"I won't go into details, but this is NOT good if the intention of the designer is low distortion and somewhere close to linear response."
But that's not the goal, there are better designs/tubes for that.
But that's not the goal, there are better designs/tubes for that.
In my opinion the 82 works fine at around 8 mA, little bit hot with 300Vdc 🙂
So the Rk will be 470-560 ohm, C=220uF
Th Ranode will be 22 kohm 3 watts
On CF the 1 k will be 470-560R
And 10k will be 15kohm 3 watt
Cout =4,7 uF
Rout=220k
So the Rk will be 470-560 ohm, C=220uF
Th Ranode will be 22 kohm 3 watts
On CF the 1 k will be 470-560R
And 10k will be 15kohm 3 watt
Cout =4,7 uF
Rout=220k
Just a suggestion, another approach is to place the volume control in between input gain stage and cathode follower. 250K pot should work. The extra blocking cap can add more "color" if that's the goal. I would also add a series resistor, not in schematic, at the output or after cathode follower, like 100R, to avoid resonance.
Will it work? Yes. Is it a good idea? No.
The input volume control is the simplest solution for a reason. 😉
The input volume control is the simplest solution for a reason. 😉
Yeah… A complete re-design is a solution. But if I did it, the ‘82 would get replaced.
Everything useful has its own downsides. The downsides of a gain control at the output havw been covered; so here is my list of upsides:
1) attenuates the noise (tube rush and hum, mostly)
2) attenuates microphonics (a particular problem with DHT preamps)
3) solves the problem of using high-power/high gain power amps with high-sensitivity speakers
For this specific circuit, the cathode follower running at only a few mA will have a hard time handling the low 10K load if the source level is high. Something like a 6DN7 or 6CM7, with the beefy section serving as the cathode follower and running 10-30mA, would be more robust.
I do like directdriver's suggestion of an intermediate attenuator; the high feedback of the cathode follower makes it more immune to noise and microphonics than the first stage. But the ultimate is to use two level controls, one at each end, enabling the end user to experiment with varying levels of "color".
1) attenuates the noise (tube rush and hum, mostly)
2) attenuates microphonics (a particular problem with DHT preamps)
3) solves the problem of using high-power/high gain power amps with high-sensitivity speakers
For this specific circuit, the cathode follower running at only a few mA will have a hard time handling the low 10K load if the source level is high. Something like a 6DN7 or 6CM7, with the beefy section serving as the cathode follower and running 10-30mA, would be more robust.
I do like directdriver's suggestion of an intermediate attenuator; the high feedback of the cathode follower makes it more immune to noise and microphonics than the first stage. But the ultimate is to use two level controls, one at each end, enabling the end user to experiment with varying levels of "color".
For this load like, the bias is around maybe 2.8V headroom. The cathode resistor should be around around 375 ohm for 8mA... I would seriously suggest considering a selected green LED for this instead of the cathode RC network. 😉
This is also perhaps one of the much more linear regions in the 'Generous Electric' published plate curves, but on purpose I did not want to replace the load resistor since it has a significant effect on the signal amplification. I just wanted to suggest a modest fix with not too much effort.
About the best area on the ECC82/12au7 plate curves is running 5-8mA around 250V on the plate with a gyrator (or CCS). But then you need something like a 380V+ supply. I did a mod very similar to this on a 'highly regarded' German name-brand amplifier (which was actually made in China) for a friend and they were very happy.
They did not want to change the valve since it was printed on the top plate of the amplifier, etc.
Last edited:
I assume that the choice of tube and operating point used makes the largest contribution to the sonic signature / "color". And Stephe has been clear that she is not concerned with those aspects of the design, however flawed they might be. So I'll try to focus on her actual question.
What I'm not so sure about is how much of the "color" is influenced by the fact that the gain section is being driven by an unattenuated signal. No doubt it has some effect but I don't know how much.
If the lack of attenuation is not a significant factor, then moving the volume control to the input seems like the better choice. But, if the unattenuated signal makes a major contribution to the "color", then keeping the volume control on the output would be better.
As I see it, the biggest problem with that is the parts values used in the original circuit. A 10k pot only measures 10k to ground if the pot is wide open, which is not a normal scenario.
I don't have any 10k volume pots but I just measured a 100k attenuator. Like most volume controls, it is an audio taper, which allows for more precise volume adjustment. When turned to maximum the resistance to ground is 100k, of course. But when set to 12:00 the resistance to ground is only 6.4k. So if the taper was the same, a 10k pot would measure ~640 ohms at 12:00.
As others have pointed out, the resistance of the pot on the output is in parallel with the input impedance of the amp and the combined resistance value is used to determine the -3db rolloff point with a high pass filter calculator. So the -3db point varies depending on the position of the pot.
