Cavalli OCL Buffer

[Dxvideo]

Hi all,

As a group of Turkish DIY ers, we started a common project that a clone of Cavalli OCL Buffer with DC servo;
http://www.cavalliaudio.com/oclbuffer/main.php?page=diy
With my PSU design, the total application schematic will be like that;

An externally hosted image should be here but it was not working when we last tested it.

And the PCB design I made;

An externally hosted image should be here but it was not working when we last tested it.

Now I am waiting for your precious comments on this..
Thanks a lot for your advices in advance..

PS: Were using Russian 6N6P instead of ECC99 tube.

 

[Zen Mod]

not as complain but more as reminding - draw on schematic grid to gnd resistor , to prevent possible confuse .

beside that - having grid resistor is always good , in case that anything goes wrong with pot contact ; not of much importance here , where triode is biased with CCS , but anyway ....

also - grid stopper is good thing to have

on link you gave - bootstrapped (cascoded) version is certainly better ( not just in test , but soundwise too) ; then it's real SLCF , as Alan Wright like to call it , and you squeeze all quality possible in that stage .

 

[Dxvideo]

Actually I was thinking about to put a grid resistor too! Thanks for reminding.
And concerning the bootstrapping issue;
I have tried a bootstrap like in this application with Pedja Rogic's famous JFET buffer stage however I found its sound a bit dull. Then I've removed the bootstrap and connect the input JFET directly to the Vcc (like this application) then the trebles came back again. May be the bootstraped version was more musical and the non-bootstraped was more boring...
Anyway, I can add a space for bootstrap to the PCB, who wants to add bootstrap just solder a MOSFET to here..
Thanks for your comments again.
More?

 

[Gordy]

An interesting project for you. However my guess is that your power supply with just CRC will give too much ripple. I suggest more RCRC, or even better to use a TL783 for a regulator.

 

[Gordy]

C14 and C15 are close to V1, so will they get too hot?

 

[Shoog]

Allan Wright claims that every mosfet replaced by a valve brings sonic benefits. My experience says I am inclined to agree with him. Unfortunately it makes the voltages a lot higher and the need for two twin triodes for a full SLCF.

Its possible to achieve a decent CCS in the tail by using a voltage divider from +B down to about 20V and then referencing a triodes grid to that. This then allows you to include a much bigger cathode resistor. This impedance is multiplied by the triodes mu and you have a decent and adequate impedance CCS. I did the same thing but with a pentode as the CCS element for a much higher output impedance. I used a simple mosfet regulator for the screen and the same resistive reference chain raised the grid to 30V. Works a treat.

This is not meant as criticism of your circuit, just passing on my experience

Shoog

 

[Dxvideo]

Thanks a lot for your comments.

Dear Gordy,
- I will move C14 and C15 more distance place against the tubes heat.
- About the CRC filter for B+.... I am confused about that because in the simulation I saw -40dB @100Hz. Isn't it enough you think?

And Dear Shoog,
I think I am not good at tube designing. And this is not my design. So I think I cannot make a critical modification the on design.

 

[Gordy]

- About the CRC filter for B+.... I am confused about that because in the simulation I saw -40dB @100Hz. Isn't it enough you think?

Perhaps reality will match the simulation.
In my own Cathode Follower I used a regulator (TL783) and it was fine. Maybe I worry too much? Maybe your CRC will be enough when combined with the PSRR of the CF.

 

[Karsten Sømand]

Maybe I´m missing something here, but wouldn´t it be very beneficial replacing the CRC with Salas´ very exciting regulator?

Karsten

 

[Dxvideo]

What benefits should we expect with adding a regulator to B+ , do you think?
And whats "Salas' very exciting regulator", is it a special thing?

 

[Shoog]

It has to be a very good regulator to not have a sonic signiture. I avoid them for that reason.
I prefer the results from a simple CLC type power supply.
If your current power supply isn't adquate then the addition of another stage of the same value will do the trick. If you can find and inductor try it - I find it makes for a smoother sound.

Shoog

 

[Dxvideo]

Hmmm...
I've tried CLC type filters in my simulator but they seemed to be oscillated in some high frequencies. So CRC s were work more stable.. Do you agree with this? Or am I making a mistake?
This is a CRC (220u+470R+220u) frequency response;

An externally hosted image should be here but it was not working when we last tested it.

And this is the CLC version of it (220u+1mH+220u);

An externally hosted image should be here but it was not working when we last tested it.

 

[SY]

It has to be a very good regulator to not have a sonic signiture. I avoid them for that reason.

With the inherent high PSR of a cathode follower and the addition of the CCS load, the power supply is almost out of the picture. I have some serious doubts that differences among good regulators in a circuit like this could be heard (as opposed to imagined).

 

[Shoog]

You may very well be right - so why go to the trouble of including one.

Shoog

 

[SY]

They're useful for LF stability and holding the operating point constant. But with the inherent PSR of Cavalli's circuit (or of the Heretical from which it's derived), a simple two transistor regulator is more than sufficient.

 

[Gordy]

It has to be a very good regulator to not have a sonic signiture. I avoid them for that reason.

I prefer the results from a simple CLC type power supply.

It has to be a very well tuned CLC to not have a sonic signature. I avoid them for that reason.

I prefer the results from a simple regulator type power supply.


So two equal but opposite opinions. Except that when the mean value of your mains goes up the operating point of your circuit changes with CLC, and it's noticeable. And when the mean value of your mains goes down the operating point of your circuit changes with CLC, and again it's noticeable. And ad-hock choices of CLC almost always produce frequency peaking, which is also noticeable. And a decent regulator will crush the ripple and noise and provide a rock-solid stable voltage from a decent low impedance which means that you can concentrate your efforts on the amplification circuit.

However I accept that some may prefer the inconsistent and inaccurate response almost guaranteed by an unregulated system.

Also (not necessarily you Shoog) please note that many judge by what I like to term associated-sentiment... they associate the word 'regulator' with over-control in their every-day lives and naturally prefer unregulated / free things in general and hence favour an unregulated supply. It's a psychological issue. Yes, really.

Also, some can not be bothered to design a regulator so they try an unregulated supply and feel it is fine and then instantly preach that unregulated is the best even though they have not tried a good regulated design. It's all psychological.

So, in my humble opinion, a decent regulator provides a very fine foundation from which to build equipment of very good performance.


...and with that Gordy takes off his rant-hat and slinks off back to the shadows...

 

[SY]

Allen Wright said it best in his preamp book- things that are supposed to be tied down should be tied down as tightly as possible. Things that should be loose should be as free as possible.

 

[Shoog]

He also said that his super-regulator was the only regulator transparent enough for his SLCF circuit.


Br Cornelius

 

[Dxvideo]

Gentlemen,

As a final word;
Can we say its enough to use just a 220u+470R+220u pi filter (like in my circuit) for B+?
Or should we use a better power supply?

 

[Shoog]

I would say yes.
If there is any residual hum add another stage - the slight voltage drop should not effect things unduly.
If you leave adequate space, you can add a simple regulator later. If you do you can compare and report your results. Empirical data is always useful to these academic discussions.

Shoog