considered cathode bias if the cathode is elevated above ground potential while the grid leak is tied to ground potential.

be careful with this.

cathode/self bias REQUIRES degeneration at DC. put another way, if the Q current through the tube goes up and the effective bias voltage does not move in a direction to counteract this, then it's not cathode bias.

although some biasing methods are inserted in the cathode circuit (ie, LED, zener, stack of traditional PN diodes, neon tube, battery, etc), they can't properly be called "cathode/self bias."

whether or not the cathode is at ground or not does not matter one bit.