Is there a minimum screen voltage?

[Jonas]

Speaking in general terms, there are a couple different was to wire up a power pentode
1) Connect the screen directly to the plate gives = equal voltage on screen and plate (pseudo triode connection)
2) connect the screen to an ultra linear tap reducing voltage at the screen compared to the plate (UL connection)

so what happens if the screen voltage is zero VDC?

I am in the process if experimenting with a VVR on my SE KT88 amp and I am getting some interesting and pleasing tones when lowering the screen voltage.

[Willabe]

Kevin O'Connor has/sells a Power Scaling circuit (VVR) for the screens of very large amps. It's hooked up to the power tubes screens, not it's plates.

Look it up, London Power amp web site.

But it's only for taken the volume down some on a huge amp. Wouldn't make sense to lug around a 200w amp and try and turn it down to 10w. 

But you can limit an amps out put by lowering the screen dcv. How far can you go? 

[tubeswell]

If a screen grid resistor is burnt open (i.e. no screen voltage), the tube won't conduct.

Once you get past that matter (i.e. once the screens are at a sufficient voltage to makes the tubes conduct), the effect of screen voltage on tetrode/pentode conduction is more significant than the bias voltage. However, the effect of screen voltage on tube current does vary (somewhat) in proportion to the (control grid) bias voltage.

But having noted all that, I've noticed through various playing around with screen voltages over the years that most output tetrodes/pentodes used in typical guitar amps won't conduct until you pump the screens up to about 150V-180V. YMMV

[HotBluePlates]

... I am getting some interesting and pleasing tones when lowering the screen voltage.

Lowering screen voltage reduces the plate current the tube can deliver.

so what happens if the screen voltage is zero VDC?

We would expect Zero Plate Current (or pretty close).

... the effect of screen voltage on tetrode/pentode conduction is more significant than the bias voltage. ...

Reconsider that.

   - 6L6GC Data sheet

   - Page 7 has curves for 400v on the screen.
   - A 20v change on G1 from the "-20v line" to the "0v line" at 400v plate results in a plate current increase of ~200mA.

   - Page 6 has the curves for "G1 = 0v" and different values of screen voltage.
   - It takes about a 200v change of screen volts to get the same 200mA change of plate current.

   - It would seem that G1 is 200v / 20v = 10 times more effective at changing plate current than G2.
   - The data sheet will usually state this as "Mu g1-g2" or "Triode Amplification Factor."
   - The 6L6GC data sheet gives a triode amplification factor of "8" though it would appear that depends on the specific conditions we consider.

[pdf64]

...
But having noted all that, I've noticed through various playing around with screen voltages over the years that most output tetrodes/pentodes used in typical guitar amps won't conduct until you pump the screens up to about 150V-180V. YMMV

It'll depend on the bias voltage.
It can be expected that a bias appropriate with the screen grids up at 400V may be close to cut off when the screen grid voltage is halved.
The bias need to be rescaled for the screen grid voltage.

[Merlin]

so what happens if the screen voltage is zero VDC?

Lowering the screen voltage progressively 'squashes' down the grid curves, so with 0V they would be squashed down to nothing: no conduction.
This image shows the curves of an EF86 as the screen voltage is lowered -all pentodes behave like this.

[stratomaster]

I believe reduction of screen voltage is part of how Marshall achieved the lower power modes of the Origin series amps.

It loses punch and dynamic in the process. Lowest power mode sounds like poo.

[tubeswell]

Reconsider that.

Good catch! I should've said 'with no screen voltage, the tube won't conduct, no matter what the bias voltage is, but once there is sufficient screen voltage, the bias voltage (obviously) affects tube current' (or words to that effect).

[HotBluePlates]

Good catch! I should've said 'with no screen voltage, the tube won't conduct, no matter what the bias voltage is ...

What's funny is I knew you knew how screen volts & bias volts interact, but other folks reading it would misunderstand.  Glad we got there in the end!

[Jonas]

thanks for all the helpful information on this one!

[AlNewman]

Lowering the screen voltage progressively 'squashes' down the grid curves, so with 0V they would be squashed down to nothing: no conduction.
This image shows the curves of an EF86 as the screen voltage is lowered -all pentodes behave like this.

Hi Merlin,
What program did you use to get that data?
Thanks.

[Willabe]

Somebody did that on a scope, doesn't look like a computer program.

[Merlin]

What program did you use to get that data?

That was done with my minimalist curve tracer:https://valvewizard.co.uk/curvetracer.html