Anyone connect suppressor grids to bias?

[statorvane]

I could have sworn that I read somewhere (but cant find it right now) that making the suppressor grid negative helps limit max plate current in extreme situations,

Mark, maybe you are thinking of Tone Lizard's Marshall Myths page. Something about connecting the suppressor grid to the bias voltage reduced the tendency for the screen grids to draw excessive current at high power - an attempt to increase power tube reliability.

[mhuss]

> Tone Lizard's Marshall Myths
Aha, you're probably right.

[mhuss]

The suppressor is there to improve the curves by reducing the tetrode kink. Evidently it works a little better if its potential is below, as opposed to the same as, cathode voltage.

That is true. I didn't mean to imply it had no effect, but rather that 'curve improvement' was probably not the primary inspiration for doing it (in guitar amps, anyway).

[bepone]

respond to both:
you can see how the top kink is increased, angle too, alowing a little bit more curve to be extracted in PP loadline setted to pass through the kink. so i never seen this for EL34 but why for EL34 or any other pentode would be different effect? can be traced by hand.

But what is the benefit specific to DC elevating g3, ie that can’t be equally well achieved by other means; so why bother?
The limit of anode dissipation still constrains what is achievable in the various operating classes.

anode voltage is low in that area. but you can extract little bit current more.

[martin manning]

I could have sworn that I read somewhere (but cant find it right now) that making the suppressor grid negative helps limit max plate current in extreme situations,

Mark, maybe you are thinking of Tone Lizard's Marshall Myths page. Something about connecting the suppressor grid to the bias voltage reduced the tendency for the screen grids to draw excessive current at high power - an attempt to increase power tube reliability.

Here's a better trace showing the effect of what is being proposed, i.e. connecting G3 to bias. I had to set up another isolated supply to run the suppressor at constant negative voltage (here -36V) wrt the cathode.

As with the earlier test, the main difference in curve shape between G3 to K and G3 to negative voltage is in the low Va, low Vg1 kink region, which is generally out of the typical operating area, unless load impedance is very high. The curves are a little more rounded above the kink region, as Vg1 approaches zero.
There is a difference in plate current. At the knee, say 75V Va-k, the reduction is about 30mA, or ~9%.
As far as protecting the screens, note that there is a difference in Ig2 when Va is low and as Vg1 approaches zero. At the knee, the Vg1 = 0 Ig2 is reduced ~18%.

Bottom line: The increased reliability claim seems to be true, but it comes with some reduction in output power.

[mhuss]

Nice work, Martin. Thanks for the hard data.

[beasleybodyshop]

Martin,

Thanks for posting this data - I'm curious how this would work if you connected the suppressor to the negative leg of your modified bipolar supply trick off the bias winding to derive a +/- 150vDC (approx, i forget the actual number) to drive a Post PI cathode follower driver?

I'm building a 4XEL34 amp and utilizing that driver - might give this a try for grins.

[roberto]

I think this question comes from some old amps practice. I've seen some EL34 with high plate voltages doing that.

I've done the opposite on some GU50 tubes, connecting g3 to a positive source compared to the cathode, in order to move all knees up and left, being able to squeeze more power from them.

[martin manning]

I'm curious how this would work if you connected the suppressor to the negative leg of your modified bipolar supply trick off the bias winding to derive a +/- 150vDC (approx, i forget the actual number) to drive a Post PI cathode follower driver?

I'm building a 4XEL34 amp and utilizing that driver - might give this a try for grins.

I think the -170V might be a bit much, and the DC bias on the cathodes of the drivers is not decoupled. I would add a simple conventional bias supply coming directly from the bias tap and use that for the negative voltage.

[FUCHSAUDIO]

I'm curious how this would work if you connected the suppressor to the negative leg of your modified bipolar supply trick off the bias winding to derive a +/- 150vDC (approx, i forget the actual number) to drive a Post PI cathode follower driver?

I'm building a 4XEL34 amp and utilizing that driver - might give this a try for grins.

I think the -170V might be a bit much, and the DC bias on the cathodes of the drivers is not decoupled. I would add a simple conventional bias supply coming directly from the bias tap and use that for the negative voltage.

Agreed, as there are no "stats" for this (although your plate curves are terrific Martin !), I can't imagine anything more than the 60-80-ish voltage of a raw bias supply being needed. Whether you are driving off a DC CF our coupling caps really shouldn't matter imho. Again, taking a pot and putting the wiper on these grids and grounding one side and taking the other side to the bias supply could allow you to see what the voltage range does. I'm thinking you hit a practical "wall" where it stops doing something or gets to some kind of cutoff point where more voltage doesn't do anything or causes trouble !! lol.

[imjonwain]

Thanks for the data and charts! Cool to see some of these type of things validated!