If an amp has 3 stages, and that amp has 3 cathode bypass caps defining 3 time constants in that amp, then how is that different from a 3 stage amp that has RC coupling between its 3 stages but no cathode bypass caps?
The DC-coupled amp has 3 time constants defined by the RC in the cathodes.
The RC coupled amp with no cathode bypass caps also has 3 time constants.
Is blocking distortion in an overloaded stage from C bypassed cathodes similar to blocking behavior from RC coupling?
Generally speaking, is DC coupling only an advantage if you can get rid of the cathode bypass caps too?
--
The DC-coupled amp has 3 time constants defined by the RC in the cathodes.
The RC coupled amp with no cathode bypass caps also has 3 time constants.
Is blocking distortion in an overloaded stage from C bypassed cathodes similar to blocking behavior from RC coupling?
Generally speaking, is DC coupling only an advantage if you can get rid of the cathode bypass caps too?
--
Cathode bypass caps create 2 time constants each, as they define a shelf not a rolloff. You are comparing apples with oranges.
However, both coupling caps and cathode bypasses can create problems during overload. The solution is to avoid overload, and have appropriate component values.
However, both coupling caps and cathode bypasses can create problems during overload. The solution is to avoid overload, and have appropriate component values.
True. I am. I'm not so much worried about the actual frequency response issues as I am concerned about overload behavior, especially blocking.
That's true as well. You're correct. However, when designing something like a 7 watt push-pull 2A3 amp (as an example), overload is all but inevitable in real world use.
My question was, understanding that the circuit is going to go into overload from time to time (and disregarding subsonic frequency response issues), whether there's any advantage to DC coupling over AC coupling, if DC coupling is going to involve cathode bypass capacitors somewhere.
--
Blocking is mainly due to coupling capacitors. Cathode bypass caps play a much smaller role. In ideal world, therefore, you would do with the DC coupling option.
How do you overload a DC circuit? If that's going to happen you made a bad decision in your calculations/ design choices. Prepare for the worst case and think in terms of an voltage envelop passing your circuit.
Here's an actual design I build yesterday:
The input voltage must fit the driver bias, the amplified voltage again must fit the output tubes bias. If the output tube runs stuck, actions must be taken to present a lower input voltage to the driver.
Further refinements can be had by examining what happens when grid current occurs, what the effect of clipping does to the audio signal etc.
Here's an actual design I build yesterday:
An externally hosted image should be here but it was not working when we last tested it.
The input voltage must fit the driver bias, the amplified voltage again must fit the output tubes bias. If the output tube runs stuck, actions must be taken to present a lower input voltage to the driver.
Further refinements can be had by examining what happens when grid current occurs, what the effect of clipping does to the audio signal etc.
Last edited:
With music signals in a real world amplifier with limited output power into a real world speaker load, overload can happen pretty easily. It's just one of those things.
I think you answered my question by showing a DC coupled circuit. (And that's an interesting way to bypass the large-value cathode resistor on the output 2A3.)
If overload is never encountered (and in a *properly designed circuit*), I suppose concerns are then mostly about frequency response, phase and such. That's where the finer points of capacitor 'sound quality' and all that come in.
If overload occurs in an RC coupled circuit, the time constant RC will need to be adjusted to minimize the period of time the signal is 'stuck.' That means shorter time constants, so possibly high low frequency pole.
If overload occurs in a DC coupled circuit, it appears that blocking is not a big concern, but all the other problems like grid current reducing input impedance of the driven stage, power supply noise under high load, etc. are still to be dealt with (of course).
That would be a volume control to control the input level to the input stage.
I suppose it's worth it to strive for the ideal of a DC coupled amplifier with a rock solid power supply, and no cathode bypass caps (if possible). At least it looks that way to me right now...
--
Last edited:
Assumptions are one thing, getting your hands dirty is another 😉
It's always necessary to compromise on one aspect or another. An unbypassed cathode for instance raises output impedance, which can conflict with the desired BW, interacting with the next stage input capacitance. So an extra stage might be called for, buffering the previous. This too plays a role in response.
The 50ms I mentioned is under the threshold of critical hearing, any hickup shorter than this period will go unnoticed.
It's always necessary to compromise on one aspect or another. An unbypassed cathode for instance raises output impedance, which can conflict with the desired BW, interacting with the next stage input capacitance. So an extra stage might be called for, buffering the previous. This too plays a role in response.
The 50ms I mentioned is under the threshold of critical hearing, any hickup shorter than this period will go unnoticed.
Of course are these small amps very limited in output power. Best results can be had when paired with an appropriate loudspeaker. Hence my aperiodic full range speakers which put the least strain on amp power.
When overdriven the above circuit plays very loud but not to a degree the distortion is unbearable. I understood the rise of 2H is masking higher order product, explaining why musical peaks appear natural although distorted. This will probably add to the popularity of 2A3.
I've listened to many SETs and the character of 2A3 is not my favourite compared to 45, RE604, AD1, PX4. The livelyness of low notes appear artificial to me. Picking a not so linear driver for compensation (like in the Darling) is not my cup of tea. So, the only way out for obtaining better peaks is the use of feedback. I'm there at the moment 🙂
- Status
- Not open for further replies.