Pentodes as grounded-grid RF amps?

[Neil Breward]

One for the engineers!

OK. We know that grounded-grid triodes were popular as the RF amplifier in VHF/FM circuits, as were cascode double-triodes in VHF TV tuners. Were RF pentodes like the EF80 or EF91 ever used in grounded-grid mode as RF amplifiers, and if so, what were the advantages and/or disadvantages of these as compared with triodes? And that's pentodes in 'real' form, not as strapped triodes!

Seems to me that in grounded-grid mode, with control grid earthed, screen grid earthed with respect to RF, and supressor grid also earthed rather than connected to cathode, the screening between anode and cathode would be extemely good, better than a triode, and the higher gain and output impedance of the pentode would also be an advantage. Any additional anode-to ground capacitance would be part of the tuned O/P circuit. Am I missing something - and not just the economics of using an ECC85 as RF and Mixer/oscillator? Was it the higher noise factor of pentodes that negated this approach?

I'd be grateful for your comments.

Cheers,
Neil

[MichaelR]

I don't think this is that easy to answer in general terms. In RF stages triodes are often used because although they suffer "miller effect" with the inter electrode capacitance and consequentially become potentially unstable unless neutralised, pentodes are physiclly much bigger as valves due to the extra structure. This makes them potentially more unstable at high RF frequency and they are noisier. They are often used at IF relatively low rf frequencies but for RF amplifiers which can be very wideband there are problems I beleive.

[Duke_Nukem]

When a triode is in grounded-grid connection, miller effect is pretty much dealt with and therfore there is no fundamental advantage to be gained from using a pentode instead. However, pentodes are inherently noisier devices.

TTFN,
Jon

[Neil Breward]

Jon,

I know it's not the Miller Effect that we're talking about here, which is minimised in grounded-grid mode whether its a pentode or triode. Does the increased noise factor totally offset any increased gain from the use of a pentode?

Cheers,
Neil

[ppppenguin]

I doubt that you would get any extra gain from a pentode (as compared to a triode) when run in grounded grid. I can't immediately find any references to support my claim but this website is interesting:

http://wb0nni.dakotamade.com/ggbasic.html

[MichaelR]

Originally Posted by Neil Breward

Does the increased noise factor totally offset any increased gain from the use of a pentode?

Cheers,
Neil

If you are looking at frequencies above 50 mhz , forget pentodes. As I said earlier it is difficult to make a general statement. The increased Noise factor will be a problem if you are looking for a low noise Rf and broadband amplifier.

mike

[kalee20]

I've been thinking about the valve connection. Usually, for 'true pentode' operation, the screen and suppressor grids are considered to be as the same potential as the cathode. So, for true pentode operation in grounded grid, the input signal will be simultaneously applied to the cathode, screen and suppressor grids (because these are decoupled to cathode).

Taking the suppressor to ground probably won't make much difference, because it's generally a coarse winding.

Decoupling the screen grid to ground will probably improve performance, because the grid will no longer be associated with the input circuit and hence capacitance from output to input will be that much less (the anode-screen grid capacitance now appearing purely as a capacitance to ground). But, for the pedants, is this true pentode operation?

For triodes, the advantage of grounded grid operation, is that Miller effect is removed. But, there's still the damping on the output circuit given by the ra of the valve. In grounded cathode operation, the output resistance is the ra of the valve; in grounded grid, it is the ra augmented by a factor dependent on the impedance of the input circuit.

So, I think that if the source impedance is high, there's nothing to be gained from using a pentode grounded-grid. Such a case might be the second valve in a cascode pair - this operates as grounded-grid, with a high-impedance feed provided by the first valve.

However, if you have a low source impedance, a grounded-grid pentode will potentially give more gain than a grounded grid triode. But in this case, why not use the pentode in grounded-cathode mode and have the best of both worlds?

[MichaelR]

Originally Posted by kalee20

So, I think that if the source impedance is high, there's nothing to be gained from using a pentode grounded-grid. Such a case might be the second valve in a cascode pair - this operates as grounded-grid, with a high-impedance feed provided by the first valve.

However, if you have a low source impedance, a grounded-grid pentode will potentially give more gain than a grounded grid triode. But in this case, why not use the pentode in grounded-cathode mode and have the best of both worlds?

Please excuse my possible understanding of this, but surely grounded grid is a low impedance input , this is ok in RF usually since you can feed from the previous stage using a transformer , I suppose it could be a problem at AF though.

Mike

[ppppenguin]

Originally Posted by MichaelR

Please excuse my possible understanding of this, but surely grounded grid is a low impedance input.....

