Triode Connected?

[Robert Gribnau]

I can hardly believe that nobody tried that before. I think I read something about it by member "Tubelab_com" on diyAudio but I could very well be wrong about this.

[barrymagrec]

Originally Posted by Robert Gribnau

I can hardly believe that nobody tried that before. I think I read something about it by member "Tubelab_com" on diyAudio but I could very well be wrong about this.

The Leak TL25A used a KT61 regulator circuit to drop the 445 volt HT to about 300 for the KT66 screens.

[turretslug]

I'm sure that I've seen an ultralinear circuit published that had stabiliser valves in series with the screen grids, likely with either EL34s or 807s with their high anode voltage capability. I did wonder if high volume excursions could result in extinguishing/re-ignition delay with audible consequences, also as discussed in another thread, odd effects can occur at below recommended currents.

[G6Tanuki]

Originally Posted by barrymagrec

The Leak TL25A used a KT61 regulator circuit to drop the 445 volt HT to about 300 for the KT66 screens.

I was thinking more along the lines of putting some sort of fixed voltage-subtracter in each of the paths from the UL taps on the transformer primary winding to each of the valve screens - so they essentially just 'took away' a couple of hundred volts (giving the screens an easier life when the HT was increased] but didn't affect the audio fluctuations.

The point about glow-regulators possibly extinguishing under certain combinations of anode- and screeen-tap voltages and its likely sonic consequences is noted; must admit, in my first mind-experiment that hadn't occurred to me.

I _have_ seen, somewhere, a single glow-regulator used as an alternative to the usual screen-dropper resistor between the HT and the screen-supply in a non-ultralinear-connected amplifier.

[Robert Gribnau]

The diagram and formula on the last page of the Tung-Sol 5881 datasheet (see post #19) give some room for optimism since Vg2 still is 228 V when Va hits 0 V. So as far as I can see, atleast some screen current will flow at all times. But the diagram and formula are probably a bit of a rule of thumb, based on a pure resistive load, so straight loadlines. With complex loads, things are different.

Attached are three oscilloscope measurements of pentodes in pentode mode (source: Philips, 1951).

The first picture shows the loadline of a loudspeaker, driven by a signal with a single frequency (the larger the ellips, the larger the amplitude of the input signal).

The second picture shows the loadline of a loudspeaker as the load, driven by two frequencies which differ by a factor of 6.

The third picture shows the loadline with a loudspeaker as the load, driven by music.

I wonder what happens to the screen grid current in UL operation with real life loads. Could it drop to (almost) zero?

[G6Tanuki]

Here's the thing I remember from "Amateur Radio Techniques" [RSGB, Pat Hawker G3VA] which used a glow-tube voltage regulator to slice some volts off the HT line to feed the screens of a non-ultralinear push-pull amplifier; part of me likes this idea because unlike a dropping-resistor, the neon/argon glowbottle will drop a _constant_ voltage irrespective of the actual screen-current.

[Radio Wrangler]

I don't know about audio applications, but RF amplifiers with high perveance valves are notorious for their screen currents going negative over part of a cycle, so shunt regulators are normal in this service.

(Eimac: Care and feeding of power grid tubes,)

DAvid

[kalee20]

Originally Posted by joebog1

With valves that have grid 3 available at the base, such as the 6CA7/EL34, most people strap it to cathode, but a much more elegant design is to tie it to the negative bias voltage.

I'm really dubious about that!

Taking g3 highly negative will put quite a retarding field between g2 and anode. Sure, it will limit anode current, but it will divert electrons back to the screen thus increasing screen grid current.

Loss of bias on g1, or a heavy positive-going excursion, might in this case over dissipate the screen grid.

There is also the possibility that with a heavy negative bias on g3, the knee of the pentode characteristic curve may actually be raised, which would reduce power output due to reduced anode voltage swing.

[Robert Gribnau]

It has been done by Yorkville/Traynor though (power tubes are 6CA7).

[joebog1]

Thanks Robert,
I have studied Traynor exactly because he was "outside the box ".
Yorkville amps ( Canadian company ) are like I said, ALMOST blowup proof. Australian made 6CA7 ( by Mullard in Adelaide I believe ),
are HIGHLY regarded, especially in guitar circles. Nope I cannot, nor ever have I played a guitar.
I CAN play records however .

Joe

[Robert Gribnau]

Originally Posted by turretslug

I'm sure that I've seen an ultralinear circuit published that had stabiliser valves in series with the screen grids, likely with either EL34s or 807s with their high anode voltage capability. I did wonder if high volume excursions could result in extinguishing/re-ignition delay with audible consequences, also as discussed in another thread, odd effects can occur at below recommended currents.

Perhaps this problem could be met by using voltage stabiliser valves with ignition electrode, making very low currents possible. Some of these valves are described in this document:

https://frank.pocnet.net/other/Telef...isers_1966.pdf

When using the ignition electrode the ZZ1010 has a control range from 0.5 to 70 mA, while the ignition electrode needs 0.2 mA to keep on working.

