I'm trying to use up a stash of 6DJ8s (I know, lucky me). I was thinking of using them for a MM phono stage, but you can't get more than about 34dB of gain from two 6DJ8 sections.
I could put a big honkin' low mu triode on the end to add another 10dB of gain, but that seems wasteful, and the THD would go much higher as well.
What about using up another 6DJ8 section as an anode follower on the output? Yes, the input impedance would be pretty low, but I could put the EQ between the first and second stages, and keep the input impedance of the anode follower output stage up at 47k ohms or higher. The output impedance should be low enough (1k or 2k, something like that).
Why not? What's not to like about it? Negative feedback?
--
I could put a big honkin' low mu triode on the end to add another 10dB of gain, but that seems wasteful, and the THD would go much higher as well.
What about using up another 6DJ8 section as an anode follower on the output? Yes, the input impedance would be pretty low, but I could put the EQ between the first and second stages, and keep the input impedance of the anode follower output stage up at 47k ohms or higher. The output impedance should be low enough (1k or 2k, something like that).
Why not? What's not to like about it? Negative feedback?
--
It's a good circuit. Just consider the Miller effect input capacitance with the input resistor.
You can use a smaller output coupling capacitor than usual, since it is within the feedback loop.
http://www.tubecad.com/2014/12/06/Plate Follower Simple.png
I cascode the two sections of a 12AX7, with anode follower feedback around them,
after a high value volume control.
Last edited:
I thought the circuit with the input to the cathode and the grid grounded is called a 'grounded grid amplifier.'
I guess there's some confusion about what an anode follower actually is. I'm referring to the circuit where you take a common cathode amplifier and return negative feedback to the grid, so-called 'plate to grid feedback.' As in the attached drawing. Maybe there's a more correct name for this?
Attachments
"Maybe there's a more correct name for this?"
Nope, that is the correct name.
With the feedback resistor removed it is a common cathode stage
Nope, that is the correct name.
With the feedback resistor removed it is a common cathode stage
"With the feedback resistor removed it is a common cathode stage"
Then it isn't an anode follower anymore.
Then it isn't an anode follower anymore.
I guess my question back to you would be this: What makes you think nobody uses it?
I've used it with great success but like anything you have to work around the circuit's drawbacks.
It's just parallel-applied voltage feedback. I tend to prefer series-applied voltage feedback in most circumstances but sometimes parallel-applied feedback is the best for the situation.
I've used it with great success but like anything you have to work around the circuit's drawbacks.
It's just parallel-applied voltage feedback. I tend to prefer series-applied voltage feedback in most circumstances but sometimes parallel-applied feedback is the best for the situation.
The reason I said 'nobody uses it' is because I was looking at phono stages and I didn't see an anode follower in any of the popular DIY projects.
There is a line stage called the Little Bear that's patterned after the line stage in the Dyna PAS, with two common cathode stages in cascade, with feedback taken off the output (plate) of the second stage and returned to the cathode of the first stage. But that's something different.
There's also the common NFB RIAA as used in the EAR 834, Dyna PAS and Marantz 7C phono stages. But again, that's a different thing.
I was thinking that if you need a line level output stage with a fairly low gain, say like 3x (10dB), and you don't want to use the 15mA or more required for a low mu triode (and you don't want to hassle with DHTs), then taking something like a 12AT7 or a 6DJ8 and wrapping parallel-applied voltage feedback from plate to grid can get you there pretty easily. Make sure the series resistor is large enough in value to keep the input impedance high enough not to load down the previous stage.
Anyway, that's what I was thinking. I've never seen a circuit like that in a popular project, and I was wondering what I might be missing...
--
There is a line stage called the Little Bear that's patterned after the line stage in the Dyna PAS, with two common cathode stages in cascade, with feedback taken off the output (plate) of the second stage and returned to the cathode of the first stage. But that's something different.
There's also the common NFB RIAA as used in the EAR 834, Dyna PAS and Marantz 7C phono stages. But again, that's a different thing.
I was thinking that if you need a line level output stage with a fairly low gain, say like 3x (10dB), and you don't want to use the 15mA or more required for a low mu triode (and you don't want to hassle with DHTs), then taking something like a 12AT7 or a 6DJ8 and wrapping parallel-applied voltage feedback from plate to grid can get you there pretty easily. Make sure the series resistor is large enough in value to keep the input impedance high enough not to load down the previous stage.
Anyway, that's what I was thinking. I've never seen a circuit like that in a popular project, and I was wondering what I might be missing...
--
Ah, okay. I have never made a phono stage myself. Never owned a turntable.
