Bias question

[newguitarsmell]

How important is it that an amp run at a specific bias? If the calculator says, for example, 29 and I am running at 26.7 is that good? bigger resistor? My feeling is as long as it’s a little lower, I won’t burn out my tubes...

[shooter]

Bias is a calibration that determines Max Plate Dissipation
PP is done "different" than SE, still Pdis the goal
the closer you get to 100% the less life a tube has, but 100% is math and experience.

SE experiment, start at 75% Pdis, go til ~110%, playing each change.
PP, really get a good "book" sense because there's more tubes and 90% increase failure when you "get it wrong"

[Latole]

Short answer ; 29 or 26.7 ; you will not hear or see a difference in sound or tube life
__________________

After setting bias, I always test with playing guitar. I fine tune for the coldest bias if amp sound stay nice. Mostly class AB amp
__________________________
Bias par Spacemen Music.com:

A correctly biased amp will run efficiently and maximizes the life of your tubes.

Under - biased (hot ) amps will lack of punch and tubes will run hotter , shorter tube's life .

Overbiased amps  ( cold ) will sound thin and brittle. with tubes running too cool for proper performance .

A correctly biased amps will sound clean and tight a moderate volumes, than at higher volumes breakup and distort musically .

[shooter]

and here's the technical easy read; 

http://www.aikenamps.com/index.php/what-is-biasing

http://www.duncanamps.com/technical/lvbias.html

[pdf64]

...SE experiment, start at 75% Pdis, go til ~110%, playing each change...

The usual objective for an operating point in SE is centre bias? Too far away from that and the power output will be compromised. As long as the idle plate is below its limit, other than in regard of tube life, the particular dissipation seems immaterial. http://www.valvewizard.co.uk/se.html

...
Overbiased amps  ( cold ) will sound thin and brittle. with tubes running too cool for proper performance...

I don't think that a plate temperature based perspective is helpful in understanding how a tube amp operates, plate dissipation being just a limit.
Crossover distortion in over biased AB amps is due insufficient plate current at idle; the temperature or dissipation which results from that is secondary, an effect rather than a cause. Provided that the tubes heater voltage is suitable, the plate can't be 'too cool'. If 30mA idle plate current gets an amp out of crossover distortion when the HT is at 450V, it will do so just as well if the HT is reduced to 300V (or whatever lower level).

[tubeswell]

The more power you can squeeze out of the tube, the more 'economical' your investment. YMMV


As pdf64 says, the goal of biasing for Class A operation is to achieve centre-bias, because that is where harmonic distortion and signal clipping are minimised.  You get optimum signal output for a given amount of signal input (at a given plate voltage, screen voltage and load). Remember than in Class A operation, tube current is always 'on'. There's either more current or less current in the tube, but never no current. Therefore you want the bias to be optimal to achieve (as close as you can get to) an even voltage swing and maximum gain, with the maximum amount of tube current you can get away with. This may mean keeping the screen about 30 to 50V lower than the plate - depending on your reflected load (Too-high-a-load tends to make the screen over-dissipate under heavy signal conditions). If you are limited by a load that is on the high side, then you can either lower the screen voltage, or cool the bias off, to deprive the tube of power.


Class AB is different, because each opposite tube is alternately off for part of the signal cycle, and this changes the reflected load through the OT, resulting in the B load line (for that part of the cycle). (Whereas the A load line requires both tubes to be conducting at the same time). To get a tube to stop conducting, the signal swing at the grid has to be sufficiently largely negative so as to prevent plate current. This means that you need to drive the grid hard with a clean signal from the PI (to get a high negative grid voltage). But in the B load line condition, the screen is still conducting, so it gets zapped with even more current (when the plate current is off). So while you want the screen voltage to be high (to maximise gm and gain), you also don't want the bias to be so hot that the screen melts when the B load line condition starts to kick in


Meanwhile the 'on' tube wants to see a clean positive signal swing at the grid, so that the (quartered) B load for that tube pulls heaps of current through the OT (and ultimately to the speaker).


So the best place to bias the tubes in Class AB is where you get maximum current, and a decent B load line for part of the cycle. This means biasing the tubes cooler than you would solely for Class A (in order to encourage cutoff to occur, and to prevent the screens melting) but not so cold that you start getting overlapping cutoff in Class B. (You want overlapping Class A, not overlapping Class B.)



That's my story and I'm sticking to it.

[shooter]

usual objective for an operating point in SE is centre bias?

usual being the key word;
took me some time to "understand" what is good engineering is NOT always good musician 
bending Class A will get "a sound", depending on tube type that will horrify an audiophile, but make a guitar player smile.  Most small bottle tubes want as stated center bias, Big bottles need interaction between tech and player for "best sound" bias be damned 

[pdf64]

usual objective for an operating point in SE is centre bias?

...for "best sound" bias be damned 

We're on the same page with that 
I guess my point was that plate dissipation just isn't a relevant metric in all this; rather it's just a secondary effect that's subject to the things that actually matter being changed.
When tinkering SE outputs, it may be more useful to set it up at centre bias, then try idling hotter, then cooler, around that reference. As that will have some known, quantifiable affect on signal symmetry etc; the finding from which could be validly transferable to other amps. Which almost certainly wouldn't be the case using a particular plate dissipation as the reference point.
Hope that made sense!

[shooter]

yup 
Pdis is an easy metric to measure and try and stay below, makes 100% more sense to me than -37