Cathodyne PI: AC coupled vs DC coupled

[Silvio55]

Hi! I'm building an Orange OR80 (2xEL34) inspired amp, it's basically the same circuit but with lower voltage PT and post PI MV.
For those who aren't famiiar with the amp, here's an schematic for the 4xEL34 version:

This schematic shows an DC coupled cathodyne PI.
There were some variations over the years, one of those being the change of the DC coupled cathodyne for a AC coupled one, like this one:

I love the simplicity of the DC coupled version, some people like it over the AC version, others like the AC coupled better.
I want to ask you what would be the tonal differences between the two versions, if someone here have ever tried both PI's and compared them.
Thanks!

PD: Sorry for the big images, I couldn't attach them in miniatures, dunno why...

[The Dude]

Maybe I'm just not familiar with the proper use of terminology and or labeling, but the first one looks like the DC coupled one and the second AC in my eyes. The first has no coupling cap from plate of previous stage to grid of PI. The second has a cap. I think you have them backwards.

[Silvio55]

Oh man! You're right!!! Sorry. Corrected now.

[The Dude]

Better.

[The Dude]

FYI, I'm not ignoring your question. I just don't have an opinion. I would have to compare the two side by side. Maybe someone else has some "ears on experience" with this.

[Gingertube]

I would strongly recommend the AC coupled version for stability of the cathodyne's operating point.
Also read this:
The Valve Wizard

If you want the theory of "tuning" the overdrive sound, you can also read reply #27 I put here:
http://www.guitargear.net.au/discuss...topic=38629.15

Cheers,
Ian

[J M Fahey]

Another vote for the AC coupled one because it´s more tolerant of tube properties, which often vary a lot.

The DC coupled one requires a specific selected tube to reach expected DC biasing points, which is a PITA in a Factory setting , even more for an individual who does not have a ton of tubes on stock to pick the most suitable ones from.

Even worse if you are NOT building the exact original version.

[Pedro Vecino]

I prefer the second. The first with that bridge rectifier can not work well ;-)

[Silvio55]

Well, so the DC coupled version is out of the question, I'll build it with the AC coupled cathodyne. Thanks!!

[KevinOConnor]

Hi Guys

The direct-coupled splitter and gain stage is preferred for loop stability when global feedback is used. This is because there is only one coupling cap within the loop along with the OT phase shifts. However, using the AC coupled circuit is fine with feedback, too, provided the coupling caps are made either very large or very small so their phase shifts do not overlap the OT's.

With the DC coupling, the gain stage idle point has to be set to a plate voltage that is about one-fifth of the supply to that stage. This means idling the tube at a low current and using a very high Ra value, or splitting Ra and decoupling it on its own. There is no real issue that I;ve ever read about nor seen regarding bias stability being better//worse in one form or the other.

The AC coupling has the advantage of providing more range for voicing the amp. Also, if you want a "post-PI-MV", you can insert a single-section pot bwteen the gain stage and the concertina. I only use concertina splitters in my amps and they are all AC coupled.

Have fun

[trobbins]

Just for background, that DC coupled configuration is pretty much what Williamson used for the input stages of his hallmark design. But the application is very different, as Williamson used a low gain tube (aka 12AU7), there was no presence adjustment, the splitter was followed by a stage that was not meant to go in to grid conduction, and the signal levels were so much lower.

That said, if the feedback level is substantial then the feedback stage operation is very much modified from standard gain assessment. There were also thousands and thousands of amps with that DC coupled configuration made and operating - so the basic biasing operation is not bad per se, but it wasn't designed to achieve balanced clipping in to the subsequent stage (as the signal levels were much lower).

[Ken Moon]

I only use concertina splitters in my amps and they are all AC coupled.

Does the Concertina splitter allow enough signal swing to drive "big bottle" tubes to full power? I'm talking about amps with "typical" power supplies (power tube plates/screens at 400-450 volts and the B+ node for the splitter at about 350V)? Or is a higher B+ used in your amps to get a larger signal to the power tubes?

Thanks!

[Mike Sulzer]

Does the Concertina splitter allow enough signal swing to drive "big bottle" tubes to full power? I'm talking about amps with "typical" power supplies (power tube plates/screens at 400-450 volts and the B+ node for the splitter at about 350V)? Or is a higher B+ used in your amps to get a larger signal to the power tubes?

Thanks!

One way to assure that it does is to add a negative supply as described here: http://music-electronics-forum.com/t41822/. The outputs from both plate and cathode are about the same as a normal stage. Also the dc on the grid is ground potential, simplifying the bias. Furthermore you really cannot overdrive it much, if at all, since it takes the full output of the previous stage.

[KevinOConnor]

Hi Guys

The standard self-biased concertina can drive full output stages - why shouldn't it? The schematics of the amps above do this, as do all the Dynaco amps with KT-88s and typical c500V supplies. The key is to run the concertina at the highest voltage available in the amp but to assure its node is especially well filtered as one output is referenced to the supply.

The use of a negative rail to make the concertina is a good one, but is not likely to be tried by many hobbyists or builders. Part of the reason is lethargy, part is "Leo didn't do it", part is a fear of negative voltages.

A Schmitt splitter can generate the same output signals with a lower plate supply voltage because there is only one signal in series with the tube. In the concertina there are two signals in series with the tube and there must always be a little voltage across the tube and the plate and cathode resistors for the circuit to be linear. The split-rail concertina does not have this disadvantage as its supply is effectively much higher.
Have fun