But from what I understand noise is a top concern. Parallel applied feedback can be difficult in some situations to get good noise performance since there is a largish value resistor in series with the signal. You'd want to make sure that the signal is large enough at the anode follower stage so that the SNR is good.
Without having attempted to calculate anything, I'm going to guess that noise is probably the reason not many people do it.
The big "why don't I see this" answer is that if you want a buffer with low output impedance to hang off a vintage phono design, you can add a directly coupled cathode follower incredibly easily and with minimal parts count.
Yes, but no gain.
A 6DJ8 with RIAA will get you no more than about 35dB gain. I want 45dB gain. How does one tack on 10dB of gain without adding noise and distortion?
--
A 6DJ8 with RIAA will get you no more than about 35dB gain. I want 45dB gain. How does one tack on 10dB of gain without adding noise and distortion?
--
The Artemis Labs LA-1 line amp uses an anode follower. It uses 1/2 a choke-loaded 5687 with a 432K resistor from input to grid and a 1.3 Meg resistor from the (cap-isolated) output to grid, thus giving about 12 dB of gain. The high resistance values are used to reduce loading on the volume and balance controls that feed the gain stage. However, these high impedances are sensitive to stray capacitance, and it was necessary to parallel the 1.3 Meg resistor with a 10-50pF trimcap, which is adjusted at the factory for flat square-wave response. By lowering the output impedance that is driving the choke, the low-frequency response is improved.
I think the reason that you don't see circuits like the anode follower or other topologies such as SRPP, etc. in classic tube amplifier circuits is the poverty of ideas and basic engineering laziness at the time. There was always time pressure to get the design out the door. Most circuits were either licensed (e.g. from RCA) or copied from other successful designs. And, in most cases, conventional R-C coupled circuits were adequate. There were only a few really creative engineers - this is still true today.
Tube amp designers today have the luxury of perusing decades of past literature and plucking ideas or circuits that may have only appeared in obscure academic publications. Also, we often have the time to refine and optimize design, since the time pressure to ship product into the high-end market is less. The nice thing about using obsolete technology is that it is already way past the optimum "marketing window"!
- John Atwood
I think the reason that you don't see circuits like the anode follower or other topologies such as SRPP, etc. in classic tube amplifier circuits is the poverty of ideas and basic engineering laziness at the time. There was always time pressure to get the design out the door. Most circuits were either licensed (e.g. from RCA) or copied from other successful designs. And, in most cases, conventional R-C coupled circuits were adequate. There were only a few really creative engineers - this is still true today.
Tube amp designers today have the luxury of perusing decades of past literature and plucking ideas or circuits that may have only appeared in obscure academic publications. Also, we often have the time to refine and optimize design, since the time pressure to ship product into the high-end market is less. The nice thing about using obsolete technology is that it is already way past the optimum "marketing window"!
- John Atwood
If I'm lazy and I assume you're doing passive RIAA between the 6DJ8s, you could use a 12AT7 or 12AX7 in the second stage instead of a 6DJ8, then use a 6DJ8 cathode follower to buffer the output.
For best noise performance the highest gain stage should be the first in the chain, I would use something other than a 6DJ8 in the first stage or use a cascode connected 6DJ8 - this would give you a lot more gain with the good noise performance of the 6DJ8. You can easily get 40dB+ of gain in that stage alone.
I actually have published several designs here that are cascode inputs with hybrid mu-follower outputs. You can do RIAA with a gain of >50dB @ lkHz at the output.
In my experience the anode follower itself does not have noise problems, it does however suffer from fairly low input impedance and pretty high output impedance compared to a cathode follower. I do find that placing the gain as early as possible in the pre-amplifier results in better noise performance provided that the tube you have chosen for that position is quiet to start with. Each successive stage amplifies both the signal and the noise from the preceding stage.
I actually have published several designs here that are cascode inputs with hybrid mu-follower outputs. You can do RIAA with a gain of >50dB @ lkHz at the output.
In my experience the anode follower itself does not have noise problems, it does however suffer from fairly low input impedance and pretty high output impedance compared to a cathode follower. I do find that placing the gain as early as possible in the pre-amplifier results in better noise performance provided that the tube you have chosen for that position is quiet to start with. Each successive stage amplifies both the signal and the noise from the preceding stage.
Hey Kevin -- I followed your Muscovite series of preamps with interest. I even purchased some 6S3P and 6N24P along the way.
Yes, passive 'all in one go' EQ between the first stage and second stage.
That is certainly true. There are lots of conflicting goals to contend with.
