Triode Connected?

[Radio Wrangler]

Originally Posted by Synchrodyne

One may infer that Dynatron saw the L77 as definitely better than a triode-strapped Z77 for the oscillator job, but that the obverse applied to the grounded grid RF preamplifier stage.

As mentioned up-thread, the strapped EF91 offers appreciably higher Gm and mu than the 6C4, but oscillators aren't usually demanding on these parameters, so the 6C4 offers plenty for oscillator service. Oscillator designers of some experience look for reliable non-linear effects to give amplitude stability, and also freedom from flicker and 1/f noise that often gets translated into phase noise sidebands. I'm not too sure how much these effects were appreciated back when Dynatron ltd was on the go.

David

[G6Tanuki]

Yes, Receiver Local-oscillator designers wanting a free running LO generally prefer a medium Mu triode over anything with crazy slopeiness, one of my favourites being a triode connected 6AK5.

Designers sometimes went off at a tangent though, the Collins TCS receiver uses a 12A6 beam tetrode as its LO.

I guess the Collins people had a good reason.

[turretslug]

The L77 also takes half the heater power of the Z77, so a watt less in the vicinity of the oscillator components- probably a peripheral concern overall but another plus point when taking possible candidates into consideration at design. I'd wondered if high gm valves, with their necessary close-spaced and fine-pitch grid wires placed very close to the cathode would be more prone to temperature-induced capacitance changes when changes of a fraction of a pF can have an effect on oscillator stability at high HF.

When Eddystone's 888 receiver (itself developed from 750 architecture) was revised into the 888A model, one change was the first LO going from triode-connected 8D3/Z77/6AM6/EF91 (etc....) to L77/6C4/EC90, so it must have been deemed a worthwhile change there, too.

[Aub]

[QUOTE=Radio Wrangler;1496453

The Eddystone EA12 uses a couple of 6C4s as its oscillators.

I've used them a few times, over the last forty odd years, in VFOs , in my amateur AM transmitters. Lovely little triode.

Cheers

Aub

[Synchrodyne]

There were some other examples of triode-strapped pentodes in Eddystone equipment.

The 770R VHF receiver of the mid-1950s used a triode-strapped 6AK5 as local oscillator. The 6AK5 was also used, in pentode mode, in both the RF amplifier and mixer positions. The S-meter driver was a triode-strapped 6AU6, and the BFO a triode-strapped 6BA6.

The 830 HF receiver of c.1962 used a triode-strapped 6AU6 as cathode follower for the 2nd IF output facility. The 2nd frequency change oscillator was a 6C4 triode, consolidating the preference for this actual triode in oscillator positions. The 1st oscillator was a 6U8 triode pentode, with the triode side used for the crystal oscillator. The pentode side, “half-strapped” as it were, was used as the VFO. The oscillator circuit used the pentode screen grid as its anode, so was effectively a triode. But the oscillator output to the mixer was taken from the anode, a low impedance source in this configuration. So the pentode was acting half as a triode and half as a pentode.

That brings to a mind a vague recollection of a circuit – maybe a microphone preamplifier – that used an EF86 with the screen grid as amplifier anode, making it effectively a triode (should we perhaps name it as a “virtual triode”), and the anode grounded, perhaps acting as an additional screen. I cannot now find the reference, which leaves me wondering whether or not it was real. But the anode grounded as a shielding screen was sometimes used for triodes employed as diodes, with the grid acting as diode anode.

There was a broad analogy in solid-state devices. When RCA, c.1968, elected to include gate protection diodes in its then still new dual-gate mosfet range, it chose not to do the same for its single gate mosfet range. Rather they were left unprotected in order to cater for applications requiring extra high input impedance, beyond that achievable with gate-protected devices. The thinking was that if anyone needed a single gate-protected mosfet, they could use a dual-gate device with the gates strapped together. The term mosfet tetrode was sometimes used for the dual-gate device, so one could say that when the two gates were strapped together, it was “triode-strapped”.


