6P14P push pull with low output

[laksaja]

Hello!

I built a tube amp like in attached schematics. I have those Russian 6p14p tubes and i decided to make a simple pp amp with them. I entered the audio signal to the grid of 6Z1P in schematics through 0.1uf capacitor. I used ECC83 in the first stage and phase splitter as well. The problem is that whatever i do, i get the clean output only to 1W@4 ohm on test load (2.1V AC). I measured the signal with windows based oscilloscope with soundcard (since i used it successfully with transistor amps, i can trust it somewhat . To the 1W the sine wave is clean, above that the sine wave get flattened like tubes limiting the signal effectively. Power supply voltage is 270V B+ with no signal and around 254 with this 1W out.
The frequency response is from 100hz to 19 khz (the gain is rising a little with frequency). Since i use cheap output transformer, its more than expected. Actually this circuit should work with negative feedback, but when i connected it to first stage cathode, the amp starts oscillating and no capacitor nor reducing the value of FB resistor thing would not work.
Can anybody suggest, how i can raise the output without raising the voltage of plates? This consruction should give at least clean 5w output which is my goal.

Thank you!

[costis_n]

One problem that I can see is that there is no power supply smoothing (resistor+capacitor to ground) between the 6Π14Π and the 6Ν2Π stage. This means that power supply noise signal is present on the anode of the 6Ν2Π. The noise is the output signal inverted. This is the first thing that "hit my eye". Try something like 2200Ω + 50 μF. About the feedback problem, you should connect the secondary of the output transformer in reverse. This will put the feedback in the correct phase.

[Arnulf]

1: Measure signal swing along its path (at the anode of first and second amplification stage, or after its coupling capacitor). Let us know what the readings say. Don't use sound card for that, use a propper voltmeter !

2: Feedback is applied to the second stage screen in your schematic, not first stage cathode. Have you followed the schematic or have you done something differently ?

3: If you have followed your schematic to the point and the amplifier is oscillating, you've probably taken the feedback signal from the wrong leg of the OPT secondary. Try using the other leg (you need to move ground connection as well) and report back.

[Ian444]

With 1W output I would expect the amp to be still in Class A, so the B+ should not drop. Something wrong there? Are you sure the output transformers are what you think they are and that the load is correct.

[45]

One thing you should never give for granted is the feedback network. If you use a different OPT more likely it will not work very well.
Secondly, that phase-splitter is never really balanced! You should take the lower output from the cathode and not between the 2K2 and 47K resistors. I notice a diffuse misconception about the Zout's of this kind of splitter. They are the same if the anode load is equal to the cathode load. Of course if the two loads are not the same the two output signals will be not the same and this usually happens because of the output stage input impedance. Also this is true for ANY phase splitter.
So, to make it simpler, I would add a 2K2 in series with the anode resistor and take the lower output directly from the cathode. Try first without feedback. It should work decently at mid frequencies anyway....

45

[laksaja]

Thank you all for your suggestion! I attach the scope froms, everything regarding to the phase splitter should work. It is measured no 6P14P in place and from grid points. The voltmeter showed 0.6 volt both tube grids before clipping.The last frequency response and feedback graph is taken where the amp worked with dummy load. I tried also switch the OPT leads (old and good trick) between each other, still the same.
And now comes the point where everybody smashing big fist into my face . Im using 100V line transformer......because i ordered 30W UL transformers from US (good paper bobbin expensive OPT-s) and it takes at least one motnh to arrive to Estonia. I was just curious how the 100v line OPT-s perform.

Thank you again for your suggestions!

[costis_n]

Originally Posted by 45

One thing you should never give for granted is the feedback network. If you use a different OPT more likely it will not work very well.
Secondly, that phase-splitter is never really balanced! You should take the lower output from the cathode and not between the 2K2 and 47K resistors. I notice a diffuse misconception about the Zout's of this kind of splitter. They are the same if the anode load is equal to the cathode load. Of course if the two loads are not the same the two output signals will be not the same and this usually happens because of the output stage input impedance. Also this is true for ANY phase splitter.
So, to make it simpler, I would add a 2K2 in series with the anode resistor and take the lower output directly from the cathode. Try first without feedback. It should work decently at mid frequencies anyway....

45

If we want similar output impedance from the two output, wouldn't the extra output impedance introduced by connecting the lower output to the 47K help with that? Because the lower part is a cathode follower and the upper part is common cathode. In that case the effect of the different connections are minimal and the mis-match is impossible to cause such a big problem , like oscillation. In any case, I would say add a 2K2 on the anode side, but take the output from the juction of the lower 2K2 and the 47K.

But all this is for after you make it work...

[45]

Originally Posted by costis_n

If we want similar output impedance from the two output, wouldn't the extra output impedance introduced by connecting the lower output to the 47K help with that? Because the lower part is a cathode follower and the upper part is common cathode. In that case the effect of the different connections are minimal and the mis-match is impossible to cause such a big problem , like oscillation. In any case, I would say add a 2K2 on the anode side, but take the output from the juction of the lower 2K2 and the 47K.

But all this is for after you make it work...

No. You have equal outputs if you have the same loads for anode and cathode and take the outputs in the same way. If you take one output directly from the anode and the other at the junction of the 2K2 and 47K resistor the splitter is unbalanced and you get distortion.
The misconception is that people think that as one output is cathode follwer and the other is common cathote then the Zout's are different. This is true only if you take one output per time if you take both outputs simultaneously the Zout's and signals are the same (as long as the loads are the same).
The oscillation could be due to the OPT which is different and so that feedback network doesn't work. Also one could use grid stoppers for the output tubes or for all of them in the worst case.

45

[DF96]

You said that you connected the signal to the 6Z1P grid via a 0.1uF capacitor. Did you leave out the tone controls? If so, did you add a grid leak resistor for the 6Z1P? The original circuit does this via the bass tone control.

Please show us the circuit that you actually built.

[laksaja]

I attach the schematic which im currently working on. I need to apologize, because the original schematic differ from the current one. So i asked questions regarding to my current setup.
This circuit cannot be simpler and i alwayz try to make clear simple things first and then complex ones. The cathode follower was mentioned here and also resitor divider was the issue. When i connected lower 0.47uF cap directly to the cathode of the splitter, the signal was weaker than the anode signal. 2.2k resistor between 100k and cathode did the miracle! When i started this thread, i had the siganl connected directly to the cathode of the splitter. Now it works like a charm Please take a look at the schematics and oscilloscope forms (100Hz, 1khz and 10khz). They have taken at full output power, which is now 6.4v AC, hence 6.4*6.4/4=10W!!!- Ten times more than before.
Only drawback that the feedback is still impossible to connect. The amp start oscillating. i think its due to my 10 dollars output transformers . I cant wait to see this small amp working with real OPT-s.


I thank you all for contributing this topic, now im happy