Biased diode clipping Q

[Chuck H]

In one of my EL84 designs I use the "Paul Ruby mod" to eliminate grid conduction and the subsequent crossover distortion. This works great, but I've always wanted to use another circuit so that "my" design didn't include someone else's known "mod" (petty vanity ). I also appreciate eloquence in design and I thought I had achieved both with an idea I had. I thought to use plain silicon diodes (that wouldn't require specific voltage selection like zeners) and bias them at the top of the power tube cathode.

Here's the design with the PR mod (that works):


Here's my idea with the diodes biased at the power tube cathode (which doesn't do anything?):


Why didn't this work or do much of anything at all?

NOTE* I only scoped the output and failed to check signal at the grids before returning the circuit to original. DOH!

[eschertron]

OK, I'll bite. It looks to me as if - in design #2 - when the grid goes ~0.6v more positive than the cathode, a short-circuit is created between grid and cathode, outside of the tube. So the cathode still sees the grid conduction current [edit: add even though the grids aren't actually conducting]. Am I close?

[Chuck H]

Thanks. That's pretty much what occurred to me. I'm going to run some ltspice stuff now to try and see what else I might be missing.

[Malcolm Irving]

Without either mod., when the output stage is not overdriven, the AC voltage across the .022 coupling cap from the PI stays low at all times.
Once the stage is overdriven, when the voltage at the power tube grid goes more positive than the power tube cathode, we get grid conduction. This means that the power tube side of the .022 cap is held (or clamped) at about the same voltage as the power tube cathode. Meanwhile the other side of the cap can go very positive and we have a big voltage difference built up across the cap. This voltage difference stays put, until the cap has time to discharge. The output stage is now blocked, as negative going voltage swings from the PI now have an additional big negative put on to them, thanks to the voltage across the cap. The output tube is now cut off far more than we want.
The Paul Ruby mod works by limiting the extent that the power tube side of the .022 cap can go negative.
The other mod cannot limit how far that side of the cap can go negative, but just adds a parallel path to the grid conduction path, increasing the clamping.

[Chuck H]

Gotcha. Nicely explained (I need that sometimes ). Thank you.

[Dave H]

In one of my EL84 designs I use the "Paul Ruby mod" to eliminate grid conduction and the subsequent crossover distortion. This works great, but I've always wanted to use another circuit so that "my" design didn't include someone else's known "mod" (petty vanity ).

Try this one. It doesn't look quite like the Paul Ruby mod and it's more elegant in that it uses fewer components.

[Chuck H]

Yupper You shared that one with me before. I do plan to use it on the next build where I need the circuit.

[Mike Sulzer]

I really like that one. Classy!

Try this one. It doesn't look quite like the Paul Ruby mod and it's more elegant in that it uses fewer components.

[ATTACH=CONFIG]40525[/ATTACH]

[Dave H]

You shared that one with me before.

Sorry about that. My memory is shot. Stick around and I'll share it again.

[J M Fahey]

I *always* say, "do not get tangled with the words", what matters is *functionality* , not *name*.
To say it in other words, it´s not about "parts" but *function*.

Or: "language is an excellent way to obscure things"

Problem with calling this "grid rectification" is that people get obsessed with the word "grid"

Fine if it´s clearly understood what it means in this particular case, but if not, it should be called: "peak detector probe made out of the coupling cap and the grid to cathode diode " (as if there were any other kind of diode )

Explain the difference between this:


and this:


when input signal peaks reach, say, 5V above diode forward conduction voltage.

Yes, both do exactly the same, the coupling cap gets charged about -4V negative compared to its original DC voltage.

Both "stolen" images show a couple extra resistors which are pertinent to the specific case, such as a grid reference resistor, very often 220k but varies from 100k (6550) to 1M (12AX7) in the tube gain stage and the extra 4M7 used in the probe to scale voltage for the meter but that does not change the way the circuit works.

So chuck h followed "the letter of the Law" (no grids were forced to rectify while filming this movie) but not "the spirit of the Law": "no diode will rectify audio voltage and vary bias", he simply relieved the grid of its *anode* futy by paralelling it with an external diode .... system stays and does the same.

[rant mode ON]
What do I think of Ruby´s Mod?

Not only "solving a non problem" but killing the good parts of "the problem".

