zintolo,
Thanks for the interesting article!
There can be other factors that affect the performance of various circuit topologies and techniques.
You mentioned 3rd harmonic distortion in your Post # 30.
My latest amplifier is Balanced input and Balanced output stage. The 2nd harmonic and 3rd harmonic distortions are both less than -50dBc,
and they are within 3 dB of each other.
Many single ended have 15dB more 2nd than 3rd harmonic distortion.
Many push pull have Lots More 3rd harmonic than 2nd harmonic distortion.
My amplifier has Almost Equal 2nd and 3rd harmonic distortion.
But I make those measurements with a non-inductive purely resistive load resistor.
So many articles show what happens when the amplifier has a non-inductive purely resistive load resistor on the output.
You can see graphs of 2nd order IM and 3rd order IM, 2nd Harmonic, and 3rd Harmonic distortions.
And then there are graphs of distortions versus different load impedance on Ultra Linear circuits, and various UL tap %.
So much for all of those curves. They will change when a real world loudspeaker is connected, instead of a load resistor.
Loudspeakers present various impedances of R, LR, LCR, and CR. Various impedances change the slope of the load line.
And reactive loads, with L, C, or both change the load line into an Elliptical Load line, with the slope of the Ellipse according to the complex value of Z.
Have Fun!
It is a Mean thing to say that Karl Friedrich Gauss was just an Average man at the Peak of his career, and at the Center of the Gaussian Curve.
Thanks for the interesting article!
There can be other factors that affect the performance of various circuit topologies and techniques.
You mentioned 3rd harmonic distortion in your Post # 30.
My latest amplifier is Balanced input and Balanced output stage. The 2nd harmonic and 3rd harmonic distortions are both less than -50dBc,
and they are within 3 dB of each other.
Many single ended have 15dB more 2nd than 3rd harmonic distortion.
Many push pull have Lots More 3rd harmonic than 2nd harmonic distortion.
My amplifier has Almost Equal 2nd and 3rd harmonic distortion.
But I make those measurements with a non-inductive purely resistive load resistor.
So many articles show what happens when the amplifier has a non-inductive purely resistive load resistor on the output.
You can see graphs of 2nd order IM and 3rd order IM, 2nd Harmonic, and 3rd Harmonic distortions.
And then there are graphs of distortions versus different load impedance on Ultra Linear circuits, and various UL tap %.
So much for all of those curves. They will change when a real world loudspeaker is connected, instead of a load resistor.
Loudspeakers present various impedances of R, LR, LCR, and CR. Various impedances change the slope of the load line.
And reactive loads, with L, C, or both change the load line into an Elliptical Load line, with the slope of the Ellipse according to the complex value of Z.
Have Fun!
It is a Mean thing to say that Karl Friedrich Gauss was just an Average man at the Peak of his career, and at the Center of the Gaussian Curve.
Maybe it's a little advanced for me, what kind of resistor should I put in the cathode of the 6v6?
In the image you can see the output and the 2 signals before reaching the 6v6 grid, this is eliminating the 100ohm common cathode resistance of the 12ax7, as you can see, the thd is 1.5%, nothing serious i guess
more details
original Schematic showing FFT
Without cathode-common resistor
and with the 300k (replacing270k)
I'm still testing in LTSPICE before taking it to the real thing, maybe it will give me an idea of how it will behave, although it's obviously not exact.
original Schematic showing FFT
Without cathode-common resistor
and with the 300k (replacing270k)
I'm still testing in LTSPICE before taking it to the real thing, maybe it will give me an idea of how it will behave, although it's obviously not exact.
Thanks to you! I still remember the first time you helped me years ago and all the effort you put in explaining where I was wrong. Very useful information!
Great! Have you shared the design on this forum?
I remember you told me so, and my guess is that this is the purpose of this amp: sound like a SE with the reduced costs of a PP.
Transconductance of 6V6GT in triode mode is 5000 uS, so the resistor should be 1/gm = 200 Ohm, while for 12AX7 you should have around 680 Ohm.
I've played with values a bit and I got this:
You get the following -15dB 3rd to 2nd harmonic SE signature, and 0,2% THD at 1 Wrms:
Code:
Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+3 3.964e+0 1.000e+0 90.53° 0.00°
2 2.000e+3 7.623e-3 1.923e-3 -174.49° -265.02°
3 3.000e+3 1.288e-3 3.248e-4 -87.91° -178.44°
4 4.000e+3 2.125e-4 5.361e-5 177.00° 86.48°
5 5.000e+3 5.419e-5 1.367e-5 92.25° 1.73°
6 6.000e+3 1.409e-5 3.555e-6 78.91° -11.62°
7 7.000e+3 1.095e-5 2.763e-6 88.96° -1.57°
8 8.000e+3 1.022e-5 2.578e-6 89.64° -0.89°
9 9.000e+3 9.103e-6 2.296e-6 89.53° -1.00°
Partial Harmonic Distortion: 0.195110%
Total Harmonic Distortion: 0.195109%
there's no need for the cathode bias cap?? (CK in my squematic)
i dont have the same results as you
i dont have the same results as you
Last edited:
No, but there's an issue with 6V6: they are designed to work in A2/AB2:
staying in A1 like this design wants, you cannot go above 4 Wrms, because anodes can't go lower than 180V.
EL84 would make it a totally different amplifier, with enough power for anything you may need, specially with something like 20-40% UL, while keeping the same harmonic signature. Having octal connections 7591 can be another option too at 40% UL instead of triode.
staying in A1 like this design wants, you cannot go above 4 Wrms, because anodes can't go lower than 180V.
