Hey Pieter, this subject was forked into a separate topic:
http://www.diyaudio.com/forums/tubes-valves/309296-common-cathode-choke-discussion.html
You'r welcome to contribute your thoughts/experience, will be appreciated.
Tell us more detail about clipping.
http://www.diyaudio.com/forums/tubes-valves/309296-common-cathode-choke-discussion.html
You'r welcome to contribute your thoughts/experience, will be appreciated.
Tell us more detail about clipping.
i have had very good results with cathode resistor bias,
slight loss of power is peanuts compared with fixed biasing...
and the sonic rewards....astounding...
Could you please post the revised schematic? It seems there are quite a few changes from Pimm's version.
I don't think "holier than pope" attitude applies here. Tony's common cathode choke idea is viable and good. I believe it is doable, and since it is functionally similar to Gary's common CCS, it does not deviate from the Tabor concept. Separate CCSes for each output tubes would be more of a deviation.
Please don't add to the confusion.
The common cathode idea is not Tony's, but mine, see #10 🙄
Tony has no intention to try a choke (see #33) but will start with cathode resistors, and eventually a solid state current source.
I'm trying to understand this. Class A push pull is different from a differential push pull -class A I suppose- in what? The differential with one input to ground will produce two antiphase signals while class A push pull won't? Or the difference is in the way it is loaded. A CCS will turn the load line from sloped to horizontal. That would make significant difference at full power -clipping- but at lower levels, distortion -mostly even order- is canceled in both push pull types or is it?
A differential amp is supposed to draw constant current from the psu. In reality things prove not to be so ideal. A differential voltage amp definately needs a tail CCS to do this. But what is happening in the power stage given the presence of the OPT as a load?
Last edited:
The difference is strict differential (CCS tail) vs. not so strict (resistor tail). CCS enforces symmetry even if halves of a differential stage are not exactly equal. Without CCS, symmetry will require identical tube characteristics and output transformer that retains symmetry in the whole frequency range. Without CCS, Class A gradually slides into AB as signal increases, whereas with CCS it is either Class A, or hard clipping when signal exceeds Class A current capability.
that is why i like the cathode resistor bias....
but who would want to drive his amp to clipping? not me...
but who would want to drive his amp to clipping? not me...
that is a handful right there.....
the object is to keep the tail current a high impedance in order to reap the benefits of long tail pair/differential amplifier operation....
That is more applicable to a differential voltage amp, but not really the case of a differential power amp when the plates are still loaded with an OPT.
that is not the case here...the load is the speaker and reflected impedances at the opt primary and is never a horizontal one...
Yes but keep in mind that this is the final stage which sums the two opposite signals and cancels residual distortion. Of course, the better the symmetry the better the result but we are talking about class A meaning rather hot biasing let alone that the tubes have to be equalized somehow since this is a requirement for the OPT.
Actually this is my query but with regard to the function of the OPT and the final diff to SE conversion. To put it roughly, if we had an OPT that operates ideally and two absolutelly matched tubes then a CCS would be necessary as it is necessary in a diff voltage amplifier even with equal halves?
A CCS anywhere in the chain will set a stable current consumption isn't it? A non linear plate load may force the tube to do other things, perhaps a different voltage swing, I don't know.
I don't know if you are missing this but the signal has only one path to take - which is down through the cathode of one side and up to the plate of the other. The current has literally no where else to go so the net result is that the current going down one side (the signal) must be equal and opposite to that going up the other side. The function of the CCS is to cut off the usual path to earth and force the signal to seek a path through the plate of the opposite valve and ultimately to earth through the power supply. It cannot in of itself eliminate distortion but the signal cannot be different from one side to the other and so the distortion is reduced and of equal character in each half of the PP pair.
Ultimately the load presented by the speaker/OT is so much smaller than the load presented by the cathodes (an opposite situation to normal PP pairs) that it must come to dominate the overall character of the load.
A normal class A amplifier operates its valves in their most linear curve and so tends to have lower distortion than a class AB operation, but the differential class A operation tends to average the residual distortion of the class A between distortions which tend to cancel each other out. This will have the effect of significantly reducing the strongest distortion component which is nearly always second harmonic, making the differential output stage intrinsically lower distortion than the class A which has a lower canceling ability. It has all of the benefits of class A PP and a considerable amount more.
This is why dropping the CCS from the output is inevitably going to result in higher distortion within the main distortion generating component of the amplifier, wasting the full potential of the design to no advantage.
Shoog
Last edited by a moderator:
The tubes don't have to "equalized" other than setting the idle bias (or DC balance), the CCS takes care of the rest - that's one of the key advantages of the differential power amplifier stage.
If everything is perfect, then the CCS would be redundant, but in practice, is that even remotely possible?
Of course, when used with a speaker, the load line is never straight but elliptical, but our point was that the load line has a slope - it isn't horizontal.
Thanks Shoog! Allthough I couldn't express it that clear this is how I also understand things about class A PP / diff. But indeed I missed something about the CCS function. So, a CCS will converse a diff stage to closed system for the psu but inside the system which includes the plate load each valve won't operate at fixed current.
Edit; Thanks Jazbo8 too! I think I've got it know.
Edit; Thanks Jazbo8 too! I think I've got it know.
Last edited:
In its amplifier there are much more powerful tubes 4-1000A tubes! http://www.tubebooks.org/file_downloads/bereskin_3kw.pdf
Last edited:
there are always two ways to look at any amplifier circuit,
first is the dc or idle operating points, what is happening to the circuit at idle and powered up,
here all stages are dc biased to the operating point, we need to look at each device/tubes and see that their rated parameters are never violated.
like cathode currents, plate voltage, and tube dissipations, if no tubes are stressed beyond their limits, then we have a viable circuit...
next comes the hard part, the ac analysis, what the circuit does in the presence of an input signal...
i have to admit that even myself find it hard to understand at times....
it is easy to mix the two and even get more confused...
Thanks Tony! I need to apologize for bothering all you with my idiosyncratic way to understand electronics. I'll be waiting for your project. Please save some space so that you could do a with/without CCS test. It would be very interesting!
Unfortunately without some sophisticated test equipment that comparison will be subjective at best and wont clarify much at all.
Shoog
- Status
- Not open for further replies.