A resistance of 640 ohms in parallel with 10k is ~600 ohms. When combined with a 1uf coupling cap a high pass calculator shows the -3db rolloff point is 265 Hz. And if the volume control is any set lower than 12:00 the -3db point is even higher.
As I see it, this is biggest problem with the original design if we assume that the "color" aspect is a desirable trait. But changing the value of the cap and pot is easy enough.
With a 100k pot (set to 12:00) and a 3.3uf cap the rolloff point would be ~12 Hz, assuming use with a SS or Class D amp whose input impedance is 10k.
If I'm misinterpreting something here, someone please correct me.
And, if having only a single control is not a requirement, then the same thing could be done with the circuit above while leaving the tube operating points alone.
What I'm not so sure about is how much of the "color" is influenced by the fact that the gain section is being driven by an unattenuated signal. No doubt it has some effect but I don't know how much.
If the lack of attenuation is not a significant factor, then moving the volume control to the input seems like the better choice. But, if the unattenuated signal makes a major contribution to the "color", then keeping the volume control on the output would be better.
As I see it, the biggest problem with that is the parts values used in the original circuit. A 10k pot only measures 10k to ground if the pot is wide open, which is not a normal scenario.
I don't have any 10k volume pots but I just measured a 100k attenuator. Like most volume controls, it is an audio taper, which allows for more precise volume adjustment. When turned to maximum the resistance to ground is 100k, of course. But when set to 12:00 the resistance to ground is only 6.4k. So if the taper was the same, a 10k pot would measure ~640 ohms at 12:00.
As others have pointed out, the resistance of the pot on the output is in parallel with the input impedance of the amp and the combined resistance value is used to determine the -3db rolloff point with a high pass filter calculator. So the -3db point varies depending on the position of the pot.
A resistance of 640 ohms in parallel with 10k is ~600 ohms. When combined with a 1uf coupling cap a high pass calculator shows the -3db rolloff point is 265 Hz. And if the volume control is any set lower than 12:00 the -3db point is even higher.
As I see it, this is biggest problem with the original design if we assume that the "color" aspect is a desirable trait. But changing the value of the cap and pot is easy enough.
With a 100k pot (set to 12:00) and a 3.3uf cap the rolloff point would be ~12 Hz, assuming use with a SS or Class D amp whose input impedance is 10k.
If I'm misinterpreting something here, someone please correct me.
Interesting . . . that's what I did with the preamp I built a while back. I used a pair of 100k mono pots on the input, because I wanted the ability to adjust the L-R balance, and a 100k stepped attenuator on the output. As I recall, when I posted that somewhere a couple of people were quite critical. Maybe I'm not a total idiot after all?
And, if having only a single control is not a requirement, then the same thing could be done with the circuit above while leaving the tube operating points alone.
Slamming the front end without attenuation creates some form of "compression" much like a guitar amp and makes the sound more jumpier or more jump factor. The cathode is sitting at 5.8V so the gain stage is not likely going to be clipped. I think the original design had subtle dynamic compression in mind besides "color." Obviously any preamp needs attenuation otherwise it's not a preamp so placing it at the output is one option. I think in the context of this, placing the volume control in the middle is the sweet spot, a happy compromise. Why not have two volume controls, in the front and in the middle just like a guitar amp?!
Does it even need that second cap? I think this mid point design might keep the original intent of this pre-amp without some of the short comings of the output volume control that have been pointed out. The resistor on the output could then be a high value and let the amp being driven set the load on the cathode follower.Just a suggestion, another approach is to place the volume control in between input gain stage and cathode follower. 250K pot should work. The extra blocking cap can add more "color" if that's the goal. I would also add a series resistor, not in schematic, at the output or after cathode follower, like 100R, to avoid resonance.
Stephen: "Does it even need that second cap?"
The cathode follower's grid has DC from the cathode so you do need a cap to block the DC otherwise it will upset the CF's bias and create noise on the pot. Marantz 7 and other vintage preamps used similar approach so it's nothing new.
The cathode follower's grid has DC from the cathode so you do need a cap to block the DC otherwise it will upset the CF's bias and create noise on the pot. Marantz 7 and other vintage preamps used similar approach so it's nothing new.
I cannot agree. This is one of those positions (i.e. that above on volume control placement) that gets repeated ad nauseam until people take it for granted. And as for the "professional" term. I am a professional with over 33 year professional experience designing both aircraft avionics systems and audio gear. And I make topology decisions based on requirements, not habit, tradition, or rules.
Don't get me wrong, I understand why the opinion stated by tubetvr is so common. And I don't fault people for their opinions or how they design their own circuits. But in this instance, the placement of the volume control was indeed intentional and used to preserve and enhance the nonlinear characteristics of the gain stage. I know that many will read the statement and say WTF? Well I have a good explanation of volume control placement at this link: https://www.cascadetubes.com/2022/01/17/about-that-volume-control/
The salient part of the 12AU7 "Color" Preamp design is, in fact, the "color" part. For those not quite getting this, I suggest reading the origonal posting on the preamp carefully. Including the statement:
And as for the frequency response issue. I suggest this post on high frequency response with volume controls in the loop: https://www.cascadetubes.com/2022/03/31/volume-controls-and-high-frequency-response/. This should clear up how high frequency response is affected by cascaded stages and volume controls.
I just wanted to make sure people understand the reasons for how the circuit in question was designed.
For the alternative design approach, minimizing distortion, i recommend looking at this schematic.
... I write my name "Stephe" because that is my name. IDK why people want to add letters to it.
The more caps - the worse the sound... 😉
Why not do a simple grounded cathode with a directly coupled cathode follower as in this example: https://darklanternforowen.files.wordpress.com/2014/01/jelsc-linepre1-b.jpg
And OK - if you like, slap a 100K pot - or stepped attenuator - on the output.
And yes, this will work with a 12AU7 as well.
And 6SN7.
Pro tip: Mills wirewound resistors here. Great preamp article here: https://worldradiohistory.com/Archive-All-Audio/Sound-Practices/Sound-Practices-Issue-13.pdf --- see page 27.
Why not do a simple grounded cathode with a directly coupled cathode follower as in this example: https://darklanternforowen.files.wordpress.com/2014/01/jelsc-linepre1-b.jpg
And OK - if you like, slap a 100K pot - or stepped attenuator - on the output.
And yes, this will work with a 12AU7 as well.
And 6SN7.
Pro tip: Mills wirewound resistors here. Great preamp article here: https://worldradiohistory.com/Archive-All-Audio/Sound-Practices/Sound-Practices-Issue-13.pdf --- see page 27.
A less obvious potential benefit of having the volume control located at the output is that the output signal’s distortion character does not change with volume level. Affording a more consistent sound character between high and low volume settings. Yes, I know that THD is reduced for lower volumes with the control located at the input. So, if minimum percent distortion is the goal of the circuit, locate the volume control at the input.
Your explanation, "About that Volume Control" concludes with the statement:
"This configuration should be acceptable for use with almost any vacuum tube power amp with an input impedance of approximately 100kΩ or more."
So clearly this design is unsuitable for use with SS or Class D because their input impedances are typically in the range of 10k to 20k. And, from what I've seen, probably 90% of those interested in adding "tube sound" / "color" / "warmth" to their systems are using SS and Class D amps.
So it's a pretty safe bet that the person who asked Stephe to build and measure this design plans to use it with a SS or Class D amp. Those who use tube amps, whose input impedances are typically 100k, generally find that their tube amps already possess "tube sound".
I'm not technically so well informed, but my understanding is that the primary factor that determines the output impedance of a single gain stage is the plate resistance of the tube. If the plate resistance is too high the most commonly used method to reduce the output impedance is to add a cathode follower stage.
But even after doing so, your calculations indicate that the output impedance, though considerably lower, is still not low enough to be suitable for a SS or Class D amp. And, based on your comments, that was probably never your intention.
A far simpler approach would be to start with a low mu tube that also has a low plate resistance. Then there is no need to add a cathode follower and a single gain stage can drive SS / Class D amps as well as tube amps.
Last edited:
I don't understand the justification, it is "color"? So why not just add a tube-based equalizer, tilt or tone stage if the whole justification of this preamp is "color" (whatever that means). Then one can actually adjust that "color" in a predictable/designed manner, instead of relying on a particular tube or manufacturing quirk, or tube brand or manufacture date, etc. to bring it, the "color".
Speaking for myself, ‘color’, is tonal gradations and contrast. For whatever reasons, some systems simply sound like they reveal more instrument tonal variation and contrasts. Some systems, obviously so. Which makes the playback of music more interesting, and more emotionally involving. In my experience, this is not an effect due to an intentional non-flat frequency response. Therefore, tone controls or EQ aren’t the solution. Applying a purposely non-flat FR instead tends to overlay the same or consistent tonal coloration for all source material, and is audible as doing such. As a result, playback sounds more un-involving, more fake, and less musically engaging.
He was explaining the other single tube pre-amp with a 250K pot on the output and no cathode follower.
https://www.cascadetubes.com/the-universal-preamp/
He came up with this other "Color" pre-amp to overcome those issues, to be able to drive a SS amp.