Adjusting pentode tube screen voltage?

[patricks]

Hi all,

I'm in the process of mulling over my first amp build, a Vox AC4 clone. 3 tubes (I don't plan on including the tremolo part of the circuit), a minimum of components and a pretty simple layout, plus it's been done a few times before with success (google "diy vox ac4" and there are about 3 decently written build threads).

So, getting to the point, my question is: is it OK to adjust the screen voltage of a pentode tube using a potentiometer?
In a build thread on the Telecaster guitar forum, the builder changed the 5.6M resistor that supplies the screen to a 1M resistor: http://www.tdpri.com/forum/shock-brothers-diy-amps/322998-vox-ac4-build-4.html, 4th post from the top) and said it gave the amp more "balls". After reading an article on ampbooks.com (http://www.ampbooks.com/home/amp-technology/pentode-gain/), I've learned that this allows higher voltage gain, but reduces headroom.

I'm thinking of putting a 5M pot in series with the 1M resistor to control screen voltage, but I want to make sure that adjusting this on the fly (while the amp's on and warmed up) wouldn't damage the tube.
One last question - would a log or linear pot be more appropriate? I'm thinking log...

Cheers
Pat

[gritz]

It depends how comfortable you are with the idea of attaching ~250VDC to a potentiometer - a lump of metal that may only be rated for e.g. 150V DC.

It also depends on whether you envisage adjusting the parameter every time you twiddle the controls on the amp, or whether it's an "adjust it once and then leave it alone" deal. It also depends on how much store you set on anonymous internet folks who claim to replace a random part and then tell the world "zomg!!!! teh awesome!!!!!"

It sounds like a candidate for (carefully) prototyping on tagstrip, but I'd lean towards using fixed resistors rather than pots or trimmers with high voltages. If you find that it's worth it then there may be safer ways to control that grid, if need be. But be careful!

Some ef86 pentode whatnot: http://www.valvewizard.co.uk/EF86.html

[patricks]

Thanks for the reply! 

I checked on Mouser and the 5M linear taper pot (no audio taper 5M pots available) is rated at 500V DC, so that's all good.
I'm not sure whether it'd be a one-off adjustment, or something that I'd like to fiddle with. As long as it won't damage the tube, I could always wire it up temporarily and see what effect different resistances have on the sound. Good point about trusting claims off the 'net, the pot would let me test it out, though.

I was planning on using tagstrips for the build, so that'll make the testing easier. Thanks for the link to the valvewizard site and the words of caution, I promise I'll be careful!

[Striker Amplification]

to adjust screen grid voltage through a potentiometer, you need to first couple the pot with a decent sized resistor the resistor being 50% smaller than the original set bias resistor for the screen grid, then you can sweep through voltage schemes, but with respect to your cathode, as long as the cathode is biased correctly, the screen grid will stay at an even current idle until its changed, if there is too much negative current flow from the cathode, the screen grid will in fact overheat, and melt.
                                                                               hope this helps! Cheers, Ryan.

[patricks]

Thanks, Ryan!

So, in this schematic, am I right in thinking that the bias resistor for the screen grid is R8 (5.6M)?


The build thread in the Telecaster forum says that they swapped this for a 1M resistor. You said that the pot should be coupled with a resistor 50% smaller than the original, but since the circuit works with either a 1M or a 5.6M in that location, would I be OK to couple the 5M pot to a 1M resistor, or should I couple it with a smaller resistor?

The only thing that keeps nagging at me is that I can't find any examples where it's been done before. If varying the screen bias with the tube warm was a good idea, someone would have tried it already. Still, I suppose that as long as I don't drop the resistance too low, I shouldn't overheat the screen.

Cheers
Pat

[PRR]

> If varying the screen bias with the tube warm was a good idea, someone would have tried it already

It's done a lot, in *other* and obscure fields.

If you are building a "good" amp, and you know what you are doing, you have a good idea of Vg2 early in design. If you don't fully grok pentodes (it aint easy), then you use the values given in the "RC Amplifier" tables in the tube data-- some 1950s white-shirt junior engineer sweated over a hot breadboard to make it *easy* for you to use his company's tubes.

Also: there's a thump if you vary Vg2 rapidly.

No, you won't kill the tube. All tube electrodes can be wobbled millions/second and the tube don't care, the electrons do whatever they do.

FWIW: 5Meg on G2 with 220K on plate is a rather extreme ratio (for most small pentodes). G2 must sit very low. Gain is as high as it gets.

1Meg in this circuit is a more book-value Rg2. This also gives good gain. (Not sure how this is "more balls".)

