All tubes create distortion. The amount of unavoidable distortion produced by a tube is known as its intrinsic distortion.It is true that the intrinsic distortion of a 12AU7is much higher than many other tubes. The total distortion produced by a particular design depends on both the intrinsic distortion of the tubes used and the amount of NFB. NFB reduces intrinsic distortion by the amount of NFB. For a given amount of NFB, a tube with a high intrinsic distortion will still have a higher overall distortion, even with the NFB, than a tube with lower intrinsic distortion. So, for example, if you build the same circuit using a 6CG7, which has similar mu and gm values as the 12AU7 but much lower intrinsic distortion, this circuit will have significantly lower distortion than the 12AU7.
if you want your build to be optimal, then you need to pay attention to both the characteristics of the chosen topology and of the active components within it.
Cheers
Ian
That uses 12AU7 push-pull. This semi-cancels the 12AU7's over-hopped-up Gm curvature. Not a good guide to SE action.
However any tube distortion short of heavy clipping may be quite acceptable and even "pleasant" in reproduction. (Ian has to watch it because in music PROduction the same errors may be stacked and repeated.)
Being that the title of this thread is "A good route to a ECC82 preamp", I figured I'd try to contribute ideas for exactly that.
I agree that the 12AU7 is not the best possible tube to start with, but it's not unusable, nor is its use a guarantee of a bad amplifier. I mean, it was used by Heathkit in some of their classic Williamson-style amps from days of yore, in which the first stage is a 12AU7 used as a plain old, single-ended, common cathode amplifier. I don't know why they didn't use 6CG7 instead. Perhaps it was because the 12AU7 requires half the heater power?
I agree that the 12AU7 is not the best possible tube to start with, but it's not unusable, nor is its use a guarantee of a bad amplifier. I mean, it was used by Heathkit in some of their classic Williamson-style amps from days of yore, in which the first stage is a 12AU7 used as a plain old, single-ended, common cathode amplifier. I don't know why they didn't use 6CG7 instead. Perhaps it was because the 12AU7 requires half the heater power?
Working at VERY tiny signal level. 3V output. And huge overall NFB (Williamson's passion).
12AU7 is distant kin to the General Purpose Triodes, in days when all audio was transformer-coupled. The most modern of these are 6C6, 6J7/6SJ7/6J5. They have fairly evenly-spaced curves, Gm changes little. In some applications higher Gm is sexy. By tweaking the grid wires they can boost high-current Gm at the expense of low-current Gm---- the curves are less evenly spaced. 12AU7 is suggested for "various applications"... they are not going to tell you not to buy and use it! But for small single-ended audio, "there are better choices".
I'm puzzled by the value of R5. If the current (4.81 mA) through R6 would be equal to the current through R5 than the bias of U2 would only be -0.27 V.
But the voltages in the schematic seem to indicate a bias of -8 V (106 - 98). The current of 4.81 mA at a bias of -8 V and a voltage between cathode and anode of 196 V is exactly right when looking at a datasheet for the 12AU7. Since the current through U1 (2 mA) at a bias of -3.9 V and a voltage between cathode and anode of 94 V is also exactly right according to the datasheet for the 12AU7, grid current can't be involved.
So shouldn't R5 be around 1663 Ohm (R = V / I = 8 / 0.00481 = 1663)?
Yes, exactly. Like a line level preamp. Figure the maximum output will be 2V rms, or approx. 3V peak. 16dB NFB as proposed.
"A good route to a ECC82 preamp".
I agree that there are better possibilities. But that's the name of the thread, so I took it up as a little problem to try to solve.
Possibly a better solution would be a common-cathode stage DC-coupled to 12AU7 White cathode follower with NFB around it. Basically the same idea, but fancier. Like this idea of yours from a couple years ago, which I've adapted for use with 12AU7, and looks like it should be excellent...
Last edited:
You could possibly bootstrap R9 from the output to increase the open loop gain and hence reduce the distortion. I am not sure I see the point of using a White follower as the output. There is no need for a phenomenally low small signal output impedance. You might just as well arrange the two triodes of a single 12AU7 as a mu follower and put some NFB around that and save a triode.
Cheers
Ian
Cheers
Ian
I brought out the White cathode follower as another good circuit that could be made using a 12AU7. I agree that it's overkill.
A variation on your idea...
Use a single 12AU7 as a 'bootstrapped pair', which is pretty similar to a mu-follower, but wouldn't need to have the heater supply elevated to quite as high a voltage.
V1's 47k plate resistor is split is into two 24k resistors, then the output of the V2 cathode follower is fed back through a large-ish capacitor to the center-tap of the two 24k plate resistors on V1.
Then wrap NFB from the output of the V2 cathode follower to the grid of V1.
A Bootstrapped Pair/Anode Follower.
A 12AU7 circuit that doesn't suck, perhaps.
A variation on your idea...
Use a single 12AU7 as a 'bootstrapped pair', which is pretty similar to a mu-follower, but wouldn't need to have the heater supply elevated to quite as high a voltage.