Cheers,

[Radio Wrangler]

Originally Posted by Synchrodyne

There was a broad analogy in solid-state devices. When RCA, c.1968, elected to include gate protection diodes in its then still new dual-gate mosfet range, it chose not to do the same for its single gate mosfet range. Rather they were left unprotected in order to cater for applications requiring extra high input impedance, beyond that achievable with gate-protected devices. The thinking was that if anyone needed a single gate-protected mosfet, they could use a dual-gate device with the gates strapped together. The term mosfet tetrode was sometimes used for the dual-gate device, so one could say that when the two gates were strapped together, it was “triode-strapped”.

Gate protection diodes are still a problem with some modern super-duper RF power MOSFETS.

Semiconductor manufacturers have totally dedicated themselves towards complex modulation schemes (usually complex in both meanings of the word) and their need for linear amplification. So they put single-ended protection zeners on the gates of their LDMOS devices, knowing that they will always be biased into conduction, and there will be no negative-going gate drive.

This locks them out of Class-C and a number of other 'switching' type modes.

Having to design a very efficient transmitter for an aviation COM radio, a service locked into the AM mode, suddenly ran into this problem. Most suitable devices were obsolete and therefore unsuited to new product development, Finding ones which could be swung negative was an interesting challenge. It was something not worth mentioning on data sheets.

Plenty of oscillators get their amplitude control by running the active devices into cutoff over a variable portion of the cycle. Good phase noise behaviour means running an oscillator at a strong amplitude and this translates into gate voltages swinging negative even with small enhancement mode devices. Any clamping action ruins the Q of the resonator, and thence the oscillator's performance.

It ain't easy trying to make anything other than a cellphone or a computer these days.

David

[Synchrodyne]

In the late 1940s, Mullard showed triode-strapped VHF oscillator circuits for both the EF91 and the EF42. In each case another valve of the same type, but in pentode mode, was used as mixer. In both cases the oscillator g3 was connected to the cathode and ground.




Connecting g3 to the cathode seems to have been the Mullard way with triode-strapping of pentodes. It was done with the EF37, for example.





Cheers,

[joebog1]

For Robert,
So I dug the book out and its scan is presented.
The amps in question are triode strapped ( which I didnt remember ) and G3 is tied to cathode, as most people do. He runs the 6CA7/EL34 with VERY low volts of 340 and 800 mA to fetch 80 watts. I can see why I sort of "banished " his book.
I prefer Davids idea of 3-500Z or even better, a 3-1000 Z, superb in winter time for high powered listening. From memory again, 10 volts @ 30 amps heater current.

So I still maintain my statement that eight off 6CA7's to fetch 80 watts is audiophool territory, OR perhaps you own your own atomic power station.


Joe

[joebog1]

A further edit:

He did use a 12AU7 as the phase splitter. I did read this book perhaps 20 or 25 years ago ( I dont remember precisly ) so my apologies for that. I still maintain its audiophool territory. 24 amps JUST to heat the output stage in a stereo amp

Joe

[Robert Gribnau]

Thank you.

So there's nothing wrong with the phase splitter afterall. Only the power consumption makes it audiophoolery for you.

I have a different idea of what makes something audiophoolery than you.

[joebog1]

Thanks Robert,
I have ALWAYS maintained I am an audiophool.
BUT, you may have your opinion, and I have mine.

AS far as the phase splitter goes ??, I still maintain to drive EIGHT valves, REGARDLESS of how they are connected, A better solution is a cathode follower AFTER the phase splitter.

But I am just a phool so I dont have a clue.

With my best regards

Joe

[Radio Wrangler]

While exploring what could be done with bipolars, MOSFETs and valves, I designed three amplifiers. The last one to be done was the valve version and the semiconductor jobs had set a hard target for it to equal. It was output transformerless and used, I think it was 20 TV line output valves in a paralleled totem pole arrangement (one channel!). These devices needed very low drive impedance to their grids, so cathode followers running significant current were needed. Without transformers and with low impedances throughout, I could equal the feedback gain/bandwidth of the transistors ones. The distortion was almost as low, but there was an issue in the differential input stage I never found a solution for. Complementary valves would have been good. Anyway, I'd engineered out all traces of 'valve sound' Perhaps illustrating that it wasn't so much the valves as the fairly basic circuits they usually got used in which was the reason. The whole amp used a stupid amount of power. Class A transistors would have been an improvement.

