question about power scaling

[katopan]

Great news! You're the first to try out such a low zener value as far as I know. Other people with your issue on 18Watt have had to resort to other methods of attenuation or as said earlier implement steps to reduce the gain (like removing the power cathode cap).

Please do post a comparison clip. And don't worry about not having a scope. If it sounds good it is good. The test is to make sure it hasn't ruined the clean tone. If you're still able to get clean at the same volume setting as before the mod then the zener isn't encroaching on the power valve before cutoff.

I'll share a link to your success over at 18Watt. Only a couple of weeks ago someone was asking if the variable zener circuit had been tried out or not because they were having trouble.

[Lindz]

Sometimes NOT knowing what you are doing is an advantage

As a follow up: I put 5v/5 watt zeners in the Ruby clamp on my other cathode biased amp and it too will scale down without swirlies now. Seems like this small zener trick just might be the ticket.

I am still curious if someone in the know can explain how this is operating so I better understand it

Also, to follow up on an earlier question I had..

Is there a recommended voltage ratio for plates vs screens with
el84's?

This 2nd amp still has screen voltage higher than plates by a few volts.

[gary]

So, if I'm following correctly, you've applied a zener that is not within 2 volts of the bias on a fixed bias AND a cathode biased amp, both of which have a Hall vvr fitted. The amps sound normal throughout the vvr range? How low are you going on the vvr?

I second the point about hoping we can get a scientific evaluation of the effects of this mod.

[katopan]

How the Ruby zener mod works is explained in his doc about it. Nothing really changes using lower zener values except the clipping by the zeners on the negative side is no longer symmetrical with the power grid clipping on the positive side. With the smaller zener it just means they are effective for longer as the VVR and therefore cathode or fixed bias voltage, and therefore power grid clipping threshold is dialled down. What else are you guys after an explaination on?

One interesting experiment would be to see how low you can go. In theory the extreme is to ditch the zener and just use the normal diode to clip at -0.7V. In theory this should be clipping before power valve cutoff and introduce some unwanted distortion when the amp is set for what used to be clean. You could then adjust up from there.

I know some amps with a choke instead of a resistor in the screen supply chain run with the screens at higher voltage than the plate. But it is harder on the screens as they draw more current. I wouldn't be comfortable with any of my amps like that, and would add either a resistor in series with the choke or up the value of individual resistors going from the screen supply filter cap to the screens. These two options will effect the feel differently.

[Lindz]

Katopan - my question about the Ruby mod is more related to my general lack of understanding of the finer nuances of tube circuits in general - not so much the crossover distortion concept in Paul's article but more that I was going so far below what he proposed for zener values and was not really sure what I was doing by taking it to that extreme

your last few posts in this thread are helping it all gel for me

Gary, essentially I started experimenting with smaller zeners that were tighter in value to those Paul Ruby proposed to kill crossover distortion because I was unable to scale my vvr down without crossover distortion appearing

He suggested 1 - 2 volts higher than grid cutoff voltage, and most threads I have seen where people have implemented Ruby clamps used the cathode bias voltage at idle to set this value - On my amp I had 15v zeners in the ruby clamp that worked great at 100% power but the distortion reared its ugly head pretty quickly when I scaled down the amp

When I started trying the lower value zeners I noticed on my amp a correlation between the bias voltage and zener value in the Ruby clamp where if the zeners were 10v, I could dial down my vvr to where the bias voltage at idle was -11v and not hear the crossover distortion that is otherwise prevalent in my amps without the Ruby clamp - a 1 volt difference (ignoring the negative).

9volt zeners, I could scale the amp down to -10 volt bias at idle with no crossover distortion - this correlation continued as I went lower in zener values

If I scaled the amp below the correlated zener value, I started to get crossover distortion progressively worse as I went lower

I tried a few lower values until I tried 3volt zeners the other day at which time I could dial my vvr down as far as it goes on my fixed bias amp

I tried 5v zeners on another cathode biased amp that was buzzy without a Ruby clamp and it also seems to work pretty well all the way down

I have not tried the 5v zeners on the fixed bias amp nor did I have every value zener under the sun to try - I just had a few random voltage zeners going down from 16v (which is 2 volts higher than my idle bias voltage of -14v with the amp at 100% b+ voltage on the vvr - again ignoring the negative) and found that a low value zener seems to solve the scaling issue

[gary]

Lindz, thanks for the additional description. Somehow, I need to school myself on these mechanisms a little more. The ultimate 18 watt scaling solution may be within our grasp!

Any chance of some clips?

[dass101]

Maybe you can devise a way in which the power scaling is not done as continous variation, but in a few steps, let's say 1/2, 1/4, 1/8, 1/16. In this way you could also switch the relevant Zener diodes.

[katopan]

Dass, others have put in a set of switchable zeners of different values as a solution to this in the past. Works well over the set ranges, but what Lindz has found is even better because you don't have to change the setting.

[Lindz]

I'll post a few clips to show the comparison between no clamp and this new clamp w/power scaling in a few days. My amps are upstairs on the bench right now but the studio is downstairs and in some disarray right now because I was re-arranging my racks.

[husky]

I have a question for some educated ears and folks that know a bit about power scaling..

I like a level control before the phase inverter and I only scale the power tubes. Then turn down the level with the power scaling. Nothing else floated my boat except for this in feel and tone and I tried every possible scenario of power scaling and Master Volumes. As long as people turn down the Level I don't have an issue.
I'm working on a programable scaling setup where we can control every voltage in the amp as well as the level going in to the phase inverter. I think that is the only real solution.

[Firestorm]

I have a question for some educated ears and folks that know a bit about power scaling..

I like a level control before the phase inverter and I only scale the power tubes. Then turn down the level with the power scaling. Nothing else floated my boat except for this in feel and tone and I tried every possible scenario of power scaling and Master Volumes. As long as people turn down the Level I don't have an issue.
I'm working on a programable scaling setup where we can control every voltage in the amp as well as the level going in to the phase inverter. I think that is the only real solution.

Hey John,

Is there no post-PI circuit you like for a level control with power scaling? It raises havoc with negative feedback, but some amps seem to gain part of their character by overdriving the PI.

[llemtt]

I like a level control before the phase inverter and I only scale the power tubes...

Hi John

I see your point and almost agree with you but that way I end up with too much negative bias shift with the power tubes and the annoying crossover distortion this causes.

I think PI headroom should be limited too, what are your PI scaling experiences? Why don't you like it?

thanks
Teo