Depends on what type of distortion you mean, and not always, depending on the circumstances. See the 1955 article:
From the foregoing experiments at least one curious result stands out: In a Class A amplifier the use of a bypass capacitor across the output cathode generally reduces the intermodulation distortion, although it may either decrease or increase the harmonic distortion. ... All of the foregoing applies to a Class A amplifier only. In the case of a Class AB amplifier the bypass capacitor is absolutely necessary if the amplifier is to perform within the modern limits of high fidelity performance.
I always have thought that IMD was too often overlooked and critically important to sound quality. Example: inexpensive solid state amplifiers with vanishingly low THD that still sound terrible. There is a lot more going on than THD, and IMD certainly is one of them, among other things.
In class AB, true. However in a class A push pull amplifier it is not true. The PP amp cancels most of the 2nd harmonic content, leaving third and higher harmonics. Odd harmonics are noted for harsh fatiguing effects.
So as I said, it is all trade-offs.
So as I said, it is all trade-offs.
I'm not going to argue about it. The article is quite clear, and their methods systemmatic and accurate. You can argue with the authors if you can go back to 1955. I'll go with their results, which I quoted above for those who don't want to read or do not understand the highly technical article. The question has been asked and answered definitively, I just needed to be pointed to the article to understand. Thanks.
Last edited:
I'm seeing quite a few sweeping conclusions drawn from possible over-simplifications, but that's seemingly always the case in audio fashion. You guys are arguing about two very different situations: a push-pull Class A amplifier, where lack (or not) of a cathode bypass capacitor is debatable (causing ideally no degeneration), and a single-ended Class A amplifier, where lack of a bypass capacitor gives degenerative current-sensing feedback. The latter case increases output impedance, reduces distortion in the output stage, and increases the required drive voltage, increasing distortion in the driving stage. The former case is much more ephemeral.
In either case, since nothing is perfect, even order distortions shift the DC voltage across a bypass capacitor, giving a shifting bias with the obvious time constant. Cancelled pretty much in push-pull at the OPT, but the summed DC shift is still there at the cathodes, even without a capacitor. The time constant is then (without a capacitor) very small, within signal band, giving Mitchell's 1955 results.
All good fortune,
Chris
In either case, since nothing is perfect, even order distortions shift the DC voltage across a bypass capacitor, giving a shifting bias with the obvious time constant. Cancelled pretty much in push-pull at the OPT, but the summed DC shift is still there at the cathodes, even without a capacitor. The time constant is then (without a capacitor) very small, within signal band, giving Mitchell's 1955 results.
All good fortune,
Chris
Last edited:
Yes, that's all I ever asked about (post #4) and it's all the 1955 paper tested, and they cite both sides of the yes-no bypass capacitor argument before testing. It's why they tested.
I'll go with their detailed bench test results, for an assortment of class A push-pull amplifiers, with an assortment of different tubes, and with some conclusions that surprised them (goes against general beliefs). Whether deemed "over-simplification" or not, it's good enough for me. We start with bench test results and the theory and then listen from there.
I didn't realize this meant that you are only interested in push-pull amplifiers where both valves have a common cathode resistor and capacitor. My replies apply to the case (push-pull or not) where each valve has its own cathode resistor and capacitor.
I have updated the original post to be more clear for anyone who reads in the future.
Because I am learning, I am choosing to study the simplest possible amplifier circuits. I realize that things change with other designs. My Dynaco ST-70 series ii and the little 6P14 amp that I am building are pretty simple examples for me to start with. It doesn't get much simpler than these two. One uses a bypass cap and one does not, which is what prompted my question. Then I found a whole bunch of blanket statements online and in other threads whether to always use one or always not use one, which didn't seem right to me.
Because I am learning, I am choosing to study the simplest possible amplifier circuits. I realize that things change with other designs. My Dynaco ST-70 series ii and the little 6P14 amp that I am building are pretty simple examples for me to start with. It doesn't get much simpler than these two. One uses a bypass cap and one does not, which is what prompted my question. Then I found a whole bunch of blanket statements online and in other threads whether to always use one or always not use one, which didn't seem right to me.
Depends on which kind of distortion we like best.
I will leave that to the superior hearing of Batz, Katz, Natz & Ratz,
My own preference, as little D% as possible. The lautsprecher will determine the bottom line in any case.
Don't sweat it, make time for other pursuits. Life is short.
😀
I will leave that to the superior hearing of Batz, Katz, Natz & Ratz,
My own preference, as little D% as possible. The lautsprecher will determine the bottom line in any case.
Don't sweat it, make time for other pursuits. Life is short.
😀
Simple as the circuit may look, it's actually quite complicated. In the configuration you are interested in, the capacitor doesn't do anything under (mathematically simple) small-signal conditions. It is only there for distortion compensation or to allow the circuit to work in class-AB on occasional large peaks. The efficacy of the distortion compensation depends on the actual waveform and how much bias shift that causes, so it is difficult to say anything generic about it.
I suspect that under some simplifying conditions, it may be possible to reduce distortion by connecting a small resistor in series with the capacitor. I'll try to calculate the required value.
I suspect that under some simplifying conditions, it may be possible to reduce distortion by connecting a small resistor in series with the capacitor. I'll try to calculate the required value.
