|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.246.123.70
So...I am trying to get a mental handle on what is happening from a Power Supply perspective during normal operation of a driver stage.Given a triode operating in Class A, with the B+ attached to the anode, the input signal attached to the grid, and a biasing resistor attached to the cathode and grounded. The output is also attached to the anode, with a DC blocking cap to prevent B+ voltage.
1) Is the voltage and current drawn from the B+ (under class A operation) always a fixed amount? Or does it vary, even slightly?
2) Can operation of the triode be thought of as a Voltage Shunt...so that a given amount of voltage is shunted through the tube, and modulated by the grid in response to the input voltage, with the remaining voltage component serving as the "amplified input signal" and traveling through the output cap?
What I am trying to understand in terms of the PSU in a Class A preamp is...does the PSU have to be critically damped to prevent ringing if voltage/current changes? Or, does this not matter, as intrinsically, the voltage/current drawn from the PSU never changes, and thus makes this point moot?
Follow Ups:
This has greatly increased my understanding. I really need to sit down and play with loadlines for the 6SN7 to increase my understanding.Would any of the TubeCad software allow me to do this? I could then easily flip through many tubes/operating points...much faster than on paper.
I think people get confused and/or mislead. A class A stage will draw a fixed average amount but it changes moment by moment. All tube stages in a preamp or power amp, other than the output tubes, are operated in class A (typically).Now a push pull stage, carefully balanced in class A, could draw a true constant current as well as differential voltage amps. A CCS can also force this and can be applied as a plate load or, for push pull and differential, a cathode load. Chokes and transformers can approach the same behavior as a CCS.
I think the way you state things in #2 shows you have an important grasp on things. Just think about what you actually need from the tube and how that is "developed" across some "load" to be presented to the next tube or load.
Good amps have been built with cap input as well as choke input filters. Most would agree that a tube rectifier and no electrolytic caps are the big points. Duncans power supply sim is worth getting to know. Learn about the step function to evaluate ringing. IMHO better to have a little leading edge spike than a rounded corner. But no way you want much ringing.
I am using Duncan's PSUD. Wonderful tool, and highly accurate.I've got a 5AR4 into a cLCLC power supply:
3.3uF -> 15Hy/60DCR -> 120uF -> 7.5Hy/30DCR -> 40uFThere is some 'ringing' when I model it in PSUD, with a 1mA current draw change (39mA vs 40mA). It is not critically damped, but underdamped.
It actually sounds BETTER when I reduce the middle cap to 20uF...which makes it even more underdamped...but I get weird in-out movements of the woofer much more regularly. They still occur, altho very uncommon, when I have the 120uF as the middle C.
And, there is a low-level 10Hz freq on the output, which causes some wobble in the woofers.
I'm trying to track down the source, and tie it back to my knowledge of how the entire circuit operates.
Sounds darn good as-is, tho...
Just a note. When looking at the step response in the PSUD remember that the longest time period that the current will increase (or decrease) from idle is 1/4 wavelength of the lowest freq. of interest. At 20hz that would be 12.5ms. After that the current change will turn around and go the other way.So the behavior for that first 12.5ms is what matters if the circuit is Class A.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The model is not really right. The current does not change suddenly. The current will increase as it follows the input signal over the first 1/4 wavelength, peaking at a 1ma. change (in your example).Tre'
a model that has two steps. We could have a step back down to the idle current 12.5ms after the first step.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Jayme> > 1) Is the voltage and current drawn from the B+ (under class A
> > operation) always a fixed amount? Or does it vary, even slightly?Yes the current varies.
The amount depends on the amount of signal and the load.
Draw a loadline to know for sure how much.< <2) Can operation of the triode be thought of as a Voltage
> > Shunt...so that a given amount of voltage is shunted through
> > the tube, and modulated by the grid in response to the input
> > voltage, with the remaining voltage component serving as
> > the "amplified input signal" and traveling through the
> > output cap?You can think of it as a voltage shunt
or a bias controlled resistor.> > does the PSU have to be critically damped to prevent
> > ringing if voltage/current changes?Definitely
> > Or, does this not matter, as intrinsically, the
> > voltage/current drawn from the PSU never changes,
> > and thus makes this point moot?Only if you are using a CCS on the plate.
A CCS forces the tube to conduct at the same current.
And also provides a high isolation from the PS noise.DanL
DanL,If a CCS is used, forcing the tube to conduct at a set, unvarying current, doesn't this then deviate from the linearity of the response that the loadline describes, as per your first statement?
Or is a triode actually slightly non-linear, and the CCS 'locks' the triode into a more linear response?
PS. Good to chat with you again!
JaymeA resistor or choke load have a slope on
the loadline or variance in the current.
A CCS has a horizontal load line showing
the same current at any voltage.
A CCS usually gives a more linear response,
this depends on the tube's curves of course.
How linear a triode is depends more on the
curves than the load presented to the plate.
The load can fine tune the linearity but
a non-linear tube will be non-linear with
about any plate load you use.
Also to be considered is the operating current,
too low of an operating current will cause
ANY tube to be non-linear.
A loadline is essential to determine
the proper/best operating point and
what you need to use on the plate.
A horizontal load line will keep an otherwise linear tube from operating in the non-linear region found to the right of the operating point. With a diagonal load line, when the current falls and the voltage rises, the tube is allowed to get into the non-linear part of the plate curves where the curves "bunch" up. The voltage does not rise as fast as it should. For example, if a one volt increase at the grid causes a 10 volt decrease at the plate, a one volt decrease at the grid should cause a 10 volt increase at the plate. With a horizontal load line this is much more likely to happen. All one needs to do is look at some plate curves and this will be obivious.Tre'
P.S. Dan, I am more or less just restating what you said.
This post is made possible by the generous support of people like you and our sponsors: