Zener on Rk to "fix" bias

[Chuck H]

I want to try using a zener across Rk on my cathode biased amp to help tame the crossover fizz at full dime. With 9.5 volts on top of Rk at idle and 18 volts at full output, I find a 13 volt zener gives good clean compression attack but snugs up nice and tight when cranked and helps a bit with crossover distortion.

My trouble is that I don't know how to figure the power rating for the zener. The whole avalanche thing and current rising while voltage is stable thing is confusing me. The "calculators" I find on line are all for a typical resistor/zener series regulator, not an absolute shunt as in this application.

I'm using a 3 watt 13 volt zener now because thats what I have on hand. It hasn't blown up yet but I fear it's very close to it's rating.

Thanks

Chuck

[loudthud]

You have to measure the current going through the zener some how. Some DVM's will measure current directly. You would need to disconnect one end of the zener and connect your meter in series with it. Some meters only go to 200mA, that would be 2.6W with a 13 volt zener. Another thing would be to connect a 1ohm 1W resistor in series with the zener and measure the voltage across it. Measure when you are playing full volume. A good rule of thumb is to double the current measurement and get the next higher power rated zener.

[Chuck H]

Thanks loudthud. That'll do.

I hadn't even considered the 1 ohm resostor trick. Perfect.

Chuck

[Chuck H]

The one ohm current sense resistor revealed 40mA. Thats with an 800Hz sine pushing the amp into a fully square wave and no headroom left. Would I multiply this by the voltage dropped across the diode like you would with a resistor? Or with a diode do you use the source voltage or something?

Either way I'm good as far as I can tell. The voltage across the diode is 5 and the voltage on top is 13 (of course, it's a 13 volt zener). Even if I factor by the total voltage of 18 thats still less than a watt. Is this because it's parallel with Rk instead of in series with it so Rk is sharing some of the work?

It seems odd because the schematics I've seen for this trick use two 5 watt zeners in series...Why? Am I missing something here?

Chuck

[loudthud]

I'm totally lost. How do you have 5V on one end of the zener? The power dissipated by a zener is simply the voltage across it times the current through it.

[Chuck H]

Maybe I stated that wrong. I know the original source voltage on top of Rk to be 18 volts at full output. Shunting the 13 volt zener across it of course regulates it to 13 volts (actually 13.25). So I have to assume that the voltage across the diode is 5 (actually 4.75). I expect if I had 100 volts on top of Rk with the 13 volt zener in place it would still regulate it to 13 volts. And there would be 87 volts across the diode. Am I wrong on this. I'd like to know what wattage to use.

Chuck

[loudthud]

In your example the diode doesn't have 5V across it. It "adjusts" the cathode voltage by sucking enough current around the existing cathode resistor to keep the voltage at 13 volts. So where did the 5V go (18-13)? The zener didn't allow it to develope.

This is wrong thinking but I'll tell you why in a second. If your cathode resistor was 250 ohms, at idle you had 9.5V/250ohms=38mA. At full volume 18V/250ohms=72ma. With the zener clamping at 13V you might assume that the current thru the resistor is 13V/250ohms=52mA so zener current is 72mA-52mA=20mA. The reason this is wrong is because with only 13V of bias, the tubes will draw more current than they would with 18V bias. I don't know what your cathode resistor actually is so YMMV. Anyway, if you just measure what the current in the zener is, multiply by 13V and you have the zener's power dissipation. Double it for a safety margin and you should be ok!

[Chuck H]

Gotcha. I'm golden then. The current through the zener is 40mA with the amp flat out and clipping hard. But thats with a generator, not a big, crunchy, 'swinging cats by the tail' guitar blaring through it Still, I think I'll be fine. I just ordered 5 watt axial leads for the job though it seems that the 3 watt would be fine.

Thanks

Chuck

[lowell]

this is an interesting thread. So do you place the zener in parallel with Rk or in place of it? If I want to fix a 12ax7 at 1.5v on cathode do I use the standard resistor that gives that w/ the zener (what voltage?) in parallel w/ it? Finally, does this rid of negative feedback that is inherent in unbypassed tube cathodes or is something else happening current-wise?

[melvin]

The way I understand it, this thread refers to a well-known(?) arrangement that uses a couple of diode drops (.6V each) to cathode-bias a triode. With such an arrangement there is no need for a bypass cap, but a snubber may be necessary because the diodes may switch on and off. And no need for a cathode resistor either.

I've never tried either myself, but it sounds like the zener used in this case is to limit the bias drift in a cathode-biased output section, therefore limiting the compression or increasing the ac bypass effectiveness at high volumes.

I suppose a 1.5V zener would work as well for a 12ax7?

[mooreamps]

Originally Posted by melvin

I suppose a 1.5V zener would work as well for a 12ax7?

It probably would work, but it would be much noisier than going with the fixed bias approach.

Yes, maybe snub it with 100pf to attempt to midigate the zener noise.


-g

[lowell]

how is it that this rids of negative feedback, if that IS the case? Wouldn't the diodes only pass the negative portion of the AC on the cathode?

[Steve Conner]

What it supposedly does is to prevent the cathode bypass cap from charging up when the stage is overdriven. Or in Chuck's original case, when it's driven into Class-AB. Even with the cathode cap clamped, the grid coupling cap will still charge up, but there's another zener mod for this. The bridge rectifier clipping circuit in later Marshalls does a similar thing.

1.5V zeners are really lousy, not much better than a red LED or two ordinary diodes in series. Some people do like to use red LEDs as cathode loads for preamp tubes.

Trying to fix the cathode voltage like this has one major disadvantage: The circuit loses the ability to adjust itself to different preamp tubes with slightly different characteristics. This is why they used cathode bias in the first place.

[lowell]

why does the cap charge up when the stage is overdriven? Is it because the guitar/ac signal is pulsating DC so it charges the cap?

[mooreamps]

Originally Posted by Steve Conner

Trying to fix the cathode voltage like this has one major disadvantage: The circuit loses the ability to adjust itself to different preamp tubes with slightly different characteristics.

you want to put some money on that ?????


-g

[Steve Conner]

Yes, it's a scientific fact. We're betting on last week's lottery results here, not next week's.

The change in bias points between tubes will be worse by a factor of whatever cathode resistor you'd have used, divided by the dynamic resistance of the LED or zener or whatever you actually used in place of the resistor. So probably about 10.

[mooreamps]

Look, maybe I'm not understanding "exactly" what type of circuit we're all talking about here ; and maybe you're a fairly educated guy and all, but when you go with fixed bias, there just are no "dynamics" that change.


-g

[Dave H]

If we are talking about a fixed bias preamp stage here, it's going to behave just like a fixed bias output stage does. When the tube is changed for one with different characteristics the bias voltage has to be adjusted for it to have the same bias current as before but a diode biased preamp stage doesn’t have a bias pot so the bias point will change.

[Chuck H]

To expand a little on what Dave said...

If you fix the bias with -V on the grid or +V on the cathode and a respective 0V reference on the other end, the tube doesn't care. It's fixed bias. It is adjusted for the tube in the amp now for a chosen operating point. Plug in a tube with different operating characteristics and that operating point will certainly be different if no bias adjustment is made. That isn't to say it won't "work". Just that the operating point will be different from the original, intended design. As far as I'm concearned this IS a drawback considering the variability between preamp tubes. And not just different types. There is a lot of difference even from one brand of 12ax7 to the next.

FWIW the original post was in regard to a diode across the shared cathode resistor for a pair of EL84 tubes. The amp plays cathode biased, with soft attack and a little breath into the onset of distortion. As the crossover distortion starts to appear the amp "fixes" and tightens up a bit. This, for me, is helping to keep crossover distortion to a minimum and get a just overall bigger sound from a 2x EL84 amp. I'm also using the "Paul Ruby mod" with the zener values set 1 volt above the cathode zener voltage. The last bit of overshoot hash is eliminated with a conjuctive filter of 10k/1500pf which is barely audible except for the beneficial smoothing of the overshoot. I'm running the tubes at 355Vp and 330 screens. I am VERY pleased with the performance of the EL84's using this circuit. With the highish voltage I'm getting good bottom end but it seems to keep a little of the jangle. Though not nearly as sweet as running these tubes at a lower voltage. But that's not what I was going for. This sounds more like a little version of a big bottle tube in the way it plays under overdrive. But it keeps the cool cathode bias vibe from clean up into lightly distorted tones. There are no audible penalties due to the diodes.

Chuck

[mooreamps]

Originally Posted by Dave H

a diode biased preamp stage doesn’t have a bias pot so the bias point will change.

Fine, go build one and find out.

-g

[Dave H]

Originally Posted by mooreamps

Fine, go build one and find out.

-g

I did build an EL84 power amp with a 15V zener in the cathodes for bias and it worked fine. I’ve never tried it for a preamp stage but I have noticed that for a resistor biased preamp the plate voltage can change quite a bit when the tube is changed so I’d expect the plate voltage of a diode biased stage to change even more without the negative feedback from the resistor. Anyway unless you always want it to be biased for max headroom the diode will most likely be close enough for rock and roll.

Dave H.