V1's 47k plate resistor is split is into two 24k resistors, then the output of the V2 cathode follower is fed back through a large-ish capacitor to the center-tap of the two 24k plate resistors on V1.
Then wrap NFB from the output of the V2 cathode follower to the grid of V1.
A Bootstrapped Pair/Anode Follower.
A 12AU7 circuit that doesn't suck, perhaps.
Hello...I have read the following posts only now because I am assembling the various parts of the preamp (at the end of the week I will finish it I think) and didn't followed the thread.
At the end of the show I pursued my path with the ccda, according to the tip from # 36 by Rongon (adapted to my 260V B+).
Obviously I am aware of the limits of the project and the potential provided by the 12AU7, but from a first listening to the prototype it seems to me not bad. Also considering that I don't want too much gain. For me 6-8 dB is enough, since my first aim was a simple buffer, so I then set a NFB with an Rf = 220K.
Since the preamp is small in size, I am using a toroidal transformer with a 13Vac + 110Vac secondary outputs with a voltage doubler for B+ and various filters (including a gyrator, maybe with a further LM317 Maida regulator). I think I can also include a ground lift switch if I ever have ground loop problems.
For now that's all, in case I will update when the job is finished to tell my impressions...
At the end of the show I pursued my path with the ccda, according to the tip from # 36 by Rongon (adapted to my 260V B+).
Obviously I am aware of the limits of the project and the potential provided by the 12AU7, but from a first listening to the prototype it seems to me not bad. Also considering that I don't want too much gain. For me 6-8 dB is enough, since my first aim was a simple buffer, so I then set a NFB with an Rf = 220K.
Since the preamp is small in size, I am using a toroidal transformer with a 13Vac + 110Vac secondary outputs with a voltage doubler for B+ and various filters (including a gyrator, maybe with a further LM317 Maida regulator). I think I can also include a ground lift switch if I ever have ground loop problems.
For now that's all, in case I will update when the job is finished to tell my impressions...
Last edited:
The amount by which NFB reduces distortion and other unwanted effects is equal to the open loop gain in dB minus the closed loop gain in dB. Positive feedback increases the open loop gain. So, for example, a 10dB increase in open loop gain will reduce distortion etc by more than 3 times.
Cheers
Ian
I think the problem would be that when the amplifier is driven to clipping (into cutoff) the negative feedback loop will cease to work, so the positive feedback could cause the amplifier to oscillate. That would make for some seriously ugly clipping behavior. However, I may have misunderstood what I was reading. I was reading this:
https://linearaudio.nl/sites/linearaudio.net/files/miller combined fb electronics march 1950.pdf
The maxim I've read is that for a stable and clean result, you first want to make your amplifier circuit as low distortion and with as linear phase response as possible open loop, before applying negative feedback.
It would follow that if positive feedback increases gain, but also causes oscillation without the negative feedback applied, that the end result would be a less stable amplifier circuit (an amplifier that would be more prone to oscillation, overshoot and/or ringing when stressed).
Am I wrong about that?
https://linearaudio.nl/sites/linearaudio.net/files/miller combined fb electronics march 1950.pdf
The maxim I've read is that for a stable and clean result, you first want to make your amplifier circuit as low distortion and with as linear phase response as possible open loop, before applying negative feedback.
It would follow that if positive feedback increases gain, but also causes oscillation without the negative feedback applied, that the end result would be a less stable amplifier circuit (an amplifier that would be more prone to oscillation, overshoot and/or ringing when stressed).
Am I wrong about that?
Not per se. It depends on how the PFB is derived and applied. In bootstrapping, for example, its effect is to increase the dynamic resistance of a plate load - much like adding a CCS. This not only increeases the gain but tends to reduce distortion by making the plate load lighter.
Of course, badly implemented PFB will lead to all sorts of issues (ditto with NFB)
Cheers
Ian
In simple terms, both positive and negative feedback affect the amplitude and phase characteristic of the loop gain and it is this that determines stability. The key determinant is that when the loop gain reaches unity, the phase shift must be less than 180 degrees to avoid oscillation.
The question of the effect of negative feedback when one or more stages enters clipping is entirely different - at this point the amplifier is clearly being operated outside its design parameters. For the majority of tube based amplifiers, the reduction in open loop gain caused by the relatively gentle tube characteristics plus the relatively low amount of NFB used, tends to mitigate this. The opposite is true in transistor amplifiers where open loop distortion is so high that large amounts of NFB are necessary to achieve low distortion.
Cheers
Ian
While assembling the preamp I noticed a strange fact that I had only noticed in old tube radios, namely a sudden bright glow of the tubes on powerup. Does anyone know how to explain the phenomenon, and if it is harmful? I assumed it was the regulator that for 1-2s supplies an overcurrent with the heaters not yet turned on...
If the NFB is crapped-out, isn't the PFB also kaput?
The PFB/NFB trick was WIDELY used in WWII servo and filter amplifiers. Seen also in Dynaco.