Absolutely. There are 3 ways I can think of to use this feature.

One is in a cascode where you feed the cathode from something close to a current source which has a high impedance. The 2nd is for RF or similar applications where you can achieve a good match to a transmission line at 50R or similar. The 3rd is the long tail pair where one half is often effectively in grounded grid mode.

[kalee20]

Yes, a grounded grid amplifier does present a low impedance load to whatever is driving it.

But, the source itself can have a high impedance.

As an example, in a cascode circuit, you have two valves - the first is grounded cathode, the second is grounded grid. So, the first stage has a high-ish output impedance (tens of kilohms) and this source drives the grounded grid stage (which presents a load on the previous stage of perhaps just a hundred ohms). And as we're talking about grounded grid, there's an example where it's fed from a really high source impedance! Not a transformer in sight! (And it works well at AF too).

[Neil Breward]

Originally Posted by kalee20

As an example, in a cascode circuit, you have two valves - the first is grounded cathode, the second is grounded grid. So, the first stage has a high-ish output impedance (tens of kilohms) and this source drives the grounded grid stage (which presents a load on the previous stage of perhaps just a hundred ohms). And as we're talking about grounded grid, there's an example where it's fed from a really high source impedance! Not a transformer in sight! (And it works well at AF too).

Yes indeed, but the first triode in a cascode arrangement provides very little gain as its load is the input impedance of the second triode - a few hundred ohms - hence the low Miller effect. Virtually all the gain comes from the second, grounded-grid triode.

Back to pentodes. Yes, pentodes were rarely much used in applications above 50 megHz, though whether this was due primarily to the increased noise factor or due to the falling input impedance and cathode inductance/transit time problems isn't clear. There were exceptions, such as some early FM sets (Pye FenMan series?) but the ECC85 -type tuner soon took over.

In grounded-grid mode, which offers a similar low input impedance with either triode or pentode, isn't the problem of falling input impedance with frequency typical of grounded-cathode operation eliminated? Matching to a low source impedance such as coax feeder isn't a problem. Output impedance for a pentode - especially in grounded-grid - must be very high, and would minimise damping on the tuned circuit.

Whatever the problems are, clearly the g-g pentode wasn't a popular circuit. Unless anyone has actually tried this, and has the data to prove it, can we assume that the noise problems outweighed any advantages?

Thanks for the comments!
Cheers,
Neil

[ppppenguin]

Originally Posted by Neil Breward

Yes indeed, but the first triode in a cascode arrangement provides very little gain as its load is the input impedance of the second triode - a few hundred ohms - hence the low Miller effect. Virtually all the gain comes from the second, grounded-grid triode.

The first triode in a cascode provides current gain, the second one gives voltage gain. Thus between them they provide good power gain.

[GMB]

I've just checked over a number of Air Ministry VHF and UHF receivers and I have to say that they all use a pentode front end albeit never in grounded grid mode.

Of course this may be because the only VHF-capable valve of the day was a pentode (usually EF50)?
Interesting that it was delevoped that way.

[MichaelR]

Originally Posted by GMB

I've just checked over a number of Air Ministry VHF and UHF receivers and I have to say that they all use a pentode front end albeit never in grounded grid mode.

Of course this may be because the only VHF-capable valve of the day was a pentode (usually EF50)?
Interesting that it was delevoped that way.

Remember how good the B set was on the No 19 set

[Neil Breward]

Originally Posted by ppppenguin

The first triode in a cascode provides current gain, the second one gives voltage gain. Thus between them they provide good power gain.

Jeff - interesting point. Isn't that more true of a cathode-coupled pair rather than a cascode though? I've always thought the 'lower' triode in a cascode was more for impedance matching rather than providing current gain?

Cheers,
Neil

[kalee20]

As a general rule, a pentode has 2 advantages over a triode.

The first is it has much lower capacitance anode-grid. This I think was the main reason for adding the extra grid, so that it could be at AC earth and provide an electrostatic shield between the control grid from the anode.

The second, is that it has a much higher anode resistance, due to the screen grid being held at a fixed positive voltage. Electrons leaving the cathode can hardly 'see' the anode, and accelerate towards the screen grid. It is only when they get there that most of them whizz though the spaces between the wires. So, the cathode (and hence anode) current is dependent on the screen voltage, and hardly on the anode voltage. And, as there is an element of randomness as to whether individual electrons pass though the screen grid or are captured by the wires, there's a source of noise.

For grounded-grid operation, the control grid provides a good electrostatic shield between the output electrode (anode) and input electrode (cathode), so there is low feedback. Adding an extra grid or grids will hardly improve this. So, no advantage in using a pentode for this reason.