I said it thumps. If you go a very wide range of G2 bias, without changing the cathode resistor, tube current tries to go high, the 220K Rp drops the plate way low, and the tube won't amplify. No harm done, but don't expect a huge range of useful operation without re-design.

The Tremolo brings in another complication. It wobbles the cathode voltage. But large range of Vg2 causes change of cathode voltage. Tremolo amplitude has been adjusted for the stock values; with very different conditions the trem will be too-much or too-little.

For your experimentation, education, amusement:



Yes the pot is insulated for a few hundred volts. I would use a plastic knob just in case.

I've drawn the pot ends "upside down" or "swopped" because going to ground will *increase* gain (up to a point), the control action is "backward" to a regular volume pot.

Lin or Audio must be a matter of experimentation.

[Striker Amplification]

WOAH!   I missunderstood, i was led to think we were in the POWER TUBE category, after all i did have nothing to go on, and its not my fault for not thinking someone actually had a modern pentode driven amplifier schematic, just when i thought i was the only one who used sub mini pentodes!  

[PRR]

> i did have nothing to go on

First post says Vox AC4 clone.

VOX AC4 is a small odd duck. It was an extra-cheap practice/small-gig amp. Reminiscent of the first Champ, in that it is a power-tube plus a one-stage preamp/driver. To get enough gain in one stage means a pentode.

There isn't a lot of difference in cost between one pentode and a twin-triode. And a 2-stage preamp/driver is much more accommodating in gain-structure. (EL84 output is better about this than Fender's 6V6.) So this plan bespeaks a very-very-low price-point.

But then the VOX AC4 throws in a Tremolo. And uses a full twin-triode to do this! So it's not the cheapest amp on the block.

The TDPRI drawing is a nearly exact copy of the VOX circuit (the trem switch is on the wrong node which defeats the kickstart). The odd 5Meg G2 resistor _is_ what the Old Guys used, and these heavy tube-heads knew their stuff.

For cross-check, see EF86 data/sale sheet. Page 2 has the RC Amp tables. Lower table for Ra=220K applies. They suggest a 1Meg Rg2, 2.2K Rk. At 250V they got gain of 180(!) and 46Vrms max output (0.25V input overload). Ik of 0.9mA and 6:1 division ratio suggests Vg2 is near 100V. The higher Rg2 should lead to lower Vg2, higher gain, lower Ik (partly compensated by a lower 1.5K Rk), and a lower input overload. I believe that hotter axes run full-up will frequently overload this stage.

After that first Champ, Fender moved to twin-triode inputs, and 99% of the market followed. Recently there is a resurgence of interest in pentode preamps, perhaps because they are "different".

Hey! VOX AC10 has variable G2 voltage on the FX channel preamp, and it adjusts "trem depth".

[patricks]

@Ryan - apologies, I should have posted the schematic first up and specified I was talking about an EF86 pentode.

@PRR - thanks for all the info! Those links are great - hey, yeah! Nice to see that the AC10 has a variable Rg2 voltage on an EF86. Now I know I'm not quite as crazy as I thought 
The reason that the author of the build thread on TDPRI gave for swapping Rg2 was that the AC15 uses a 1M resistor in that spot. It does (http://www.el34world.com/charts/Schematics/files/Vox/ac151960.pdf), but some of the other resistor values aren't textbook compared to the spec sheet, either (Rk and Rg1 aren't used in the AC15 and some of the capacitor calues are different, too).

At least I know I'll be safe when Vg2 is between 1M and 5M, so I might try a 2M pot in series with the 1M and see how it sounds at various points around the dial
I'm using a G&L Tribute Comanche; the MFD z-coil pickups aren't quite as hot as regular humbuckers, but they are hotter than regular single coils, so I might end up being better off with a lower Rg2 (lower than 5M, anyway) for a little less gain and more headroom. But I'll experiment, anyway!

[Striker Amplification]

If you want Patricks i can design bout 5 simple single ended tube amp designs and send them to you??

[patricks]

Wow, thanks Ryan, that'd be sweet!
What can I do in return? My main job is writing/reviewing/editing for a scientific journal, so if you need any technical material written up for your website/blog/facebook page then that might be a good skill swap

[teemuk]

If varying the screen bias with the tube warm was a good idea, someone would have tried it already

It has not only been tried but also patented by several companies (e.g. Claret, Trentino, Blackstar, Marshall, etc.)

Screen voltage modulated tremolos excluded, the earliest examples of applications still date back 30+ years.

So, getting to the point, my question is: is it OK to adjust the screen voltage of a pentode tube using a potentiometer?

No.

