I'm not a great tube freak so can anyone tell me the pros and cons of Cascoded tube front ends? So far I have noticed lower gain than some circuits and a tendency for negative waveform clip and lower noise.
I recently purchased and put together a Cascoded 6sn7 tube driver board for a Dynaco. I noticed a lack of life in addition to sounding just so so.
Was this just a **** poor design or is this a common fault to be found with this circuit?
SY?
I recently purchased and put together a Cascoded 6sn7 tube driver board for a Dynaco. I noticed a lack of life in addition to sounding just so so.
Was this just a **** poor design or is this a common fault to be found with this circuit?
SY?
Like a Pentode's screen, a cascode stops V1's control
grid from seeing intrinsic negative feedback from the
other plate that drives the actual load.
Like a Pentode, you can run a cascode in any number
of ultralinear configurations. Or feedback directly to
V1's grid as O.H.Shrade (of RCA fame) did in 1938.
You have a lot of options to shape the curves to your
personal liking.
grid from seeing intrinsic negative feedback from the
other plate that drives the actual load.
Like a Pentode, you can run a cascode in any number
of ultralinear configurations. Or feedback directly to
V1's grid as O.H.Shrade (of RCA fame) did in 1938.
You have a lot of options to shape the curves to your
personal liking.
The remark about poor PSRR is SPOT ON.
Cascodes have a very high O/P impedance. The gain of a cascode is approx. = (lower element's gm) (net ZL). The 'SN7 is a superb tube, but its gm is (IMO) only marginally adequate for service in a cascode.
Cascodes have a very high O/P impedance. The gain of a cascode is approx. = (lower element's gm) (net ZL). The 'SN7 is a superb tube, but its gm is (IMO) only marginally adequate for service in a cascode.
PSRR is near perfect, as long as you like the noise
reference common to B+ rather than ground.
Fine if you intend to ultrapath your output stage
cathode back to B+, then helps PSRR that the
drive from the cascode is also common to B+.
-------------------------------------------------------
But thats not the only way to abuse a cascode.
Here we ultralinear a cascode to compute 2xMu,
and then Mu follow. PSRR common to GND.
The parallel 6dj8's merely a scheme to have
intrinsic feedback of V2 correct microphonics
of V1 and vice-versa. Probably A and B triodes
of entirely different envelopes. No attemt to
match.
reference common to B+ rather than ground.
Fine if you intend to ultrapath your output stage
cathode back to B+, then helps PSRR that the
drive from the cascode is also common to B+.
-------------------------------------------------------
But thats not the only way to abuse a cascode.
Here we ultralinear a cascode to compute 2xMu,
and then Mu follow. PSRR common to GND.
The parallel 6dj8's merely a scheme to have
intrinsic feedback of V2 correct microphonics
of V1 and vice-versa. Probably A and B triodes
of entirely different envelopes. No attemt to
match.
Attachments
burnedfingers said:
PROS:
* Higher gains.
* Greatly reduced Cmiller
CONS:
* Worse PSRR -- Needs clean DC
* Output swing is thin compared to Vpp
* Hi-Z output -- needs to work into a Hi-Z, Low C load.
Got a schemo of the circuit? Here, we get into all that subjective opinion concerning how the thing sounds. You're not gonna get much consensus on that score, and I'm sure there are others out there who say it sounds just fine, and who'll rave about the improvement it made to their Dynaco's.
The cascoded front end I did was a cascoded LTP voltage amp/phase splitter. Of course, a balanced topology is going to get rid of lots of distortion that a SE won't. The cascoded LTP worked out just great, and made for a great sounding final project.
A cascode does not necessarily have to use the same device on top and bottom. You can use a lower Z tube at the base, and feed it a portion of B+ from a resistor to give it the desired op point.
The 6DJ8 is designed with cascode service in mind.
The 6DJ8 is designed with cascode service in mind.
Darn skippy those devices don't have to be the same.
Course, the current behavior tends to be dominated
by the lower device, and voltage behavior by the top.
Unless you impose some feedback scheme that sais
otherwise.
Course, the current behavior tends to be dominated
by the lower device, and voltage behavior by the top.
Unless you impose some feedback scheme that sais
otherwise.
Schematic? Maybe something obvious is there. A cascode can be quite excellent; Joe Curcio's ST70 (for example) is very highly regarded.
You might omit the 620R between upper und lower tube to gain more juice, its purpose is local feedback.
KaDe said:
No, it provides bias to the upper tube. Take a closer look.
Note that the operating point is pretty far from optimum for a 6SN7 - can you provide some measured voltages, if nothing else the voltage across the bottom 620 ohm resistor?
I ran it about 6 hrs and removed it. I have voltages around here somewhere. I will look for them.
nice one Kade,
so the 12AU7 has a transconductance of 1600 at ~50volts and 2mA plate and a stage gain of 160....this is a self-bias cascode...
so the 12AU7 has a transconductance of 1600 at ~50volts and 2mA plate and a stage gain of 160....this is a self-bias cascode...
Kevin's got it- the 6SN7 is a poor choice for cascode service. The lower tube is looking into a very steep loadline.
burnedfingers said:
6DJ8 comes immediately to mind..
Depending on how the tubes are arranged a 6SL7 mu follower might fit the bill as well.
Local feedback with minimum phase shift could and should
have been utilized to linearize that first cascode stage.
620R cathode resistors serve to provide bias, but totally
insufficient by themselves to assure any sort of linearity.
Rather than allow gross non-linearity to further propagate.
Phase shift the spectra of distortion products with reactive
coupling. Distort and shift again. Then hope that a strong
global negative feedback is going to fix everything. Yeah,
right...
Also perfect example of a circuit with the poor PSRR that
others have been talking about. The phase splitter that
follows is not going to appreciate hearing any grid noise
common to B+.
have been utilized to linearize that first cascode stage.
620R cathode resistors serve to provide bias, but totally
insufficient by themselves to assure any sort of linearity.
Rather than allow gross non-linearity to further propagate.
Phase shift the spectra of distortion products with reactive
coupling. Distort and shift again. Then hope that a strong
global negative feedback is going to fix everything. Yeah,
right...
Also perfect example of a circuit with the poor PSRR that
others have been talking about. The phase splitter that
follows is not going to appreciate hearing any grid noise
common to B+.
- Status
- Not open for further replies.