• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

New: Equal impedance output split-load phase splitter

O

oldeurope

Active Member
A cathodyne with equal output impedances is possible by using a unity mu triode.
I am waiting for this triode since 2007 and I don't give up hope.

During waiting, negative feedback makes it possible too.
In the schematic a 12AX7 example is given.
More schematics will follow soon.

Blog:

http://1mucathodyne.blogspot.de/2014/09/das-ein-katodyn.html

blog is under construction, tnx for looking. :tresbon:

Note that the output signal is able to floate. :yes:
The Achilles Heel in the high µ split-load circuit is solved this way.
Klick and scroll down.

LG Darius
 

Attachments

  • 1_mu_Cathodyne_12AX7_schema.png
    1_mu_Cathodyne_12AX7_schema.png
    14.6 KB · Views: 88
Register to hide this ad
Okay, I'll bite. I read Broskie's article that you linked. Summary I got from it is a split load phase inverter, if used to directly drive an output stage, can make apparent the dissimilar impedances in the output stage when the output stage is over driven. Okay, but that situation exists with a long tail pair, too. Maybe this is why some more exotic designs place cathode followers in between the phase splitter and the output stage.

So if the designer is cognizant of it, it seems like cathode followers are the way to go. Seems like you could build those cathode followers from silicon as well. Or...am I missing something? It's possible I may have missed the point of Broskie's article.

How does your circuit work? Does it eliminate this problem somehow?
 
I agree with kward. I'm not sure I'm seeing the advantage here. Using absurd conditions to illustrate:

With a traditional splitter:

1. If the top output is shorted to ground, the bottom output will see little if any increase in output.

2. If the bottom output is shorted to ground, the top output will increase by the (restored) effective gain of the splitter stage.

With the new arrangement:

1. If the top output is shorted to ground, the bottom output will increase by a factor equal to the loss of FB in the loop.

2. If the bottom output is shorted to ground, the top output will increase very little due to the FB loop compensating for the increased gain.

In other words the scenarios are basically reversed.

In driving a class AB output stage, what happens to the "other" side due to one side being clipped is of little relevance: That side is already cut off anyway.

Now in class A output stages, this could be an issue. But otherwise, I agree that the ultimate wrapping to place on an inverter stage package is to use driver or cathode follower stages after it.

Problems only occur with a split load inverter when the following stage load impedance placed on the inverter's two outputs start to become unequal, as at the onset of clipping. As long as the load impedance is equal in every aspect on both outputs, then the output is equal in every aspect as well.

Dave
 
kward said:
...

How does your circuit work? Does it eliminate this problem somehow?
Click to expand...

Hello,
the feedback loop from the anode of the cathodyne to the cathode of the gain stage
increases the cathode output impedance and decreases the anode output impedance of the cathodyne.

If the feedback loop is will aligned, impedances will be equal.
The µ=1 Triode does this job without feedback.
 
What happens when say the output tubes draws current? Does the output impedances still stay the same?
 
jazbo8 said:
What happens when say the output tubes draws current? Does the output impedances still stay the same?
Click to expand...

Well, besides the impedance issue- the OTHER problem that occurs, that none of the proposed inverter topologies from the OP address, is blocking distortion caused by changes in effective tube static bias, when grid current causes the coupling cap to charge (as in when the amp clips). This drives the tubes farther into cutoff, earlier, causing notch distortion.

Until you address that, there's not much point in worrying about the differences in AC impedance of the previous stage... this blocking distortion will swamp any effect of dissimilar driving impedance, as long as there are coupling caps.

The solution? Just don't drive the amp far enough into clipping, to get grid current. If the amp isn't big enough to do that at a reasonable level, get a bigger amp, or convert it over to DC coupled buffered grid drivers on the outputs (i.e, cathode followers or such). Both problems (dissimilar impedance AND blocking distortion) are solved at once, with the buffer solution...

Regards,
Gordon.
 
GordonW said:
The solution?

Regards,
Gordon.
Click to expand...

Klick

GordonW said:
Well, besides the impedance issue- the OTHER problem that occurs, that none of the proposed inverter topologies from the OP address, is blocking distortion caused by changes in effective tube static bias, when grid current causes the coupling cap to charge (as in when the amp clips). This drives the tubes farther into cutoff, earlier, causing notch distortion.

