Cathodyne phase splitter with CCS?

[andrew_whitham]

cathodyne phase splitter with CCS?

A request this one.

Currently scheming out a P-P amp having ignored this topology so far...

I plan to use a cathodyne phase splitter direct coupled to the input triode a-la williamson I think. Ok there's nothing odd there.

I'm intending to baby it a little by using CF's to drive the (potentially paralleled) output tubes, so at the moment I'm not tremendously worried about odd loads upsetting balance.

What I am concerned about is distortion, fave method for knocking that on the head is with a constant current source or sink.

and finally I get to my point - HOW?

Is this even possible? The CCS would be semiconductor based, to keep V+ sensible, But adding them above or below the tube - the increased resistance (Ra or Rk whichever) would appear to mess up the balance.

presumably I cant add one on both sides of the tube (can I?)

I really can't get my head around this .My usual trick of trawling the internet to find out whats been done before has so far come up blank. Perhaps its just a dumb question

Any links or information would be really appreciated

Thanks

Andy

 

[ray_moth]

No genuine question is dumb (although some answers are ) I think in this case a CCS cannot help. Balance is the key requirement and a CCS at either end (or both ends) of the concertina would knock balance out of the park.

Distortion is also important but I think the answer to that is probably to use a medium mu triode at fairly high current, for optimum linearity. Use a high B+ rail, so that there is a decent plate-to-cathode voltage. Loading the splitter with cathode followers doesn't hurt. You might be interested in this rather complicated solution in
De Palma's article.

 

[Geek]

I don't see why a CCS on the cathode and on the anode couldn't give you the best an old cathodyne can offer.... they're close to ideal loads and will allow you to keep your current up, like Mr. Moth suggests.

 

[SY]

Because matching CCS is a whole lot harder than matching resistors?

The distortion is rarely an issue here- degeneration is 100%. I use CCS everywhere I can, and that's one place I wouldn't.

 

[radtech]

Seems to me if you put a CCS on the plate, and assuming the grid current is negligable, the current through the cathode resistor will be constant, so adding another CCS there would be redundant, and like Sy said, you'd have to have them pretty well matched or you're going to violate Kirchhoff's "Seventh and a Half Law of Current Sources"*

I'm not sure off the top of my head what a CCS would do to the balance, you could always try simulating it. I think I'd go with the old tried and true resistor approach though.. if you can't find any examples of using a CCS here there's probably a good reaon.




*"The rotational velocity of Gustav Kirchhoff in his grave is directly proportional to the difference between any two constant current sources in series."

 

[SY]

You also have to have them well-matched to keep the circuit in balance. The beauty of a split load is that the balance is completely determined by the matching of the plate and cathode loads. Resistors are really the best choice in that situation.

 

[Tubelab_com]

In one of my more blonde moments I tried puting a CCS in the cathode and another CCS in the plate of a triode amplifier circuit. The result..... It is called a bistable multivibrator in the digital world. The plate voltage would rail to B+ or near zero, no place in between. One CCS will win. Even with a 10 turn pot to finely adjust the CCS current, it wouldn't work.

Even if you could get the circuit stable, the split load relies on equal load impedances. The CCS is a very high impedance. Now the circuit balance is dependent on stray capacitance and the grid impedance of the following stage. Here resistors are your friend. They swamp out all of that high impedance imbalance stuff.

 

[Brian Beck]

Yes, current sources dueling in series is like voltage sources dueling in parallel. One always “wins”. Kirchhoff and Thevenin roll in their graves.

In one of MY blonder moments, I thought I could balance plate and cathode CCSs with a feedback servo to keep the whole thing from going astable. Then it occurred to me that the feedback made the “CCS” no long a current source. Doh! Resistors to the rescue...

 

[Wavebourn]

In one of MY blonder moments, I thought I could balance plate and cathode CCSs with a feedback servo to keep the whole thing from going astable. Then it occurred to me that the feedback made the “CCS” no long a current source.

Shmidt trigger with feedback thtough an integrator gives nice oscillator!

 

[Geek]

If you want a CCS to help give you perfect balance, you'll probably need to change the topology and include it in a common-mode feedback loop.

Imbalance in this circuit was 20mV P-P with a 60V P-P output. Tube sections weren't even matched.

 

[andrew_whitham]

Wow,

Sorry to dissapear like that, DiyAudio was suddenly unreachable from what passes for the internet here (China) - it happens.

A big thanks to you all for the help, I suppose if its never been done before theres a good reason...

OK So CCS's are not a great idea here and I actually *think* I can see why... this is above and beyond the extra complexity they entail.

So to phrase it another way. I think I'm looking for a method of providing a high AC resistance while still getting power to the tube. That should keep it stable at DC but allow the advantages of a big plate load.

So as an alternative, what's wrong with using a couple of chokes instead of (or to supplement) a resistor pair? OK. theres all the problems associated with choke loading and so on, You need a big (expensive) choke too, but could it work? Would it be beneficial?

On the other hand it is difficult to see how to 'beat' a couple of matched resistors. particularly for what might be very little improvement.

Thanks again

Andy

 

[kevinkr]

Chokes would have to match perfectly at all frequencies in the desired pass band or unbalance would result.

You have pretty much answered your own question..

Stick with resistors or choose a topology like the LTP with a CCS in the tail.

 

[lndm]

Y'know, it seems Williamson had similar concerns in mind when he built the third stage of his amp.

The result..... It is called a bistable multivibrator

LOL

 

[andrew_whitham]

Chokes would have to match perfectly at all frequencies in the desired pass band or unbalance would result.

You have pretty much answered your own question..

wry smiley needed... resistors are a heck of a lot cheaper too.

thanks again I think I'll go work through this (again) reasonably happy that I'm not missing the next bg thing..

If you want a CCS to help give you perfect balance, you'll probably need to change the topology and include it in a common-mode feedback loop.

Imbalance in this circuit was 20mV P-P with a 60V P-P output. Tube sections weren't even matched.

Although I have theological misgivings about this differential topology (the inverting part gong through a second triode) The 20mv mismatch is not to be sniffed at. Would you care to explain how the the CMFB loop is working? Or point to where I might look for some more info.

Cheers

Andy

 

[ray_moth]

Fred Nachbaur's design is quite well explained. He used a pentode for his CCS but the principle is the same. I use a pentode without the CMFB bit and it works OK.

 

[Geek]

Fred Nachbaur's design is quite well explained. He used a pentode for his CCS but the principle is the same. I use a pentode without the CMFB bit and it works OK.

That's the one I derived the SS CCS LTP from

 

[Wavebourn]

Fred Nachbaur's design is quite well explained. He used a pentode for his CCS but the principle is the same. I use a pentode without the CMFB bit and it works OK.

I suppose he put his RV503 and an ampermeter on the front pannel, right?

 

[Geek]

And that means.... ?

 

[Wavebourn]

And that means.... ?

That means HAFB











(Human Aided FeedBack)

 

[ErikdeBest]

Now that you guys are talking about 'conflicting' CCS's, I have a question regarding a schematic by Morgan Jones (fig. 2.49 in the third edition) It is basically, for those who do not have the book, a double triode in LTP configuration, with a CCS on the tail and two CCS on top, one for each triode, the idea being that they achieve a high plate resistance, with just a small voltage drop from the B+.

My question is...won't the two CCS on top conflict with the CCS on the tail? I mean, the current thorugh both CCS at the top should be equal equal equal to the current drawn at the tail (Kirchhoff's law), and adjusting this, as many commented here, is almost impossible to accomplish. What am I missing?

Thanks all

Erik