The ignition electrode also makes it possible to shunt these valves with as much capacitance as one wants, without the risk of oscillations (but I don't think this feature is usefull for the application we are discussing now).

[Radio Wrangler]

Originally Posted by kalee20

Taking g3 highly negative will put quite a retarding field between g2 and anode. Sure, it will limit anode current, but it will divert electrons back to the screen thus increasing screen grid current.

Indeed.

Look up Transitron oscillator and Dynatron oscillator to see it in use. The Dynatron was invented by Hull at GE - he also came up with the magnetron, the split anode version which Randall & Boot developed the cavity magnetron from.

The Miller Transitron timebase found in simple oscilloscopes is another example.

All this mention of oscillators re something being discussed for amplifiers ought to be making people feel nervous...

David

[G.Castle]

I was recently discussing this suppressor connection question with a friend who is designing a small amp which has DH output valves, and is going to try to use a trick filament supply.

Originally Posted by Radio Wrangler

All this mention of oscillators re something being discussed for amplifiers ought to be making people feel nervous...

(If you want an oscillator just try to design an amplifier?)

Funny this should be mentioned just after a diagram of that York amp has been posted, it gave me a well needed larf!

If anyone is interested I can circle the four resistors, two valves and the transformer that will go straight to guitar amp heaven when it goes unstable and isn't turned off PDQ.

It's out of the box OK, or at least the voice coils are out of the air gap!

Wasn't funny for the owner though

Thanks everyone! An interesting and entertaining discussion as always.

Greg.

[joebog1]

Interesting you say that Greg, I have worked on a couple ( they are very rare amp ) and have never had any stability problems with them. Like all 6CA7/EL34 valves though, they dont last long at high power.
They were more for the hi-fi world. ETI published a design called the "130 watts of rich warm valve power" It was warm all right, I have never seen one last more than about an hour when being pushed.
It didnt use Traynor's idea though.

Joe

[Synchrodyne]

Some curves for pentode suppressor grids may be found here: https://www.radiomuseum.org/forum/du..._pentodes.html.

In particular, those curves show that for some, but not all pentodes, suppressor grid negative bias down to -40 V has but minimal effect on the basic valve curves. On the other hand, positive suppressor grid bias introduces tetrode characteristics.

If the EL34 suppressor grid characteristics fall into the “not much affected by negative bias to -40 V” group, then one might deduce that the -37.5 V bias applied in the Yorkville/Traynor amplifier is essentially harmless.

That amplifier used the EL34 in pentode mode. I am not sure that the suppressor grid characteristics when measured in pentode mode would also be applicable to the triode mode.

As an aside, Lowther, in its “Lowther Linear” circuits, used the EL34 suppressor grid as part of an additional feedback loop. This appears to have been with positive bias, though.

Re the question of ultralinear with different HTs for the anodes and screen grids, I think that the first diagram in the Hafler & Keroes patent provided one solution, albeit perhaps difficult and costly from the transformer perspective.


Cheers,

[joebog1]

The ETI amplifier used a separate supply for the screens derived from a separate secondary winding on the power transformer. Without digging out the circuit I seem to remember it was 225 volts or thereabouts.

Synchrodyne, The circuit from Hafler is more a cath amp design very similar to Quad. Requires another power supply voltage.

Joe

[Robert Gribnau]

Originally Posted by Synchrodyne

Re the question of ultralinear with different HTs for the anodes and screen grids, I think that the first diagram in the Hafler & Keroes patent provided one solution, albeit perhaps difficult and costly from the transformer perspective.

Menno VanderVeen from The Netherlands sells output transformers with seperate windings for the screen grids. A link to one of them:

https://www.mennovanderveen.nl/index...1070-uc-detail

[joebog1]

I have his book " Make your own audio valve amplifiers ".
Some interesting stuff in the book. LOTS of type errors or maybe typesetting.
VERY expensive transformers. I wonder about the coupling of his transformers though, No interleaving, and no separation between sections. Yes they are toroids with pretty good coupling to begin with, I am still a bit suspect about his specifications. Especially at 100 Khz.

Joe

[Robert Gribnau]

I have three of Menno's books but the one you mentioned didn't ring a bell. Can't find it when searching on internet either.

Do you perhaps mean "Build Your Own Audio Valve Amplifiers" by Rainer Zur Linde? If so, his book contains some serious mistakes. There's a thread about this here:

https://www.vintage-radio.net/forum/...d.php?t=175065

I'm using Menno's transformers in some my builds and are very happy about their performance (but I didn't perform elaborate measurements on my builds). Designing and manufacturing good transformers, especially toroids, seems to be (one of) his goal(s) in life since 1990 of even earlier than that. I have no reason to doubt the data he provides about his transformers.

[joebog1]

You are correct Robert, my apologies.
I do have one of Menno van der Veens books though. It has some stupid designs too!! like eight off EL34 valves in push pull parallel. What is he driving ? a modulated welding machine ?. IN fact, in that book its just designs where he starts with a single pair of EL34 and then goes to four, then six and lastly eight. He explains phase splitters that have uneven output impedances and it doesnt use a buffer to drive the output valves ?.

Joe