Use an active voltage regulator circuit for the dirty work and put the potentiometer where it belongs; to its low voltage base, grid, gate or whatever circuit. You don't want to put high voltage DC across a generic potentiometer. If you really must adjust screen voltage so simply then buy one of those high power rheostats. Don't use a generic potentiometer. It will fail sooner or later, probably somewhat catastrophically.

There are several good reasons why you pretty much always see such stuff done with something like a MOSFET voltage regulator instead of a simple resistive divider with one resistor being adjustable.

[patricks]

It has not only been tried but also patented by several companies (e.g. Claret, Trentino, Blackstar, Marshall, etc.)

Cool, thanks - good to know!

Use an active voltage regulator circuit for the dirty work and put the potentiometer where it belongs; to its low voltage base, grid, gate or whatever circuit. You don't want to put high voltage DC across a generic potentiometer. If you really must adjust screen voltage so simply then buy one of those high power rheostats. Don't use a generic potentiometer. It will fail sooner or later, probably somewhat catastrophically.

Ah, thanks for that advice, I'll look into voltage regulator circuits. Out of curiosity, in the AC4 schematic, the voltage at the anode of the EF86 is listed as 215V - do the volume and tone pots have to deal with this voltage or will there be a sufficient voltage drop across C6?

[tubegeek]

C6 will block the DC voltage and the pots will be referenced to ground on one end, so they will sit at 0 V DC with only the signal AC on them.

Remember: caps block DC and pass AC, inductors are their evil twin, and pass DC while opposing AC. The frequency where "block" crosses over to "pass" depends on the value of the cap or inductor.

C6 will therefore see a difference of 215 V from one end to the next, so you can easily see how it must be rated for at least that voltage. In actuality, it must be rated for the peak power supply voltage - because it will have the full power supply voltage across it before the pentode starts to conduct and brings the voltage at the plate end of the plate resistor down.

The phrase "voltage drop" almost always refers to a DC voltage drop caused by current flowing through a resistor. Ohm's law would be the applicable statute there.

-j

[patricks]

Remember: caps block DC and pass AC, inductors are their evil twin, and pass DC while opposing AC. The frequency where "block" crosses over to "pass" depends on the value of the cap or inductor.

Ah, of course - now that you mention it, the article did talk about that, but I'd forgotten 
Thanks very much!

[PRR]

> AC15 uses a 1M resistor

It also has a whole extra stage between input and power stage, so is not as desperate for gain.

re teemuk's warnings on pots: in the old days we ran serious voltage on pots. See any old oscilloscope. However pots have gotten cheaper. The plan I proposed will work, but.... I bet half the range of the pot just slams the pentode out of bounds (you could shim the ends of the pot with a few hundred K to limit the range; and I probably would not use the common dollar-pots for anything I took out of the workshop, I'd get the "mil-spec 2 Watt" pots that cost $15-$30 each.

[patricks]

It also has a whole extra stage between input and power stage, so is not as desperate for gain.

Ah, right. I remember reading in the build thread that the author described the sound of the amp as "wooly" with the 5.6M resistor for Rg2, so maybe all that gain meant that it was overloaded.
Thanks for the tip about the pots, I checked on the Mouser site and they have some that should fit the bill. I'm reading up on active voltage regulators, too, although most of what I've found surpasses my limited knowledge so far 

[tubegeek]

Thanks very much!

Well, aren't you polite! You DO realize this is the internet, right?

You're quite welcome. I wouldn't know jack s**t if it weren't for RG sorting me out anyhow.

Whippersnappers: he began that work almost 25 years ago now. Damn.

And he's not nearly done yet...

So - I'd sure as hell BETTER pass on whatever he's managed to teach me!

-j

[patricks]

Well, aren't you polite! You DO realize this is the internet, right? Wink

Ha ha! I can always start trolling or BUMP this thread if no-one responds within 5 minutes

I've checked out RG's site a few times for various electronics/fx things, lots of good knowledge.

[DougH]

Here's an idea that might help. If you are trying to control compression on the EF86 by varying screen voltage- this is a way of doing it with low voltage:

http://www.ax84.com/static/corepreamps/Blues/AX84_Blues_Preamp_Schematic.pdf

(Check the "Squish" control on page 2.)

This is the ax84 "blues preamp". It uses a 7199 pentode for the preamp. Don't know (remember) the reason they chose this over the EF86, it may have to do with the effectiveness of this kind of control, so YMMV. But you basically want to stick your variable resistor on the other side of the screen filter cap, between it and ground. HTH


Edit:

You may want to adjust R8 accordingly to compensate for the added series resistance due to the pot.