So wilder valve amps can be done. There are things to be learned from them, not least of which is their limitations. Traditional circuits soon come up against the output transformer as a limiting factor, and its limitations trade off into all other areas. Triode and ultra-linear modes offer less forwards-path distortion, and aren't so demanding on feedback, but they are less efficient than beam tetrodes in terms of output power versus the size of valve needed. For the transformerless job, valve matching was a major issue.

As for me, I built a stereo set of the MOSFET version, and the others turned into parts. The finished thing has sat in the living room for just over 40 years and it just works.

For anyone wanting a valve amplifier, I recommend sticking to the trad circuits like the Mullards, Leaks and Quad. Exotic circuits are possible, but more trouble than they're worth, they also engineer out the valve sound effect. You might just as well go straight to transistors.

David

[Aub]

Originally Posted by joebog1

Thanks Robert,
I have ALWAYS maintained I am an audiophool.
BUT, you may have your opinion, and I have mine.

AS far as the phase splitter goes ??, I still maintain to drive EIGHT valves, REGARDLESS of how they are connected, A better solution is a cathode follower AFTER the phase splitter.

But I am just a phool so I dont have a clue.

With my best regards

Joe

The GEC 400 Watt, 5 pairs of KT88 amp, does use cathode followers to drive the output valves, so you're not alone Joe

Aub

[joebog1]

Thanks Aub,
The GEC amp also makes 400 watts, NOT 80. Plus I "think " it is designed as a PA rather than Hi-Fi. That makes the output transformer much more manageble, both to make AND to carry.
Having said that, I have been aksed to make 500 watt Hi-Fi output transformers. I did the basic design
( obviously E/I core ) and they would have been 10" square cubes. No, I didnt make the traffos after the audiophool customer got the quote. He went back to his american amp which was 4 off 6146's in push pull parallel AND pentode mode. I think it was 200 watts or so.

Joe

[Trevor]

Many years ago I made the large 10 valve KT88's The GEC circuit was followed and to be Honest it worked very well ! It was used by a Jamaican friend of mine with 4 off 18" goodman spk in large reflex cabs on a DUB Reggia sound system
To be honest the sound was really rather good
Trev

[Robert Gribnau]

Originally Posted by joebog1

AS far as the phase splitter goes ??, I still maintain to drive EIGHT valves, REGARDLESS of how they are connected, A better solution is a cathode follower AFTER the phase splitter.

Could you post the rest of the schematic? You posted the part with 8 x EL34 in triode mode, but not what is driving them.

[Robert Gribnau]

Found it myself in the book "Modern High-End Valve Amplifiers" from 1999.

[cathoderay57]

I have recently finished restoration of an HMV 650 receiver that employs a KTZ63/6J7G pentode connected as a triode as AF driver. I noticed that not only is g2 strapped to the anode, but also g3. In other circuits such as triode connected EF86, g3 is connected to cathode. Does it make any difference and, if so, what? Cheers, Jerry

[Synchrodyne]

There was extensive discussion of this issue in this recent thread, with quite a few examples given:

https://www.vintage-radio.net/forum/...d.php?t=193910

I don’t think that any hard conclusions were reached, other than perhaps the disposition of the g3 connection in a triode-strapped pentode was likely a circuit designer choice.


Cheers,

[cathoderay57]

Thanks. I did search but didn't find this thread. Makes interesting reading. One additional fault that I had with the HMV 650 (not mentioned in the other threads) was that the set erupted into severe crackling and reduction of audio. Tapping the 6J7G AF driver (the valve with g2 and g3 strapped to anode) made things worse. Initially I was confused by the fact that the valve still tested 100% OK on the VCM 163, while a replacement valve worked perfectly in the set. Eventually I found that on the troublesome valve an internal short had occurred between pin 1 (M) which is strapped to chassis and pin 5 (g3) which subsequently settled to a leakage resistance of about 2k. I could have continued using the valve by strapping g3 to chassis but used the replacement valve instead. This circuit feeds into an inductive load (phase splitting driver transformer for the push-pull output stage) so maybe the g3 circuit design configuration was associated with that. Cheers, Jerry