If he likes linear amplification so much, why doesn´t he plain use bipolar transistors?
He also kills bite by adding an output snubber ... same thing.

Of course, such mods, and other horrors like snipping off bright caps, etc. are immensely popular among bedroom rockers, who sadly seem to be 90 to 95% of guitar players, and whine all day long about buzzy distortion, earpicking highs, etc. and turn amps into soul less boxes.

My only answer to all those whines is: play your amp full tilt , and if possible in a large room, because that´s what´s expected of them .... or ... get a modeler, designed to sound the same at any volume.
[rant mode OFF]

Next time some tree hugging type insists I boil or microwave my steak because all that carbon crust is bad and grease falling in the embers smokes it with unhealthy fumes I´ll kindly kick him with my combat boots in his arse until he is expelled from the dining room or patio.
Let me enjoy my grid rectification and biting spikey wave edges in peace.

This is what killer Marshall 18 does when overdriven:


and this is what an overdriven SS amp does :

"Look Ma !!!! .... no crossover !!!!! ..... rounded edges !!!!!! "

[Chuck H]

Ha ha!!! Juan.?. You're killin' me.

The PR mod is only the diodes on the PI to EL84 grid circuit. THE REST IS MINE including the "snubber" which doesn't kill "bite" in this case (Look at the cap value!?! 1500pf=tiny). It alleviates a voltage spike occurring for it's own reasons that I had to solve for because this amp lives dimed and the constant spike found it easier to dissipate in the screen grids than the plates because the OT impedance at that frequency is so high. It would have helped to use 1K screen resistors, maybe, but in the end it wasn't necessary.

I didn't include all voltages in the diagram but there's about 10V on the cathode. The voltage rises to 13V as the wave form starts to flatten. So I clamped the bias there with that diode across the cathode resistor to minimize bias cooling and keep the OD tone more like fixed bias, while still keeping the cathode bias sound for cleaner tones. It works a treat. Then I valued the PR mod zeners GREATER than the cathode clamping diode to minimize, but not eliminate loading (and crossover distortion). Notice that the PR circuit zeners are at 15V. I could have valued them the same as the cathode clamping diode BUT I WANTED A LITTLE CROSSOVER DISTORTION, but not a lot. With EL84's (not EL34's) excess crossover distortion doesn't sound like a touch wha, as it can with big bottles. With EL84's it sounds more like a can of bees. Instead of BWaaaahhhh you get FZWIizzzzz. A lot of players are aware of this and my circuit with the cathode clamp and method for valuing the PR circuit get's a little mention here and there WRT 18 watt types and AX84 discussions. It's just a way to make the tubes behave that we didn't consider before recently

But thanks for beating up my design

[hylaphone]

Try this one. It doesn't look quite like the Paul Ruby mod and it's more elegant in that it uses fewer components.

[ATTACH=CONFIG]40525[/ATTACH]

Neat! Which Zener is appropriate for this?

[Dave H]

What do I think of Ruby´s Mod?

Not only "solving a non problem" but killing the good parts of "the problem".

If he likes linear amplification so much, why doesn´t he plain use bipolar transistors?
He also kills bite by adding an output snubber ... same thing.

Of course, such mods, and other horrors like snipping off bright caps, etc. are immensely popular among bedroom rockers, who sadly seem to be 90 to 95% of guitar players, and whine all day long about buzzy distortion, earpicking highs, etc. and turn amps into soul less boxes.

I posted that schematic because I thought it could be a possible solution to Chuck's problem but I have never used it myself. I have gigged a 2 x EL84 amp for over 20 years and the standard circuit has always sounded just fine to me. Bright caps on the other hand should be banned from the face of the earth

[Chuck H]

Neat! Which Zener is appropriate for this?

It depends on other voltages. The cathode voltage on that circuit will rise as the amp is pushed, effectively cooling the bias. At that point the amount of negative swing necessary to place the power tubes into cutoff also rises. Since you probably DON'T want to hear any diode clipping in your tube amp you'll want to push the circuit as hard as it will commonly be pushed, measure the voltage on top of the cathode resistor and choose a zener voltage one volt higher.

If you use a clamping diode, like my design above, you can simply value the zener a volt higher than the cathode clamping zener voltage.

[nosaj]

Explain the difference between this:

i have to say all I see is a sad tube.

nosaj