Can someone please help me understand how the introduction of a screen grid between the anode and control grid reduces the Miller Effect feedback capacitance between them?

There are plenty of web resources on how to design with a tetrode and how to connect the screen grid (even in AAC own pages). I understand how to design with the screen grid and the beneficial effect it would have on extending frequency response, BUT I don't see how the screen grid reduces the unwanted capacitance between the control grid and anode, that exists in a triode valve.

So it's held at a potential similar to the anode and is shorted to earth by a capacitor - all that means is that some of the ac signal of interest never makes it to the anode (though aparently it only takes a fraction of the signal), how does that reduce the parasitic capacitance?

It's a bit like trying to reduce the collector base capacitance in a common emitter BJT circuit I suppose.

Thanks for any light you can shed on this.

 

You need to understand the significance of the Miller effect. In a triode common-cathode amplifier, the plate voltage is inverted relative to the control grid, and is Av(gain) times the input (grid-to-ground) voltage. Therefore, the grid-plate capacitance (Cgp) will have Av+1 times as much voltage across it as the actual input voltage. The source resistance (Rs) of the input will therefore "see" that capacitance and create a time constant of Rs*Cgp*(Av+1). This obviously has a detrimental effect on high frequency response.
With the introduction of the AC-grounded screen grid, the input capacitance is simply the capacitance between the two grids, because the screen grid shields the plate from the control grid.
Look up "cascode" to see ways of accomplishing the same thing with transistors. The same can be done with triodes, although I don't know if there is any advantage over using a tetrode.

 

Thanks, I think I get it now. So you mean that with the introduction of the screen grid of a specific physical geometry they can control the capacitance by setting the known parameters that control capacitance (like plate area, distance etc) - and I guess they would design these parameters to give smaller capacitance than you would otherwise get between the control grid and anode plate - and as it only takes some of the signal away most of the electrons get through the mesh and make it to the anode. That's pretty ingenious.

 

Here's another way of looking at it. The screen provides a feedback path to GROUND rather than back to the control grid.

One configuration of very high power transmitters, in fact used quite commonly in TV transmitters is the grounded screen amplifier. Certain power tetrodes are designed with a screen ring that bolts directly to the "deck" of an r.F. amplifier. A NEGATIVE screen voltage has to be supplied to the cathode, in addition to the positive voltage applied to the plate. This really makes for some "hairy" topology as far as the power supply is concerned, but it makes for a VERY stable and broadbanded amplifier.

Eric

 

Thanks, I think I get it now. So you mean that with the introduction of the screen grid of a specific physical geometry they can control the capacitance by setting the known parameters that control capacitance (like plate area, distance etc) - and I guess they would design these parameters to give smaller capacitance than you would otherwise get between the control grid and anode plate - and as it only takes some of the signal away most of the electrons get through the mesh and make it to the anode. That's pretty ingenious.

I'm still not sure you get it. Have you read the Wikipedia entries on the tetrode, Miller effect, and cascode? It's really all about putting an AC-grounded electrode between the control grid and the plate, so AC current cannot flow between them, for the reason I mentioned in my first post. Also see Eric's first sentence:

Here's another way of looking at it. The screen provides a feedback path to GROUND rather than back to the control grid.

Of course, if it's to ground, it's really not a feedback path, but let's not split hairs.

 

I'm gonna have to think about it - I thought I got it, but after making my second post, I thought it through again and my mental model had some problems, especially to do with the capacitor connected to earth. I'm going to sleep now, I'll probably get it in the morning.

 

Of course, if it's to ground, it's really not a feedback path, but let's not split hairs.[/quote]

That's the whole idea!

 

That's the whole idea!

Touché!

 

I think my understanding may be right now:

So when you introduce the screen grid you no longer have the parasitic capacitance between control grid and anode, but in theory you introduce 2 capacitances in series, between anode and screen grid and then between screen grid and control grid. So you hold the screen grid positive to keep the electrons on their way to the anode and use the capacitor short to earth because if it wasn't there some of the signal on the anode would merely propagate across the 2 capacitances back to the control grid. So using the screen grid without the capacitor short to earth is pretty useless for getting the high frequency response improvement.

 

I think my understanding may be right now:

So when you introduce the screen grid you no longer have the parasitic capacitance between control grid and anode, but in theory you introduce 2 capacitances in series, between anode and screen grid and then between screen grid and control grid. So you hold the screen grid positive to keep the electrons on their way to the anode and use the capacitor short to earth because if it wasn't there some of the signal on the anode would merely propagate across the 2 capacitances back to the control grid. So using the screen grid without the capacitor short to earth is pretty useless for getting the high frequency response improvement.

Yes. In fact, most R.F. tetrodes are so designed as to have minimum effective inductance, so that when you "ground" the screen, either "officially" or through a lot of capacitors, it has essentially 0 AC impedance to ground.

eric