EL84 would make it a totally different amplifier, with enough power for anything you may need, specially with something like 20-40% UL, while keeping the same harmonic signature. Having octal connections 7591 can be another option too at 40% UL instead of triode.
Something like this (playing with R5 (your R9) will give 7 Wrms at 2% THD still in class A and less than 0,2% THD at 1 Wrms:
Code:
Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+3 3.205e+0 1.000e+0 90.71° 0.00°
2 2.000e+3 5.012e-3 1.564e-3 2.02° -88.69°
3 3.000e+3 3.190e-3 9.954e-4 94.18° 3.47°
4 4.000e+3 1.030e-4 3.213e-5 -146.73° -237.44°
5 5.000e+3 6.100e-5 1.904e-5 -87.98° -178.68°
6 6.000e+3 3.052e-5 9.522e-6 -89.06° -179.76°
7 7.000e+3 2.603e-5 8.122e-6 -90.08° -180.79°
8 8.000e+3 2.297e-5 7.168e-6 -90.03° -180.73°
9 9.000e+3 2.043e-5 6.374e-6 -90.02° -180.73°
Partial Harmonic Distortion: 0.185444%
Total Harmonic Distortion: 0.185451%I am not sure what is going on.
The results are not what I expected with the modification of the original schematic, but with the 100 Ohm resistor removed, and the cathodes connected together.
The 680 Ohm without any bypass capacitor is a poor man's differential that is combined with a paraphase splitter.
Probably not a good combination.
Time for me to think some more, and see if I can understand all the recent posts.
The results are not what I expected with the modification of the original schematic, but with the 100 Ohm resistor removed, and the cathodes connected together.
The 680 Ohm without any bypass capacitor is a poor man's differential that is combined with a paraphase splitter.
Probably not a good combination.
Time for me to think some more, and see if I can understand all the recent posts.
Do you have a reference for that? I doubt the 6V6 was designed for control grid current operation. The curves, like for many other tubes, show operation with positive grids, but all datasheet examples (and schematics I have seen) are for classes A1/AB1.
That does not mean, of course, that it cannot be used for A2/AB2.
I just noticed that with this design we can't go over 4 Wrms due to how the curves of the 6V6.
I switched to EL84 (or 7951, being octal) in UL to have linearity, low rp and double the power.
It is indeed, but this design has the purpose to have 2nd harmonic.
By design is 1 kOhm, that I lowered to 680 Ohm.
yes, and switched to EL84 or 7951 because they can give much more power in A1/AB1 and, from previous experience, they have very good rp at 40% UL compared to 6V6 in triode.
Last edited:
I have deducted from the fact that curves go up to +20 V for g1, that his big brother 6L6GC was designed for AB2 and other tubes not intended for AB2 do not show curves above Vg1 = 0. I have used them in AB2 for one of my Baby Huey amps and they worked fine.
I don't have datasheets that show A2/AB2 operation.
Some tubes, especially some RF tubes are made to work well with grid current.
Many audio Beam Power tubes, and many audio Pentode tubes are Not made to have any grid current.
You are on your own if you do that.
With RC coupling between the driver tubes and the output tubes, any grid current will change the voltage across the coupling cap (the 2 in AB2 means grid current).
That causes the output tube bias voltage to change, and it takes time to recover to the original quiescent bias voltage.
If it is bad enough, it not only is distorted, it is called Blocking.
The same is true for RC coupling and A2 single ended.
Note: All the schematics in this thread are RC coupled.
AB1 is OK. No grid current.
AB2 is Not good. grid current.
A driver that is DC coupled to an output tube that is made for grid current, can do a fair job of AB2, IF the driver has a low output impedance, plus the driver must have enough output current.
If not . . . stay in AB1 operation.
Just my un-distorted opinion (pun intended).
Many audio Beam Power tubes, and many audio Pentode tubes are Not made to have any grid current.
You are on your own if you do that.
With RC coupling between the driver tubes and the output tubes, any grid current will change the voltage across the coupling cap (the 2 in AB2 means grid current).
That causes the output tube bias voltage to change, and it takes time to recover to the original quiescent bias voltage.
If it is bad enough, it not only is distorted, it is called Blocking.
The same is true for RC coupling and A2 single ended.
Note: All the schematics in this thread are RC coupled.
AB1 is OK. No grid current.
AB2 is Not good. grid current.
A driver that is DC coupled to an output tube that is made for grid current, can do a fair job of AB2, IF the driver has a low output impedance, plus the driver must have enough output current.
If not . . . stay in AB1 operation.
Just my un-distorted opinion (pun intended).
Last edited:
I opted for your solution, 6A3sUMMER.
The voltages became even, I couldn't get a 300k resistor, so I put the closest thing on my market, 330k.
Unfortunately I cannot get higher quality OPT in my country to see that improvement that you talk so much about, maybe on some trip I will get something.
I've been listening to music and I still haven't noticed much difference, maybe a little more bass. I attach a final image, maybe it will help someone, the harmonics that are seen are ok, I see that the 2nd harmonic passed to the 3rd harmonic, Without the modifications, the 3rd harmonic was superior to the second
The voltages became even, I couldn't get a 300k resistor, so I put the closest thing on my market, 330k.
Unfortunately I cannot get higher quality OPT in my country to see that improvement that you talk so much about, maybe on some trip I will get something.
I've been listening to music and I still haven't noticed much difference, maybe a little more bass. I attach a final image, maybe it will help someone, the harmonics that are seen are ok, I see that the 2nd harmonic passed to the 3rd harmonic, Without the modifications, the 3rd harmonic was superior to the second
I would not consider A2/AB2 for this simple circuit. My class AB2 6V6GT amp has dc coupled mosfets (supplied at +50 and -100V) to push them into AB2.