1) I want about 45dB gain. That seems to be a sweet spot in my setup. Keeps the subjective playback level of records about even with the levels from my DAC or CD player.
2) This time, I want a really simple circuit. That's why I wasn't thinking of a cascode. But how about a cascode made with an nJFET on bottom and a 12AU7 (or something or other) on top?
2) I want low input capacitance. That's why a 6DJ8 as input seemed like a good idea. But of course the cascode accomplishes that too.
3) I want to use up some 6DJ8s, because I have them. Three of them to build a preamp seemed like a good idea.
4) The anode follower on the output also might allow for adjustable gain from the preamp. Maybe a switch can be used to switch in different values of parallel NFB resistor. Or maybe that's a bad idea. There should be no DC on that resistor, since it's coming after the output cap, and it's going to the grid of that stage, after the DC blocking cap from the second stage's plate.
As the gain is dialed down, the output impedance decreases. The input impedance is set by the series NFB resistor.
As the gain is dialed up, the output impedance increases, but the input impedance is still set by the series NFB resistor -- correct?
I don't think a phono preamp needs output impedance of less than 2k ohms. No higher, though.
A cascode 6DJ8 first stage to all-in-one-go RIAA EQ into a hybrid mu-follower would be an excellent idea, I think.
So many circuits....
Yes, passive 'all in one go' EQ between the first stage and second stage.
That is certainly true. There are lots of conflicting goals to contend with.
1) I want about 45dB gain. That seems to be a sweet spot in my setup. Keeps the subjective playback level of records about even with the levels from my DAC or CD player.
2) This time, I want a really simple circuit. That's why I wasn't thinking of a cascode. But how about a cascode made with an nJFET on bottom and a 12AU7 (or something or other) on top?
2) I want low input capacitance. That's why a 6DJ8 as input seemed like a good idea. But of course the cascode accomplishes that too.
3) I want to use up some 6DJ8s, because I have them. Three of them to build a preamp seemed like a good idea.
4) The anode follower on the output also might allow for adjustable gain from the preamp. Maybe a switch can be used to switch in different values of parallel NFB resistor. Or maybe that's a bad idea. There should be no DC on that resistor, since it's coming after the output cap, and it's going to the grid of that stage, after the DC blocking cap from the second stage's plate.
As the gain is dialed down, the output impedance decreases. The input impedance is set by the series NFB resistor.
As the gain is dialed up, the output impedance increases, but the input impedance is still set by the series NFB resistor -- correct?
I don't think a phono preamp needs output impedance of less than 2k ohms. No higher, though.
A cascode 6DJ8 first stage to all-in-one-go RIAA EQ into a hybrid mu-follower would be an excellent idea, I think.
So many circuits....
Last edited:
Anode follower has the gain of 1 (0 dB) just by the definition of the "follower" 😉
If the gain is not 1, then it isn't follower anymore, just the common cathode stage with parallel feedback.
I'd go for the cascode 6DJ8 running at about 8 - 10mA and about 40dB of gain in that first stage, then passive EQ, and finally a 6DJ8 common cathode - how about LED or battery bias and a 6DJ8 as a cathode follower. It'll be fun, you can fine tune the cascode for the gain you want.
You could also use a linear systems jfet in the first stage, but I think you would have more gain than you know what to do with if you do that.. (a couple of well matched low noise nch fets in parallel could make this good for a LOMC direct in without headamp/SUT.
You could also use a linear systems jfet in the first stage, but I think you would have more gain than you know what to do with if you do that.. (a couple of well matched low noise nch fets in parallel could make this good for a LOMC direct in without headamp/SUT.
That sure would be fun. OK, time to give this a spice run... Wait... Isn't that the Mini-Muscovite III? 😉
Now things are getting advanced. But that would be fun too.
--
Last edited:
The Faylon BK30 guitar/instrument amp uses anode to grid feedback in both channels (and has an ultra linear power stage with an OPT of Hi-Fi quality).
Faylon was a Belgium manufacturer of guitar/instrument amps (closed 1977). The Dutch band "Golden Earring", for the rest of the world best known by their hit "Radar Love" (1973), used Faylon equipment for some years.
I own one and love it for it's clean sound and it's quietness (no hiss or other noise).
Faylon was a Belgium manufacturer of guitar/instrument amps (closed 1977). The Dutch band "Golden Earring", for the rest of the world best known by their hit "Radar Love" (1973), used Faylon equipment for some years.
I own one and love it for it's clean sound and it's quietness (no hiss or other noise).
Attachments
Last edited: