El 84 fixed bias amp & Paul Ruby zener mod.

[Alf]

Some time ago I read about the Paul Ruby zener trick for cathode biased amp to reduce the harshness when diming the amp ( el84 tubes).
Can this also be done with a fixed bias amp and how should I connect the zeners ( about 12 volt , I think) ?

Can I connect the zeners to ground after the coupling caps and before the 220K bias resistors ? Does this influence the bias ?



Alf

[nickb]

As I understand it the purpose of the zener (+ regular diode) is to limit the maximum negative excursion of the grid signal to match the positive excursion (limited by grid conduction) so that the coupling capacitor gains no net charge and so avoids blocking distortion.

It seems to me that you could achieve exactly the same effect with less components by using two back to back zeners connected between the power tube grids. The peak to peak swing will be limited to the zener voltage + ~0.7V. Choose the swing to be about the bias voltage x2. This shows it for a cathode biased output but it works just as well for fixed bias with no changes. Note, the source impedances need to be similar.


Another way of looking at this is that it's just an appropriately sized diode clipper that limits the signal applied to the output stage and could in principle be anywhere after the master volume.

[Alf]

Okay Nickb so I can leave the original bias as it is and just add these two back to back zeners.
I'm gonna try that , thanks.

Alf

[trobbins]

nickb, I don't think the back-to-back zeners are going to achieve the same effect as the ruby style mod. For example, if both coupling caps get charged by the same voltage level by equivalent levels of grid conduction from each grid, then the zeners don't see that.

[frus]

BTW, I tried it with fixed bias amps before, the way Alf suggested in the first post (zener+1N4007 to ground after coupling caps). works fine.

[Alf]

I'll try both ways !
Frus, what zener did you use ?

[frus]

it was a 6L6 amp, but I would use something like bias voltage, maybe a few volts more

[Steve Conner]

For a fixed bias amp, you need to clamp the negative peaks at the grid to a little more than twice the bias voltage. Say -120V for a 6L6 amp.

[frus]

why twice?

[Alf]

For a fixed bias amp, you need to clamp the negative peaks at the grid to a little more than twice the bias voltage.

Yes , why , I am wondering too !
In my el84 amp I have seperate biaspots and use - 13 V bias so I would have to use 26 volt zeners ?

[Steve Conner]

OK, here goes.

In a cathode biased amp, the grids normally sit at 0v, and the cathodes at +Vk. At full drive, the grid voltage swings from +Vk (clipped by positive grid current) to -Vk where the Paul Ruby Zener kicks in.

In a fixed bias amp, everything is moved down by one Vk. The tube cathodes are grounded, the grids are biased to -Vk, and so the full power grid swing is from 0 to -2Vk, around a quiescent value of -Vk. Except we would call it Vg1 or Vc or Vbias or whatever.

You can also use a symmetry argument. In a push-pull amp, if one grid swings up from -Vc to 0, the other one must swing down from -Vc to -2Vc, in order for the PI output not to be clipped prematurely.

Most guitar amps are Class-AB, so the tube will cut off quite a while before the grid voltage reaches -2Vc. But clipping one side of the PI will affect the signal on the other side. This is most obvious with the cathodyne PI, where clipping the cathode output increases the gain of the plate output.

So to sum up, if we keep the Zener voltage greater than Vk in a cathode biased amp, or 2Vbias in a fixed bias amp, then we won't hear any diode clipping, it will just act to stabilise the DC conditions.

[Alf]

A very clear explanation, thanks Steve !

[frus]

oh yeah... Thanks mister!!!

[nickb]

nickb, I don't think the back-to-back zeners are going to achieve the same effect as the ruby style mod. For example, if both coupling caps get charged by the same voltage level by equivalent levels of grid conduction from each grid, then the zeners don't see that.

Ah ha! I thought you might ask. It works because it's magic

The single zener clamp has the same problem if you make it too big. Anyway you did make me work on it some to be sure and here are the simulation results comparing Ruby ("Clamp"), back to back zeners and unmodified . Pretty convincing, I think.
Clamp.pdf

[jimboyogi]

Thanks for posting the scope and analysis nickb - very clear as you say and both the 'clamp' and 'back-back' very effective at reducing xover distortion.

I wanted to ask, is xover distortion solely responsible for the 'harshness' that the OP spoke of?

And how do the overdriven EL84's sound with either of these mods installed?