Until you address that, there's not much point in worrying about the differences in AC impedance of the previous stage... this blocking distortion will swamp any effect of dissimilar driving impedance, as long as there are coupling caps.

...

Regards,
Gordon.
Click to expand...

Thanks Gordon.
But there is something bevore heavy blocking. The PA goes from Class AB to Class B.
The common cathodyne is not able to handle this without distortion.
The reason for this is that the difference signalvoltage changes.
If you touch the the grid current area with the cathode side, it will enlarge
the whole signal.
This problem is solved by the 1µ-Cathodyne. And the circuit is a very easy one compared to other solutions.

For the AC-Component the 1µ is similar to a transformer winding.
The conventional cathodyne not.

It makes sense to my to build a line preamplifier that has a floating sym-aysm output.

Regards,
Darius
 

Attachments

  • 1mu_cathodyne_line_pre_2014_schema_darius_001.png
    1mu_cathodyne_line_pre_2014_schema_darius_001.png
    19.8 KB · Views: 27
rage said:
I ilke a simple LM334 tailed differential pair. :)
Click to expand...

OK, I don't like the LM334 semiconductor constant current source in the tube circuit.

You'll get a time delay between the outputs.
(This is the reason why Williamson used a cathodyne in front of.)

Note the output impedance it is cathode-follower like! :yes:
 
Please help me translating this text into good english.

A feedback loop from anode of cathodyne to the gain stage provides equal output impedances with respect to ground. Kathode resistance parallel anode resistance. The impedance between the inverting and the non inverting output is reciprocal value of mutual conductance of the cathodyne triode (i.e. Cathode follower) under all circumstances such as symmetrical, asymmetrical, inverting or non inverting and floating.

http://1mucathodyne.blogspot.de/2014/09/das-ein-katodyn.html

(blogtext written in green.)

tnx
 
oldeurope said:
OK, I don't like the LM334 semiconductor constant current source in the tube circuit.

You'll get a time delay between the outputs.
(This is the reason why Williamson used a cathodyne in front of.)

Note the output impedance it is cathode-follower like! :yes:
Click to expand...

Oldeurope, what is the nature of this time delay you are describing? I've never heard of this phenomena before when using a ccs (either tube-based or silicon) in a tail of a long tailed pair.
 
inv signal goes from grid to plate of the grounded chatode amp.
non inv signal goes from cathode follower to a grounded grid stage
and the CCS must let it pass.
The LM gets it current source charakteristik from a circut that contains a few transistors.
I "natural" current source like a pentode's anode or positiv grid triode's anode
or a long tail resistor has not to react on something.

A complementary diff-amp does not suffer with such problems because it needs no
current source like the cathodyne.
You can see the cathodyne as a special kind of complementary diffamp that can be made with tubes.
 
So it sounds like you are saying that because the first tube in the cathode coupled differential drives the second tube through it's cathode, it causes this phenomena (time lag) when used with a transistor based CCS? But why does the first tube driving the second tube in this way cause a time lag only when using a transistor based CCS? Is it because of the inherent capacitance of the MOSFET device?

Lastly, I think you are saying that when using a cathode coupled differential pair phase inverter, that you would rather sink it with a pentode rather than a transistor based device such that you avoid this time lag?

And I apologize if this is taking your thread off topic. I'm just trying to understand these things you are describing.
 
A collector or drain output is a current output from its nature.
Triode is a good example:
Grid negative means voltage output, grid positive current output.
In between resistor like.
Often you see triode circuits using a negative grid (respect to cathode) and a cathode resistor.
In this case the triode "transforms" the cathode resistance µ-times to the plate
and acts as a current source. Truly it is a driven voltage source.
Hu, this is difficult in english for me, sorry.
 
mucathodyne-line-pre

some additional pics ... :music:
 

Attachments

  • 2014_0925mucathodyne_line_pre0002.jpg
    2014_0925mucathodyne_line_pre0002.jpg
    92.4 KB · Views: 9
  • 2014_0925mucathodyne_line_pre0003.jpg
    2014_0925mucathodyne_line_pre0003.jpg
    84.5 KB · Views: 11
You must log in or register to reply here.
Back
Top Bottom