Low voltage and output tubes

[dude]

I can't find much on a search but before I start tweaking an 18 watt 6v6 lite I'd like some advice on low B+.

With a B+ of 300/302dc hot, the amp is too creamy, buttery or brown tone-ish. Just not enough clean Fender like tone.  I know the circuit has a lot to do with it but I have other 18 watt "el84" amps that have that Fender-ish sparkle with a few cap and resistor changes.

So the question before I start tweaking this 18 watt 6V6 conversion, would low B+ of 300v be one of the causes? Cathode biased a little hot, 250 ohm, bridge rectifier,  used old stock 6V6s.

Just wanted to know if I'm trying to make something that's - what it is...?  I do have another PT for a Princeton that would raise that voltage by at least 75v, I'm assuming?  Or do you think I can tweak this "low voltage" build closer to a black face. I could post the schematic but I'm more concerned about the 300v before I get into it.

Thanks as always and learning everyday,
al

[HotBluePlates]

... 18 watt 6v6 lite I'd like some advice on low B+.

With a B+ of 300/302dc hot, the amp is too creamy, buttery or brown tone-ish. Just not enough clean Fender like tone. ...

People probably build the 18w because they don't want "another Fender." The preamps are entirely different. If you take out the 18w preamp and replace it with the 2 gain stages with volume & tone stack in between like a blackface Fender, then you'll get a lot closer to that sound.

But there is more to it...

...  I know the circuit has a lot to do with it but I have other 18 watt "el84" amps that have that Fender-ish sparkle with a few cap and resistor changes.

So the question before I start tweaking this 18 watt 6V6 conversion ...

I think the voltage is not the issue, but 6V6's vs. EL84's. I know everyone says that, but here's why:
6V6's are beam power tubes, EL84's are true pentodes. Beam power tubes have very little 3rd harmonic distortion compared to pentodes. And if the push-pull output stage were perfectly balanced, all even harmonic distortion would be cancelled, too.

In general, odd harmonic distortion adds an "edge" to the sound which makes it "present" or "bright" compared to low-order even harmonic distortion (and anything above 6th harmonic would be tiny regardless, so it really doesn't count here). To a limited extent, that may account for the "dark" or "brown" character that you're hearing.

Beyond that, I'm sure others can help. I'm no 18w expert.

[printer2]

Bias the tubes colder and see how you like it. I messed around with the bias for my 2W 5E3, the hotter I biased it the smoother it sounded. Part of the reason was that when the output is clipped the bias voltage goes more negative on the cathode cap. The more negative it goes the more closer you are to Class B and the more crossover distortion. It give more of an edge to the sound, more of an aggressive the sound. Too much and it sounds fizzy.

Other than the outputs maybe you can reduce the dropping resistors in the power supply so the preamp tubes run at more voltage. Maybe run a parallel node off the first stage for the preamp. Drop the voltage to about 285V and use a big cap, at least 22uF for the preamp stages. No reason they have to get their voltage off the screen node. Should help brighten up the sound. If you have too much hum, half the resistance and do two stages of filtering for the preamp.

If it helps, great, if not you can go with the higher voltage.

[HotBluePlates]

... when the output is clipped the bias voltage goes more negative on the cathode cap. The more negative it goes the more closer you are to Class B and the more crossover distortion. ...

More negative on the coupling cap.


The voltage on the cathode bypass cap can only ever go more-positive than the voltage at idle, unless both side of the output section shut off. But the grid current that comes when you drive the grids hard with a peak input signal greater than the bias voltage leaks out the grid and charges the grid-side of the coupling cap negative.


The fix is drive the output stage less-hard. Or have a low value resistor from grid/coupling cap to ground. Or maybe a bigger grid stopper resistor. Or a smaller-value coupling cap. Most of these fixes, though, reduce low end response, while a huge grid stopper could roll off highs by interacting with the tube's Miller capacitance (though for an output tube, it'd take a frickin'-super-huge grid stopper).

[printer2]

... when the output is clipped the bias voltage goes more negative on the cathode cap. The more negative it goes the more closer you are to Class B and the more crossover distortion. ...

More negative on the coupling cap.


The voltage on the cathode bypass cap can only ever go more-positive than the voltage at idle, unless both side of the output section shut off. But the grid current that comes when you drive the grids hard with a peak input signal greater than the bias voltage leaks out the grid and charges the grid-side of the coupling cap negative.