Is a general solution possible? Seems like it would be signal level dependent, at first blush.
All good fortune,
Chris
All good fortune,
Chris
I (we) don't "like" any distortion. Per the article and tests, there are circumstances where IMD improved without increasing THD, contradicting the blanket statements I found online stating "always use" or "always do not use" or "always increases distortion." It contradicted what the experimenters expected and contradicted what now appears to be "generally accepted" online - that it always is a bad idea and is always best when eliminated. Apparently not.
Note to self: find a new friend with an oscilloscope. The shop that used to have one is long gone.
Last edited:
for an Oscilloscope, there are some pretty inexpensive options such as this computer add-on:
2 channel USB scope
You may have to buy 10X probes for high voltage, not clear from the advertisement.
or full stand alone unit:
Rigol for less than $200
2 channel USB scope
You may have to buy 10X probes for high voltage, not clear from the advertisement.
or full stand alone unit:
Rigol for less than $200
Last edited:
This is a mostly fix-bias amp. The 15.6 Ohms in the big Dynaco is not enough to matter. The Gm of EL34 at this current is about 8 mA/V. So the internal cathode resistance will be 1/8 mA/V or about 125 Ohms. Another 15 Ohms of resistance is very little difference.
The Philips datasheet, under "class AB", shows 130 Ohms to actually bias a pair of EL34. Compared with 125 Ohms internal cathode resistance, gain would fall. In common-cathode push-pull, not much. (In single-ended, to about half; but SE was excluded on post #4.)
Ping.
Going to have to read through this several more times.
In building a 300b SET I have collected a dozen some-odd schematics and plans from around the web. I already own one Tektron amp and see (and hear) what their designers placed between ground and cathode in that circuit (100 uF and 1k resistor) which in my ham radio circuitry mind I see as a simple RC circuit with a value of something like 1.8 Hz (IIRC, I ran the numbers a few days ago), while one or two other circuits I found had much higher value capacitors (or lower-value resistors) bringing this number down to under 1 Hz.
What this actually MEANS in terms of sound I don’t know, but I don’t understand people saying the cathode bypass capacitor-resistor pair doesn’t affect SIGNAL as it’s not in the SIGNAL path, but since electrons (current) flows from ground/cathode to PLATE and then OPT’s, how can the bypass circuit NOT affect signal?
This is mostly rhetoric; no need to belabor the point or get all ”aspy” in replying, like I said, I’ll re-read this post from OP on and figure it out. Just wanted to vent, I guess.
There is SO much information available online regarding this topic and I’m a hobbyist, not an electrical engineer.
NB
Going to have to read through this several more times.
In building a 300b SET I have collected a dozen some-odd schematics and plans from around the web. I already own one Tektron amp and see (and hear) what their designers placed between ground and cathode in that circuit (100 uF and 1k resistor) which in my ham radio circuitry mind I see as a simple RC circuit with a value of something like 1.8 Hz (IIRC, I ran the numbers a few days ago), while one or two other circuits I found had much higher value capacitors (or lower-value resistors) bringing this number down to under 1 Hz.
What this actually MEANS in terms of sound I don’t know, but I don’t understand people saying the cathode bypass capacitor-resistor pair doesn’t affect SIGNAL as it’s not in the SIGNAL path, but since electrons (current) flows from ground/cathode to PLATE and then OPT’s, how can the bypass circuit NOT affect signal?
This is mostly rhetoric; no need to belabor the point or get all ”aspy” in replying, like I said, I’ll re-read this post from OP on and figure it out. Just wanted to vent, I guess.
There is SO much information available online regarding this topic and I’m a hobbyist, not an electrical engineer.
NB
Hi Never Get Old,
A couple of thoughts that will improve the musical quality.
1. I never go by a rule of thumb as I have never found that method to be worthy. For more accurate/natural
musical reproduction, in this case, you will have to test various size capacitors, the optimum depending upon the OPT,
power supply, the design, and other parts quality. (Smaller than optimum ufd will reduce bass, but with more margin to
prevent low frequency oscillation. Over optimum value ufd will tend to give a flubby, tubby bass. (DA, ESR, Internal inductance,
materials, termination techniques are problems one has to deal with.))
2. In your final design, improvement in musical quality will occur if you do not use one large cathode bypass electrolytic capacitor
but multiple capacitors in parallel for desired optimum ufd. Much higher voltage rating than the actual cathode to ground voltage is
also advantageous, especially in case a low frequency oscillation occurs.
3. A final thought; it will take time, but there is no other way to properly optimize the ufd value.
Good luck in your ventures.
pos
Last edited:
There is but it involves equations, knowledge & extensive testing., happily there.s a shortcut - bung in whatever value you've got, listen/test try something else until your happy and the amp is stable with your speakers.
Bypass capacitors and related subjects for a 300B SE amplifier stage:
I attached a discussion.
I hope that will help for understanding of some output stage factors.
Have Fun!
I attached a discussion.
I hope that will help for understanding of some output stage factors.
Have Fun!
I covered such in points 1 and 2; actually there is not due to many variables I pointed out, if one is dealing with high end.
My apology for not being clearer.
Cheers and all the best. 🙂
pos
Last edited: