Yet another output power thread

[Dave H]

With sine wave input , 500 Hz , at onset of clipping the screen voltage dropped to 95V . Is this normal ? Spec sheet suggests 5ma increase in screen current time two tubes through 22k dropper would explain this . Need the dropper to get below max screen voltage . Is there any way around this ?

You could try replacing the 22k with ~40V zener

[Helmholtz]

50L6 with 4k plate to plate load .

The loadline construction shown for class AB looks correct (even shows the class A kink) .

Means an optimum Raa of 4k with Vak = 160V and 110V screen voltage both measured at full output.
The theoretical output would be around 6W.

But B+ and screen voltage drop at full output, both pushing the knee of the Vgk = 0 curve down to lower currents.
So the actual loadline needs to be flatter, meaning a higher optimum Raa.

Lower screen and plate voltages strongly lower available output power.

Also the cathode voltage (measured at full output) lowers the available Vak, thus reducing output power by maybe 2W.
Fixed bias would avoid that effect.

[loudthud]

At idle, guitar amps typically run screen grids at higher DC than the data book maximum in push pull class AB power amps. That's not what stresses the screen grids. At maximum signal, an unbypassed resistor drops the screen Voltage to a safe level, but you need to look at screen Voltage with a scope to see it.

If you look through Fender tweed schematics, you can see the evolution of screen grid circuits.

[Helmholtz]

BTW, the cathode voltage also lowers effective screen voltage.

[35L6]

So it turns out the problem wasn't the screen voltage dropping too low under signal but rising too high at rest . Once I biased it hotter it had enough current to keep it low at idle . I put in a 100R cathode resistor and it biased at 8.5V . That gave me enough screen current to put in a 5k6 screen dropper . Screen voltage at idle is 122V and 105V at onset of clipping . All the other voltages dropped a little too . Measured directly , the voltage from anode to cathode was 135V . With the 4k tap I got 5.3V at the speaker out for 3.5 Watt . That takes me from " can you please turn that down " to " turn that damn thing down " . Its loud in a small room . Blow the roof off a doll house !

[35L6]

[35L6]

Schems with new component values and voltages . Notice cathode bias . Power supply is totally different . I will draw a complete schematic at some point . I like the zener idea . It also occurred to me that Leo was right to set the screens high because they drop under load . This was definitely a case of getting experience right after I needed it . And of course the Hi input was bright because I forgot the 100p from plate to cathode . Put it in and it sounds better . I might go for 150 !

[35L6]

Full schematic , hope you can read it . If not let me know . Not my artwork , I traced it . Don't know who drew it , but I want to thank them anonymously . Non-standard values my own . Standard values by Marshall of course .

[35L6]

I got some zeners ,1N5938 , 2W 36V . Install in series with the screen supply replacing the 5k6 resistor . I'm only dropping 24V across it . It seems that voltage changes with current . Should I have gotten .5 W zeners ? The screen voltage comes down to 124V at onset of clipping but I get no more power than with the resistor . Perhaps screen voltage sag wasn't the limiting factor ?

[Helmholtz]

I got some zeners ,1N5938 , 2W 36V . Install in series with the screen supply replacing the 5k6 resistor . I'm only dropping 24V across it . It seems that voltage changes with current . Should I have gotten .5 W zeners ? The screen voltage comes down to 124V at onset of clipping but I get no more power than with the resistor . Perhaps screen voltage sag wasn't the limiting factor ?

I guess a zener doesn't make sense anymore.
Also a series resistor improves ripple filtering while a zener doesn't.

What is the voltage drop across the 5.6k at idle and at full power?

[Enzo]

Yes, a zener needs a certain minimum current through it before it will "zene" accurately.

[Helmholtz]

Yes, a zener needs a certain minimum current through it before it will "zene" accurately.

True, but a quality 1N5938 should drop around 35V even if the current is only 1mA.
https://www.mouser.de/datasheet/2/30..._D-2309189.pdf

[Enzo]

Do we know the source of his zeners?

[mozz]

People use zeners on the power transformer CT to drop voltage, they never drop what they are rated at.

[Helmholtz]

People use zeners on the power transformer CT to drop voltage, they never drop what they are rated at.

You can't expect a zener to drop DC output by exactly the zener voltage when used with pulsed currents on the AC side.
Nevertheless results should be repeatable.