Can a 12AU7 designed clean gain stage be as loud as a 12AX7?

[GlideOn]

I'm looking to utilize an unused side of a V3 triode in my single channel amp for a dedicated, switchable clean channel (see other thread on how to get the switching)

I currently have my 12AX7 in V3 functioning as a "faux phase invertor" which allows a smidgen extra gain/compression and the ability to have presence, negative feedback in my single ended. It works extremely well here.

The other side of the tube has the heater hooked up already, but no components.

Previously I had thought of the merits of the JJ ECC823 tube which has the 12AX7 on pins 1, 2, 3 and the 12AU7 on pins 7, 8, 9.

I don't hear a lot of feedback regarding the 12AU7 as a dedicated clean stage, just anecdotal tales here and there. It's a very curious utility tube mostly.

I realize it is a gain factor of 100 to 20, but could a purely clean channel be developed on the 12AU7 side to be conceivably just as loud as a 12AX7? Playing with coupling caps, plate resistors and cathode values?

Or am I better off keeping a 12AX7 and compensating the values to be cleaner?

[R.G.]

You're confusing signal size with gain. The 12AU7 has a lower gain (that is, size of the signal out divided by the size of the signal in), but can swing as big a signal on its plate as the higher gain variants.

Edit: I realized I didn't finish the thought. In an amplifier, it would be hard to get as much voltage gain out of a 12AU7 as you get out of 12AX7, or any of the others in the 12Aa7 series. You would need more stages to make the voltage signal as big - that is, as loud at the output. It might be possible to surround the 12AU7 with fancy circuitry to goose this up, but that's again more effective stages to get as big a voltage signal. The AU7 can do much more current - up to 20ma - than the AX7 (~1-2ma), so it's better for current drivers than voltage gain.

[Stevem]

Big big current difference!

Put a 12ax7 in a Marshall major in place of a 12au7 and watch the over 200 watts of rms output drop down to 90.

[GlideOn]

You're confusing signal size with gain. The 12AU7 has a lower gain (that is, size of the signal out divided by the size of the signal in), but can swing as big a signal on its plate as the higher gain variants.

I understand that signal and gain are two separate things, yes.

Gain is not distortion. Gain is GAIN or arguably signal as you say.

I am admittedly much more experienced with high gain applications.

I've experimented with using higher plate values than normal for 12AX7 stages and keeping the cathodes roughly 1.5 to 3.3k with less than 1uf cathode value. That's the other side of the coin though as you get more distortion but you're actually limiting fidelity and losing lows and highs at the expense of added mids. The volume is surprisingly around the same or just a smidgen more.

This proposed using of 12AU7 would be in the complete opposite direction.

I'm curious whether using lower plate values like in the realm of 22k to 10k would raise the voltage enough to provide a sufficient "volume" or signal increase that would be similar to a normal 12AX7 with 100k plate, 1.5k/25uf config. Keep the coupling cap large-ish around .082uf or .1uf

Of course, I'm not just working in the dark, I calculated some load lines and they look fairly favorable compared to a 12AX7.

My B+ after 4.7k preamp dropper is around 300v.


12AU7, 10k anode @ 300v:

Screenshot_20241223-114251~2.png

Screenshot_20241223-114244~2.png

12AX7, 100k anode @ 300v:

Screenshot_20241223-123516.png

Screenshot_20241223-123539.png

[R.G.]

I understand that signal and gain are two separate things, yes.
Gain is not distortion. Gain is GAIN or arguably signal as you say.

Apologies - no disrespect intended. What I was getting at was that "loud" corresponds to "signal voltage size" in my mind. So yes, a 12AU7 stage can output a signal voltage as big, peak to peak, as a 12AX7 can, or very nearly so. My caveat is that to get a signal voltage that big on the plate, the 12AU7 will generally need a bigger input signal to drive its grid.

I've experimented with using higher plate values than normal for 12AX7 stages and keeping the cathodes roughly 1.5 to 3.3k with less than 1uf cathode value. That's the other side of the coin though as you get more distortion but you're actually limiting fidelity and losing lows and highs at the expense of added mids. The volume is surprisingly around the same or just a smidgen more.

This proposed using of 12AU7 would be in the complete opposite direction.

I'm curious whether using lower plate values like in the realm of 22k to 10k would raise the voltage enough to provide a sufficient "volume" or signal increase that would be similar to a normal 12AX7 with 100k plate, 1.5k/25uf config. Keep the coupling cap large-ish around .082uf or .1uf

Again, I'm confused by your terminology. It is somewhat foreign to my understanding of tube operation to quote operating points in terms of cathode and plate resistances. I would be much more at home with the resistances stated, but the plate, grid, and cathode voltages listed.

In fact, this is so outside how I see tube operation usually described, I can't help but ponder whether your question comes from an AI chatbot. No disrespect intended if you're a person, but the question is way outside the norm for this kind of discussion. That's how I got to thinking you might be a beginner and just not understand the terms.

[GlideOn]

I understand that signal and gain are two separate things, yes.
Gain is not distortion. Gain is GAIN or arguably signal as you say.

Apologies - no disrespect intended. What I was getting at was that "loud" corresponds to "signal voltage size" in my mind. So yes, a 12AU7 stage can output a signal voltage as big, peak to peak, as a 12AX7 can, or very nearly so. My caveat is that to get a signal voltage that big on the plate, the 12AU7 will generally need a bigger input signal to drive its grid.

I've experimented with using higher plate values than normal for 12AX7 stages and keeping the cathodes roughly 1.5 to 3.3k with less than 1uf cathode value. That's the other side of the coin though as you get more distortion but you're actually limiting fidelity and losing lows and highs at the expense of added mids. The volume is surprisingly around the same or just a smidgen more.

This proposed using of 12AU7 would be in the complete opposite direction.

I'm curious whether using lower plate values like in the realm of 22k to 10k would raise the voltage enough to provide a sufficient "volume" or signal increase that would be similar to a normal 12AX7 with 100k plate, 1.5k/25uf config. Keep the coupling cap large-ish around .082uf or .1uf

Again, I'm confused by your terminology. It is somewhat foreign to my understanding of tube operation to quote operating points in terms of cathode and plate resistances. I would be much more at home with the resistances stated, but the plate, grid, and cathode voltages listed.

In fact, this is so outside how I see tube operation usually described, I can't help but ponder whether your question comes from an AI chatbot. No disrespect intended if you're a person, but the question is way outside the norm for this kind of discussion. That's how I got to thinking you might be a beginner and just not understand the terms.

My question does not come from an AI chatbot, no and I am most definitely human last I checked.

When I propose a question, I try to give some context and background from which it comes from so others can understand some of my thought process and have an easier time responding in kind. This comes from my experience and shortcomings of trying to get a normal 12AX7 to behave in a given setting, normally when trying to achieve additional gain and distortion.

To illustrate further, I've had mediocre results with attempting to get a 12AX7 stage to behave cleanly in a V1 setting (as I'm sure many have before me). Tried 82k, 56k plates and it simply neuters the tone too much. Even with 820ohm and 25uf cathodes to keep it lush and full it is still lacking. Alternatively, keeping 100k plate and trying huge cathode values like 4.7k to 10k, it still doesn't quite sound right and loses too much volume and sparkle.

There's really not that many variables or confusing terminologies at play - we have our anode, cathode and whatever signal is incoming which is the guitar via the grid resistor. It's 2 or 3 things that can be manipulated. It's V1 Gain stage going into classic Marshall typography; V2 Cathode Follower and then correspondingly V3 Phase Invertor, then Master Volume/Power Amp.

I'm not terribly experienced with this AU7/ECC823 tube type and am admittedly a 'beginner' with it as you say. It's not a commonly seen in guitar application (other than as low watt power tubes), hence the reason of the post.

I think I may understand where you're coming from but just as you to me; I may not fully understand your thought process unless you provide some context as well

It could be a classic instance of "there's no free lunch" where the solution will always be some sort of compromise, but again this was a shot to see just so.

Perhaps more of the HiFi crowd would have some more experience with this?

[GlideOn]

Some insight as to others exploring the merits of using 12AU7:

https://el34world.com/Forum/index.php?topic=9714.0

[Suckseeder]

Vox used a 12AU7 in the first stage of the AC50 and AC100

[Suckseeder]

https://schematicheaven.net/voxamps/ac50_1.pdf

[cdemike]

I think it could be worth figuring out where your amp is overdriving and adjusting things accordingly. Assuming V1 is set up with known workable values for 12AX7 or isn't working at a strange operating point, I'd hazard a guess that overdrive is occurring primarily at the third gain stage with the second also contributing at higher volume levels. I'm not clear on how the amp is set up, though, with reference to V1 going into classic Marshall topology: do you mean with V1 cascaded like in JCM800 or is parallel like in the 5F6A and associated JTM and early JMP derivatives? If V1 is cascaded (i.e., if there are 4 total gain stages in the preamp), those stages will overdrive very early assuming that the topology of the amp isn't significantly different than a JCM800. If, on the other hand, V1 is in parallel, then I think the third gain stage would be most likely to be overdriving first.

The reason I'd want to know which is overdriving first is to identify which stage would more likely send a cleaner signal to the power amp if set up with a higher headroom stage than is currently in the amp. Because 12AU7s can be set up with higher cathode voltages while remaining center biased, you could say that "input headroom" is higher than 12AX7, so it'd be able to amplify a comparatively larger signal while remaining clean. As has been mentioned, though, the total output may be significantly quieter, depending on what the signal path looks like downstream of the gain stage(s) in question. For example, if the signal reaching the power stage is loaded down due to the comparatively high output impedance of 12AX7, a 12AU7 may actually cause the output stage to distort more than with the old 12AX7 even though the preamp stage in question might be cleaner depending on the operating point (just an example -- it doesn't sound like you're doing anything strange in terms of grid leak resistor value in the output stage).

I agree with RG, though, that more information about the amp would be helpful. Without knowing operating points of the tubes currently in the signal chain, it's hard to say how to approach a possible plan.

[GlideOn]

https://schematicheaven.net/voxamps/ac50_1.pdf

Yep, that's one and although the schematic is extremely blurry and difficult to decipher I can faintly see a 47k anode coming from 300v (same as mine) on V1A. I don't know how in relation to a 12AX7 it sounds, but it shows a modest attempt at optimizing this lower gain/lower resistance tube. Seems it would generate a decent amount of voltage swing, but at what volume?

I think it could be worth figuring out where your amp is overdriving and adjusting things accordingly. Assuming V1 is set up with known workable values for 12AX7 or isn't working at a strange operating point, I'd hazard a guess that overdrive is occurring primarily at the third gain stage with the second also contributing at higher volume levels. I'm not clear on how the amp is set up, though, with reference to V1 going into classic Marshall topology: do you mean with V1 cascaded like in JCM800 or is parallel like in the 5F6A and associated JTM and early JMP derivatives? If V1 is cascaded (i.e., if there are 4 total gain stages in the preamp), those stages will overdrive very early assuming that the topology of the amp isn't significantly different than a JCM800. If, on the other hand, V1 is in parallel, then I think the third gain stage would be most likely to be overdriving first.

The reason I'd want to know which is overdriving first is to identify which stage would more likely send a cleaner signal to the power amp if set up with a higher headroom stage than is currently in the amp. Because 12AU7s can be set up with higher cathode voltages while remaining center biased, you could say that "input headroom" is higher than 12AX7, so it'd be able to amplify a comparatively larger signal while remaining clean. As has been mentioned, though, the total output may be significantly quieter, depending on what the signal path looks like downstream of the gain stage(s) in question. For example, if the signal reaching the power stage is loaded down due to the comparatively high output impedance of 12AX7, a 12AU7 may actually cause the output stage to distort more than with the old 12AX7 even though the preamp stage in question might be cleaner depending on the operating point (just an example -- it doesn't sound like you're doing anything strange in terms of grid leak resistor value in the output stage).

I agree with RG, though, that more information about the amp would be helpful. Without knowing operating points of the tubes currently in the signal chain, it's hard to say how to approach a possible plan.

Fair point, so I will post my layout for reference and with the proposed changes highlighted in "Robinette" fashion in hot pink:

AX84 SEL V26 maybe.png

AX84 SEL V26 maybe.pdf

IMG_20241222_133659~3.jpg

This is currently a working amp, but to summarize the changes going on:


Single Ended Parallel. EL34s @ approx 20w.

Classic Marshall style preamp.



Unique features:

- Series/Cascade switch (V1B series switched in or grounded).
- Bright Cap/ Cascade Peaker cap switch.
- V1a Cathode Cap
- V2 Cathode 'Fat Cap' on Mids push/pull pot.
- V2 Cathode Follower protection diode.
- Gain Pot 'Tight' push/pull, also engages V1B cathode cap.
- Master Volume (post PI) 'lean' (parallel for 150k grid leak) push/pull.
- Treble Pot with 330pf/56k or 500pf/39k 'Slope' push/pull.
- Bass Pot 'More Bass' push/pull.
- Presence Pot with 'Era' push/pull.
- V3 Faux Phase Inverter using 1/2 V3 stage.
- Master Volume after the Faux PI Stage.
- Negative Feedback circuit.
- Resonance Pot on rear.
- Artificial 6.3v Center Tap.
- 220r Output Transformer protection.


Proposed changes:

- Utilize unused half of V3 becomes new dedicated Clean channel.
- (*) ECC832/reverse 12DW7 tube, 12AU7 side with 10k anode resistor.
- Center biased cathode RC for maximum headroom
- Small 10k value Grid Input Resistor.
- Large value Coupling Cap.


(*) Open to considering utilizing other tube types, or alternative for optimizing existing 12AX7.

[GlideOn]

Hope you all are recovering okay from food and booze past few days! I know I am!

Some more thought to generate maximum volume:

I can simply attach to the main B+ 405-410v before the 4.7k preamp dropper, use the same 10k to 12k resistor, go directly to the anode connection, presuming there's a bit of 10ma + current suck from the higher demand of an AU7, get somewhere realistically in the neighborhood of a healthy 280-300v which puts me in Fender Twin territory. Full-bodied center bias of Cathode like a 1.5k/22uf. That would allow maximum clean headroom and volume for a clean channel, no?

[Bergheim]

Hi GlideOn! The requested operation points are measured voltages through the amp, most importantly supply node voltages, anode voltages and cathode voltages.
It would also be nice to have a as-is schematic of the amp. The faux PI and the parts around it is particularly interesting.
A classic 100k/1.5k 12ax7 input stage with >250V B+ is gonna be spanky clean on its own, if fed by a normal guitar signal.

[GlideOn]

Hi GlideOn! The requested operation points are measured voltages through the amp, most importantly supply node voltages, anode voltages and cathode voltages.
It would also be nice to have a as-is schematic of the amp. The faux PI and the parts around it is particularly interesting.
A classic 100k/1.5k 12ax7 input stage with >250V B+ is gonna be spanky clean on its own, if fed by a normal guitar signal.

Yes, the basis of this was to 100% utilize the capabilities and more forgiving nature of a Single Ended (paralleled) guitar amp.

The Faux/Fake PI only uses 1/2 a side and to my ears works as well as any push/pull with presence, negative feedback. The only catch is that because it is non-inverting being "fake," the tap needs to go to the start of the cathode, not after like in a push/pull as it will generate positive feedback and oscillation.

As of right now I haven't worked up the confidence to write a full schematic, which I had gone into detail on another thread. I'm not quite sure how to map out 4PDT and DPDT switching schemes in schematic form, layout is many factors easier and allows a more 1:1 typography in case of oscillations, etc. It's still very much an experimental amp.

It's a given to me having worked on many Fenders, especially that utilizing a 12AX7 at a fair high voltage will result in a strong, moderately distorting clean tone and further know that center-biasing a cathode, using sufficiently large coupling caps also tends to lessen distortion as it amplifies a fuller frequency. Hence going with a large-ish coupling cap on that stage on my proposed layout.

What remains still entirely unknown or unanswered about in this thread so far is what a 12AU7 specifically would do if it were being met with the ideal B+ voltage, ideal Anode Resistor and ideal Cathode Biasing.

Earlier it was mentioned that a Vox AC50/4 utilizes it in the V1 and even goes with a better matched 47k Anode, but I have never heard in person nor can find sound demos of that specific amp being played in a clean fashion.

If I cannot sufficiently find a solid case for utilizing, then I suppose falling back on a trusted 12AX7 for cleans is still a solid bet.

[R.G.]

What remains still entirely unknown or unanswered about in this thread so far is what a 12AU7 specifically would do if it were being met with the ideal B+ voltage, ideal Anode Resistor and ideal Cathode Biasing.

By now a reasonably intelligent and aware human should have figured out that the participants in the thread can't tell you what the ideal voltage, resistors, etc. and so on for a 12AU7 are, nor specificallyat it would do.
So go do it, and tell us what and how your did it - if actually figuring it out is your objective.
It's beginning to seem to me that your objective is to keep the thread going as long as you can.