General Understanding of screen vs plate voltage reading interpretations

[jojokeo]

***My 3 main questions here are underlined below***
I made an amp a while back that can run either cathode or fixed bias. I haven't switched to cathode bias in a while but noticed a slight red plating while in this mode. We're talking 6V6s & a standard 250r cathode resistor used - "just because it's standard for this combination." My measurements came out to be at ~14.11 watts pDiss. The original cathode resistor measured 246r so I raised it to 279r which dropped pDiss down to 13.5w and no more red-plating. Another lesson learned to not simply (blindly) use standard (normal) values for things like this.

In performing the above measurements my voltages are: node A = 362, node B = 360
Plate to Ground = 348, but plate to cathode resistor = 325v SO the tubes actually "see" plate to cathode resistor and not to ground. Correct?
Here's leading to my main question regarding Screen voltages and what the tubes "see":
Using 1K screen resistors (actually meas - 983r) Ground to resistor = 358v
Ground to screen grid = 354v
BUT:
if I measure Cathode to Screen resistor = 336v, & Cathode to Screen Grid = 333v : IS this what the tube actually "sees" as the plate voltage does above OR does is see the voltage as it's measured to ground????

Also, the tube data sheet says max Screen Voltage should be 315v. I assume it's generally okay to have the screen voltages higher than their max ratings as long as Screed Dissipation is not exceeded???

[tubeswell]

The tube 'sees' plate-voltage-to-cathode-voltage in terms of measuring plate dissipation. If your screens are higher than the plate voltage, the amount of screen current will get progressively more out of kilter with the plate current as the signal gets bigger, so you probably should drop the screen voltage a little

[jojokeo]

So to keep the math simple if I drop 4 volts using 1K Screen resistor, then I could drop 40 volts using a 10K. But, would the 10K limit too much current for a screen resistor or it's ability to draw more than 4mA through it?

As an easy remedy would it be best to simply attach a 10K resistor at Node B (after the choke) and feed this lowered voltage to the two 1K screen resistors feeding the two 6V6s to get them to the lowered voltage wanted?

[tubeswell]

Try 4k7 for the plate-to-screen node dropper. If it works out, you could have a switchable 1k5 in parallel with it, which switches in when you switch to fixed bias- if you want to preserve the fixed bias mojo. These figures are just guesses I'm bandying about, based on how 5E3 and AA1164 circuits compare (since you mentioned 6V6s)

[PRR]

> okay to have the screen voltages higher than their max ratings

Technically, no. If life is short, and you violate any rating, you can't yell at Cunningham for making bad tubes.

Of course you can't yell at anybody who makes tubes in this world.

I would not be the least bit worried about 333V on a 315V tube.

I'd be startled if any *modern* "6V6" showed color at 14.11W. That's not even over the old rating (less than 1%, and we didn't have 1% meters back in the day). And Fender abused so many 6V6 that all modern "6V6" are beefier than the old spec. I wonder if you mathed it wrong.

In general, on self-bias amps, you really should reduce the B+ to the OT rather than jack the cathode resistor and starve the tube.

[jojokeo]

Of course you can't yell at anybody who makes tubes in this world.

I would not be the least bit worried about 333V on a 315V tube.

technically speaking for the screen grid voltage measurement & reading - what does the tube "see": Ground to screen grid OR Cathode to screen grid?

[tubeswell]

The screen grid dissipation is the screen-to-cathode voltage x the screen current

[jojokeo]

The screen grid dissipation is the screen-to-cathode voltage x the screen current

Thanks for confirming this for me. I've been racking my brain on it for some reason because of a biasing spreadsheet I came across and have been working on which needed correcting. All current passes through cathode resistor and it only makes sense since plate current does, so would screen current. Same for voltage readings on both plate and screen too.

The amp has loads of power early on the dial like a Fender BFDR but - at extreme full power w/ all gain & vol settings it does have an EMF type noise created. It could be attributed to what you're saying about screen/plate relationship OR the trannies are too closely mounted together? Either way the amp's overall volume is such that this extreme setting isn't needed anyway but as always we strive for perfection. This was one of my earliest amp builds so not overly surprising considering.

[tubeswell]

The noise when its overdriven could be anything. It could be the transformer you've used.

[jojokeo]

It's interesting in that it definitely achieves it's total output power while being very quiet up through to that point mentioned whether or not the extra preamp overdrive stage is activated. It's as if it has an excess of power/voltage level like going to 110% or 120%.

[tubeswell]

It could be PI distortion.

[jojokeo]

It does this with or w/out any signal applied. With all the volumes & gains up and the master on about 8 out of 10 or 3 o'clock if you're thinking of chicken head knobs, the overall powerful EMF starts to wake up from being quiet. All full up, it sounds like some serious overhead power lines carrying a few thousand volts type of sound. It's not super loud but it gets your attention. If you were troubleshooting it while like this, you'd think twice before you touch something and if you made a wrong move it'd kill you. But it does play fine in this condition w/ the normal associated distortion. I just don't turn everything up like this as there no need to. I would like to confirm the how, what, & why this occurs but it's not a necessity.

[tubeswell]

It could be excessively-lengthy signal wires coupling with something (How about I just keep on suggesting things and you test them out?  )

[jojokeo]

  okay but that's not it either I can't imagine. I always reverse the layout design of my LTPI components in keeping w/ Gerald Weber's golden "lead dress, layout, and design rules". He does have some good experience & recommendations on things that I do trust such as keeping low signal high Z grid wires short as possible & high signal low Z plate wires longer. The typical average layout has this exactly opposite of how it should be done in complying with this. I know many amps are fine the "standard" way but an ounce of prevention kind of thing....ya know!? Here's a photo of things with the increased cathode resistors in series and the short grid wires to the power tubes. Hell lets blame it on the cap cans? 
I still think my trannies might be too close to each other as I didn't know to use the listening trick back then and just placed them to fit the situation. Nowadays my OT is on the other side of the chassis.
I'm thinking the unshielded OT could be getting slammed by the PT at max power the way the laminations face directly at it. In this situation a protected/shielded OT would've been the better call here. See 2nd photo of what I'm describing.

But at the end of the day, the amp sounds awesome and it's not used w/ everything on max anyway so it's a non issue in that regard. It just bugs me because I don't know exactly what causes it.

[Willabe]

I know many amps are fine the "standard" way but an ounce of prevention kind of thing

OK, so, with that, shot in the dark, surf'in the wave, I'd still twist these 2 yellow wires, with a PVC or what ever modern wire insulator, not heavy cloth cover.

Per KOC, twisted pairs where ever you can. Plate send/feed from PI to power tubes, he has said doesn't need, many/most times, shielded cable, just a twisted pair.    


             Brad