However, there's still the issue of output resistance causing damping on any tuned circuit in the anode. Unlike the grounded-cathode configuration, the output resistance in grounded-grid is highly dependent on the sourse resistance. A cascode circuit works well with little damping, because the grounded grid stage is fed from a high source impedance. No advantage in using a pentode for this stage! But, if you feed your grounded-grid stage from a low impedance source (eg a 75 ohm dipole or a step-down transformer), then yes, the triode can present a lowish output resistance, and in this case, a pentode would be a better choice.

The only thing now, is, having chosen a pentode, why use it in grounded-grid? You get the same high output resistance and low feedback capacitance in grounded cathode, without the low input impedance. So, you get more power gain.

So any way I can see, there seems to be little advantage in using a pentode in grounded-grid configuration.

[Neil Breward]

Originally Posted by kalee20

The only thing now, is, having chosen a pentode, why use it in grounded-grid? You get the same high output resistance and low feedback capacitance in grounded cathode, without the low input impedance. So, you get more power gain.

So any way I can see, there seems to be little advantage in using a pentode in grounded-grid configuration.

OK. Suppose you're matching to a 75-ohm coax from, say, a VHF dipole. The classic ECC85 grounded-grid RF amp circuit has a broadband low-ratio input transformer, the output being tuned. However, the gain of a single grounded-grid triode is not very high - cascodes are much better, it seems, and are less prone to cross-modulation. Using a pentode like an EF80 in grounded cathode mode causes problems as the input impedance at 100 MHz is fairly low, probly only a few K's, and (as has been noted) pentodes are inherently noisier due to the 'partition effect' of the screen grid. In grounded-grid mode the input Z is low and will match closely to the 75-ohm input. The control grid will be an effective screen, and a bypassed screen grid will add to this screening - perhaps only marginally. The higher ra figure should give less damping on the o/p tuned circuit and a higher gain ... but...

Now, was the g-g pentode circuit not adopted because provides little improvement in gain/noise figure relative to a triode? Most replies seems to support this view, and I'm happy to accept this explanation, though firm evidence seems to be lacking.

I'm still left with the nagging feeling that the ECC85-type tuner, like the EABC80 combined AM/FM detector/AF amp circuit with all its compromises, were driven by economic convenience rather than best-practice design . . .

Sorry if this is drifting off-topic!

Thanks all,
Neil

[MichaelR]

Originally Posted by Neil Breward

I'm still left with the nagging feeling that the ECC85-type tuner, like the EABC80 combined AM/FM detector/AF amp circuit with all its compromises, were driven by economic convenience rather than best-practice design . . .

Sorry if this is drifting off-topic!

Thanks all,
Neil

I agree entirely , remember also for the purposes of domestic radio, Companies such as GEC marconiphone at that time also made valves.Valves intended for the domestic market were designed like any commmodity to be as inexpensive as possible, the market for valves was not for the end user but the producers of the sets. Show the producer a component that can be used to minimise the cost of his design he would be interested.

The mass market Set producers wanted good performnce but enough to satisfy their mass market user not specialist users. People bought sets for their sound, looks and price.In a mass market get any one factor wrong and you lose !

Mike

[kalee20]

If we ignore economic considerations, but baseline a requirement for perhaps a gain of 20 at the best possible noise figure, it would be possible to do a comparison of techniques.

I think the 'classic' ECC85 circuit RF stage won't achieve the gain. But, it does isolate the oscillator from the aerial so there's minimal radiation.

I'd expect a pentode would achieve the gain, albeit at a worse noise figure.

My own preference would be to use a cascode circuit (as used in Band 3 TV tuners) - more expensive, but you'd get the gain at a better noise figure than the pentode.

The single triode (grounded grid) would lose out, but you can make up the gain in the IF amplifier.

For true comparison, you'd need to do the calculations, and at this point I'd have to say I've never done any RF noise calculations (or made a 100MHz RF amplifier). But, this is my gut feeling - best practice would be use a cascode double-triode amplifier, a triode mixer, and a separate triode oscillator. Four triodes!

[Neil Breward]

I suspect you're right Kalee. I remember some late 1950s 405 TVs had Band II coverage featured, using the classic PCC84 (cascode) - PCF80 tuners. They were probably pretty good, but sadly the quality of the transmissions was probably ruined by the awful AF stages and tiny loudspeakers which were typical of such sets!

- like single-ended PCL83's with no negative feedback, driving 5-inch elliptical speakers via very small output transformers.

Cheers,
Neil