moving DC screen supply to PI/preamp node on vintage amps

[mxrshiver]

besides a few dB volume drop, any reason why not to? seems a minimally invasive, simple way to help power tubes survive and last long, in amps where the screens were run far past their voltage ratings and/or are held at the same voltage as the plate.

lookin at those black and silver era Champs and Princetons... in most cases, not even a bias adjustment would be needed afterwards, as i've most commonly found them idling around 18-19W on the plate if they're original, so a cooler bias would be welcome seems the only downside is changing a power supply dropper or two, to handle the extra current and still provide the same PI/preamp voltages. one could drop the whole B+ down at one point or another, but that would usually involve adding several components, more noticeably limiting volume or headroom, or some combination of the three. and i've found that vintage 6V6's can usually handle that 420V on the plate without too much trouble, as long as the screens are run much cooler (and the bias setting isn't insane).

of course, screen stoppers, lower grid resistance, lower B+ and likely a bias adjustment would be ideal to bring em fully within their max ratings. and just switching to those fabulous JJ 6V6S's would solve a lot of those problems right off the bat. but i'm trying to address situations where the owner wants to keep the amp as original and original-sounding as possible, wants to use vintage 6V6's as part of that, and wants to crank the amp... AND wants it to be reliable.

i've tried it a few times with good results, but i'm wondering if it's a practice that might sometimes result in increased noise or phase issues depending on other factors. haven't experienced that so far, and i would think not, since the screens aren't part of the audio circuit and are fully filtered, but i wanted to get some other takes on the idea. thanks in advance ✨️

[Merlin]

This is one of those things that you can probably get away with most of the time, but I can envisage situations where a sudden power chord will cause a sudden current demand at the screen grids, leading to modulation of the preamp supply node and therefore potential unwated feedback effects / oscillation.
I think a safer solution is to insert a Zener diode(s) in series with the feed to the screens to drop the voltage as much as you want, without changing the supply node.

[tubeswell]

Or you could set up a separate dedicated screen supply filter cap node (that is separate from the rest of the PS rail I.e. a separate rail) and use a voltage divider to set Vg2 to whatever you want (as long as it’s enough voltage to get the tubes to conduct)

[mxrshiver]

very sensible ideas, thanks to both of you! i hadn't thought about an oscillation or phase issue with large current demands like Merlin describes, but it makes a ton of sense. i suppose it's not only a good idea to have at least 10-30V between plate and screen nodes, but between the screen node and nodes further down as well...

the separate node seems to have no downside besides an extra cap and resistor! but i am guessing it might lead to a slightly different feel due to reduced sag and higher voltages for the preamp. great if someone's looking for a bit more clean headroom and a stiffer response, but perhaps a bit less 'original'. although the effect may be subtle to unnoticeable in amps with lots of preamp tubes and not a ton of screen current.

and the zener is a great idea, i hadn't been considering it because i'm used to using them to drop the whole B+, which leaves limited options for how much voltage you can drop with each diode because there's so much dissipation. but feeding just the screens sounds much more achievable. in terms of wattage headroom... thinking i'd give them about as much as i'd normally give power resistors. so however many watts they'd dissipate with the screens at the peak current under drive, i'd get zeners with twice that wattage rating at minimum, ideally more like 3-5 times. so if i expect 100mA peak screen current, a 20V zener should be rated 4W or better, ideally 5W. does that sound about right?

[Merlin]

thinking i'd give them about as much as i'd normally give power resistors. so however many watts they'd dissipate with the screens at the peak current under drive, i'd get zeners with twice that wattage rating at minimum,

That's certainly enough. Average current is what matters for dissipation, rather than peak current, and the average is much less. You could of course use one Zener per tube so dissipation is shared. I would think 1W Zeners would be more than enough for that.

[tubeswell]

the separate node seems to have no downside besides an extra cap and resistor! but i am guessing it might lead to a slightly different feel due to reduced sag and higher voltages for the preamp. great if someone's looking for a bit more clean headroom and a stiffer response, but perhaps a bit less 'original'. although the effect may be subtle to unnoticeable in amps with lots of preamp tubes and not a ton of screen current.

The amount of ‘Stiffness’ will depend partly on the Vg2 voltage and partly on whether the g2 supply is fully bypassed, partly bypassed or not bypassed.


In terms of the amount of voltage, a lower Vg2 will squash the grid curves down (and if it’s low enough, it will put the knee of Vg0 below the load line - making the signal more ‘linear’/ less asymmetrical. Whereas a higher Vg2 will do the opposite - but if it’s ‘too high’, it will expose the screen to over dissipation. Methinks the goal in a guitar amp would be for the knee of Vg2 to be on the loadline. YMMV


As to whether the screens are fully, partly or not bypassed, that will determine the amount of compression you get when driving the amp.