The fix is drive the output stage less-hard. Or have a low value resistor from grid/coupling cap to ground. Or maybe a bigger grid stopper resistor. Or a smaller-value coupling cap. Most of these fixes, though, reduce low end response, while a huge grid stopper could roll off highs by interacting with the tube's Miller capacitance (though for an output tube, it'd take a frickin'-super-huge grid stopper).

Yes, more positive on the cathode cap which results in more negative on the coupling cap. Just never said it properly.

As far as fixing the problem, depends if you actually want it fixed. A little crossover distortion is not always a bad thing.

[HotBluePlates]

You've got the right idea, and the right net sonic effect. The info below will flesh out the details, if it's handy for you.

It's like this:
A little of the current on the way from the cathode towards the plate may hit the grid instead of passing through the wide spaces between the grid wires. If you apply a peak input to the output tube bigger than the tube's bias, the grid has become more-positive than the cathode. Now some of that cathode current is being attracted by the grid instead of being repelled by its negative voltage.

Those electrons have to go somewhere. The resistor from grid to ground is that path. But that resistor's relatively high value was selected assuming the grid was negative of the cathode at all times, and that the grid current was in the pico-amp or low micro-amp range. The resistor also needs to be large-ish to provide a load easy for the phase inverter to drive.

But once the grid is positive, grid current may be a milliamp or more. Now the relatively high grid resistor value is slows the bleeding off of grid current, which physically are electrons flowing from the cathode and striking the grid. Resistances slow the movement of electrons (look up the definition of "ampere" and you'll see it's a rate of electron movement), so electrons build up on the plate of the cap attached to the grid. Electrons carry a negative charge, so that plate of the coupling cap builds up a negative charge.

So that's how the grid winds up negative. It tends to bias the tube off, until enough time has passed for the excess charge on the cap to drain away. The severity of the effect is mostly dependent on how hard you slam the output tube grid (50v peak signal to a tube with 15v of bias?), and how big the cap and resistor are (these work together to lengthen the time to drain the cap).

[dude]

OK, I used 220K Power tube grid leaks, I'll change them to 100K.

I use 100K anode load resistors, I'll change one to 82K, to unbalance.

I used a 2200uf 50v cathode bias cap (I know it's way large but I read somewhere  that a large biasing cap would increase tightness...? I'll go to 220uf.

"One change at a time" with one plate load resistor change to 82K first (other 100K), play, then try changing power tube grid leaks to 100K and finally get the 2200uf cap out to 220uf.

Sound like a plan?

I know it's not a Fender layout, more Marshall but I was hoping for a Marshall Plexi like 6V6 tone. I guess I can change to fixed bias, next if tone isn't what i want, if I'm going to do that maybe put the 325-0-325 PT in.

al 

[dude]

 But there is more to it...

I think the voltage is not the issue, but 6V6's vs. EL84's. I know everyone says that, but here's why:
6V6's are beam power tubes, EL84's are true pentodes. Beam power tubes have very little 3rd harmonic distortion compared to pentodes. And if the push-pull output stage were perfectly balanced, all even harmonic distortion would be cancelled, too.

In general, odd harmonic distortion adds an "edge" to the sound which makes it "present" or "bright" compared to low-order even harmonic distortion (and anything above 6th harmonic would be tiny regardless, so it really doesn't count here). To a limited extent, that may account for the "dark" or "brown" character that you're hearing.

Beyond that, I'm sure others can help. I'm no 18w expert.

You're right the voltage is not the issue, I think you're the expert. I unbalanced the push-pull output stage with  82K/100K plate load resistors and changed the tube grid leaks to 100K from 220K. Big difference, still the Marshall grind with the gain up, led Zep, with gain down nice and clean a lot of presence, nice blues tone and nice harmonic's very touch sensitive. I put a 5791 in the first gain stage, even better and with gain up still have the crunch.

I'm completely happy, thank you HBP, your the man. I'll shoot a few pic's and try to get a MP3 posted.

The only bad thing is the Crate VC cab is small and sounds boxy with a ten, the real tone comes from a GH1230 in a bigger cab. Looks like I might have to get the table saw ready and make a bigger 12" cab. That partical board is very heavy too, pine would be lighter.

Thanks again,
al