cathode follower: why the bad rep, and when sounds good?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
> 1/gm?

Of course. Internet line noise.

> With resistive loading you see mostly second

Adjust the top resistor, trim the DC voltages, 2nd will null. (Well, in SPICE with exact-match pairs....)

> keep the plate voltage constant, the growth of high-order distortion products is drastically reduced

The WCF works best when both top resistor and load are very small. In that sense, Vp-k is fairly constant.

The flip side is that your power supply "impedance" tends to be at least Mu (amp factor) higher than load, and efficiency is Mu times lower than perfection even before you allow for sine waves. It's awful lossy.

> Why not use a choke?

If you allow a choke, you may as well allow a transformer. Given a perfect transformer, which will support plate impedance, the WCF is an awful alternative. White wuz workin with pulse/video signals, 50CPS-5MC or whatever, That's tough at 50R:50R, impossible at 1K:50R. We don't use transformers for video if we can avoid it; when we must, we have to work them far below plate impedances. In audio, transformers are nominally acceptable, though fussy audio gets fussy about its iron, and cheap audio has a hard time buying audio iron on today's market.

Allowing just a choke, the comparison is between two triodes wired as: 1) a WCF working at 200V 10mA, and 2) parallel CFs working at 100V 20mA. Now it is a fair fight. The power consumption is the same. The WCF output impedance is near 1/(2*Gm), and so is two CFs working at the same current in each unit. The output large-signal is grossly the same. The 2*CF will show large 2nd for low-Z loads, the WCF can do better. The WCF's power demand (higher voltage lower current) may fit system power better. The WCF does not want a choke.

> The plate voltage would be essentially constant due to the low DCR of the choke.

No. Set the tube idle voltage to 100V. Swing 10V onto the load. The tube's Vp-k will swing 90V to 110V. Or for big fun, try to swing 40V peak load: tube Vp-k swings 60V-140V. A kiddie explanation is that tube's Rp changes with Vp-K, and that changes the total load the tubes are driving, and causes distortion. I see more insight into how Gm is varying (a lot) and what factors work to reduce the nonlinearity this implies. But whatever: if you are swinging voltage to the load, you are swinging voltage drop across the tube.

> something like adding an emitter follower (or source follower) to drive the anode of the upper tube?

Gimmicky bootstrappers, which were silly in 1955 and IMHO fairly silly today.

It is an easy trap to fall into. Riccardo, a desperate war department, and Rolls-Royce went a similar path with the Crecy engine. Riccardo thought poppet valves limited engine power, and played with sleeve valves. And added 2-Stroke working, a supercharger, direct injection, and stratified charge, and turbocharging was in the works. The test engines could almost keep up with the old-tech Merlin, but the 2-Stroke and porting and blower and lean-burn gave HUGE exhaust flow. When aimed backward, and flown past 400MPH, the jet-effect was good for another 15MPH (on paper) over the Merlin. The obvious next steps would be to increase the turbocharging ratio, feed some of the turbine power back to the crank and prop. And you don't have to go far before MOST of the power is from the turbine. The finely honed pistons are just hot gas generators. Some other guys (Whittle and von Ohain) did the blower and turbine with a tin-can gas generator, and started the Jet Age, and flying fighting pistons vanished almost overnight. Commercial flight also fled rubby/bangy pistons darn quick. And that FVP5 kinda looks, to me, like a Crecy sandwiched by turbines.

If you call Cathode Followers "Miss Piggy", then do an Anode Follower. Use plate-grid feedback to set a grounded cathode stage's gain near unity. The drive, distortion, and impedance are essentially the same as a CF.
 
But the grid on an anode follower is a virtual ground, so the input resistor (think inverting opamp) is the input impedance. Sometimes that's not an issue, sometimes it is.

Peace,
Paul
 
Would anyone mind to comment on the 12AX7 cathode follower circuit used in the Eden WT-series bass amplifier? I would like to know more about the tonal characteristics produced by this particular topology. I posted a thread in The Lab with a partial schematic showing the input section.

http://www.groupdiy.com/index.php?topic=292672&highlight=#292672

BTW, I've since replaced U1 with two AD825s, and the coupling electrolytics were replaced with film capacitors.
 
[quote author="CJ"]Looks like a paralell triode, where do you get CF, or am I on PCP?[/quote]

You may be correct..... I don't know much about tube circuits. However, a couple of people who seem to know what they are talking about said it was a cathode follower.
 
[quote author="PRR"]> [i> The plate voltage would be essentially constant due to the low DCR of the choke.

No. [/quote]

Poor choice of words... I meant, the plate supply would be essentially constant, assuming a stiff voltage source at the other end of the choke.

Here's an idea. Since, in a classic WCF, the drive to the lower grid is quite a bit lower than the input at the upper grid, it seems it'd be quite practical to drive the WCF asymmetrically from a common-cathode amplifier with an unbypassed cathode resistor--using this cathode to drive the lower grid--and eliminate the WCF plate resistor and the coupling from upper plate to lower grid.
 
The Eden circuit is indeed a cathode follower, if a peculiar one. The plates of the parallel triodes receive regulated DC from the emitter-follower regulator, and have no signal connection to subsequent stages. The signal instead is taken from the cathode, so it's a cathode follower. But a very weird circuit nonetheless.

Peace,
Paul
 
[quote author="pstamler"]The Eden circuit is indeed a cathode follower, if a peculiar one. The plates of the parallel triodes receive regulated DC from the emitter-follower regulator, and have no signal connection to subsequent stages. The signal instead is taken from the cathode, so it's a cathode follower. But a very weird circuit nonetheless.[/quote]

Thanks Paul,

Any comments on the sonic signature of such a circuit, distortion characteristics, etc...?

From what I can tell, it doesn't seem to contribute a whole lot to the sound of the preamp. Many people have said that changing tubes doesn't have much effect on the sound. Personally, I've noticed some subtle differences between tubes. The stock Eden tube seems to have a bit more clarity and extension in the highs, but mine is now microphonic. Replacing the op amps and coupling capacitors in the rest of the preamp helped open things up considerably.
 
[quote author="NewYorkDave"]
Here's an idea. Since, in a classic WCF, the drive to the lower grid is quite a bit lower than the input at the upper grid, it seems it'd be quite practical to drive the WCF asymmetrically from a common-cathode amplifier with an unbypassed cathode resistor--using this cathode to drive the lower grid--and eliminate the WCF plate resistor and the coupling from upper plate to lower grid.[/quote]

The WCF has the property, to some extent, of adapting to load Z "automatically". If the load Z is well-characterized then different schemes such as your idea, tuned for the known load, would work.
 
> in a classic WCF, the drive to the lower grid is quite a bit lower than the input at the upper grid

When well-optimized, the two grid-cathode drive voltages are equal but opposite. It self-adjusts very nearly to "true push-pull" operation.

That's at low load impedances. For high load impedances the signals are very different, indeed it becomes a push-push amplifier, and loses all advantage. You could do as well, maybe better, with one triode.

The "classic" White is often drawn with the top resistor very much larger than 1/Gm. This makes the bottom grid drive much larger than the top (the bottom plate pushes the top cathode up and negates the input voltage). Aside from loss of push-pull action, the over-large top resistor is dead-loss as far as large-signal work goes.
 

Latest posts

Back
Top