pse help, unity µ triode as a phase inverter

[oldeurope]

Originally #23 posted by Michael Koster
#22

"I think the attachments from this thread
are interesting for you."

Very interesting. The signal path outlined in red helps a lot

...

Michael

Hello Michael,
thank you very much for your interest in the "fourth circuit topology".
Do you have the update_1.May_2007 ... zip from the vintage repair and discussion forum?
It contains the ECC83 PCL85 dc coupled line pre schematic.

Kind regards,
Darius

[kenpeter]

Originally posted by oldeurope
Hi kenpeter,
I tested the arrangement in post #32.
µ'< 1/µ .

How far off the 1/µ mark did it land? (With your 6080?)
And the overall cascode? How far from 1?
----------------------------------------------------
What constitutes 'special construction'?
I was kinda leaning towards 6N30Pi,
especially for a smaller tube.
----------------------------------------------------
I was never too certain the Mu of the lower
triode as it is not driving constant current,
but sees the same current as the top tube.
A transformed (4th circuit) load resistance?

[oldeurope]

Originally #24 posted by Wavebourn
Take any tube and select load resistance and idle current to get exactly unity gain. No feedback, the problem solved.

oh oh, Wavebourn,
loading down a triode is a very bad design.
It disappoints me to hear something like this from you.
Have a look at post #12 and click on the hyper link
well known factor. Hope it helps ...

Originally #25 posted by jon_010101



So true! -- I would wire up a unity-gain plate follower and call it a day

Hi Jon,
See the advantages listed in #12.
The plate followers impedance depends in the impedance
of the feedback loop and I don't like feedback.
The unity µ triode stage see post #6 has no feedback
thus the input resistance is not reduced by negative
voltage feedback like it is in a so called plate follower.
BTW Look at this list, funny isn't it?


Hello kenpeter

Originally #42 posted by kenpeter


How far off the 1/µ mark did it land? (With your 6080?)
And the overall cascode? How far from 1?
...

I tested some ECC82. I didn't write down the results,
it was some weeks ago.
See post #12
From the graph I get µ'= 0,15 @ Ia'=100mA, Ug'=-10V for the 6080.
Thus in this example you should expect something &asymp; 0,3.

Originally #42 posted by kenpeter

...
----------------------------------------------------
What constitutes 'special construction'?
I was kinda leaning towards 6N30Pi,
especially for a smaller tube....

I don't know this tube.

Originally #42 posted by kenpeter
...
I was never too certain the Mu of the lower
triode as it is not driving constant current,
but sees the same current as the top tube.
A transformed (4th circuit) load resistance?

A transformed constant current source is still a constant current source.

Kind regards,
Darius

[ilimzn]

Looking at the ECLL800, it's phase inverter triode has a muo slightly higher than 1, to account for the fact that actual gain is derived through a divider comprising of the plate resistance and plate resistor.
Not sure i know of a triode with anywhere near the required mu, the one in the ECLL800 is really special that way. But, depending on the swing you need, you could simply make the plate resistor an actual divider. Not the best approach as you are amplifying, then attenuating, but given the scarcity of triodes that would fill the need, it may be the way to go. Start with something low-mu, like 12B4...

[oldeurope]

Originally #44 posted by ilimzn

...Not sure i know of a triode with anywhere near the required mu, the one in the ECLL800 is really special that way. But, depending on the swing you need, you could simply make the plate resistor an actual divider. Not the best approach as you are amplifying, then attenuating, but given the scarcity of triodes that would fill the need, it may be the way to go. Start with something low-mu, like 12B4...

Thanks, 12B4 µ = &asymp; 6
I was thinking about the EF98 µg2 &asymp; 4

The voltage divider loads down the triode,
thus a good compromise between output resistance
of the divider and high well known factor must be
found.
It is sad that there is nothing like the triode section
of the ECLL800. This triode is a clever one.
The 6080 performs excellent. I expected hum problems
from AC heating. But, surprise, no hum.

Kind regards,
Darius

[smoking-amp]

"why do you ignore my posts and hyper links?"

I've read them. I commented that the scheme Ken came up with was related to your 4th topology. Then I went on to try some variations.

"Why do you tell porkies in this thread?"

Because it is like a Greek comedy or the Odyssey, we go around in endless circles trying to accomplish what is easily done conventionally. I'm not complaining, just enjoying the ride. It might be helpful if we could tie down the rules a little though. Best I can tell are:

1) no intentional feedback allowed
2) this (1) includes cathode followers and concertina splitters and LTP inverters
3) no resistive attenuation allowed to reduce excessive gain
4) no low value plate resistors allowed that would increase distortion
5) no cheating allowed such as just using the un-inverted signal to drive a subsequent cathode instead of the grid
6) no SS parts in gain path (so no N type mirror into a P type mirror to invert phase)
7) did I miss any? (are CCS's allowed?)

"µ'< 1/µ"

Looking at the 6AS7 data, I see normal Mu at 2 and inverted mode Mu' at 1/5 Some of this discrepancy may be due to the lack of space charge beyond the positive grid in this mode. Hence reducing the inter-electrode capacitance to the plate versus normal operation. And likely increased space charge between cathode and grid1 due to the concentrated current conduction there.

This is one reason I proposed using a pentode to play off screen Mu versus plate Mu (using a low Rp pentode). This would preserve the normal space charge distribution if operated in positive g2 input mode. Likely to still give some overall gain though.

Or, by using a CCS in the cathode path, one can get high input impedance for a plate input using normal (positive plate) mode operation.

What's wrong with the beam deflector tube solution by the way?
We could probably scheme up some dual control tube solutions too maybe. (6BU8,6KF8,6MK8,6GS8,6HS8,6LE8 ---- 6BV11,6GY6,6HZ6,6AS6,6DB6,6DT6,9KC6 ---- or even pentagrid mixer tubes: 6BY6,6CS6,6BA7,6BE6,5750,6SA7,5915 ...) With all those grids available, there's gotta be some way to get unity gain. Then there's the shadow grid tubes, could be a very interesting dual grid solution there: 6FG5,6FS5,6GU5

Don

[Wavebourn]

Originally posted by oldeurope


oh oh, Wavebourn,
loading down a triode is a very bad design.

Yes, it is a very bad design, but it is the single one that satisfies your main condition: "No feedback added"!

It disappoints me to hear something like this from you.
Have a look at post #12 and click on the hyper link
well known factor. Hope it helps ...

I have to disappoint you even more, because triodes are always loaded in order to work...

[SY]

I commented that the scheme Ken came up with was related to your 4th topology.

Landee, Davis, and Albrecht just refer to that topology as "plate driven tube." It doesn't sound as mysterious that way.

[oldeurope]

Originally #48 posted by SY


Landee, Davis, and Albrecht just refer to that topology as "plate driven tube." It doesn't sound as mysterious that way.

We have three well known topologies for transistors, pentodes ...
and a fourth one for triodes. Thus it don't sound mysterious.
Please note, this is only possible for triodes not tubes generally.
I am very interested in Landee, Davis, and Albrecht text
and schematics about coupling triodes.
I'd like to add this to the blog.
Please show it to me. Thanks.

Kind regards,
Darius

[smoking-amp]

I looked in my Electronic Designer's Handbook by Landee, Davis and Albrecht and found this for "plate-driven tube" on pages 3-6 and 3-7: