CEC 780CD tube buffer based on 6n1p cathode follower

[yagoolar]

Hi,
The thread should be in two sections I guess: Tubes and Digital Source, but posting here first.

I have decided to follow some of DIYers and improve the sound of a CDP, i.e. C.E.C. 780CD. I thought I would add a simple tube buffer to LC7881. It is built with Russian one 6n1p (6H1Pi) - each triode per channel. Did some simulations and those seemed correct. Yet, after connecting grids to Ch1Out /Ch2Out (pins 1/20) or after opamp voltage buffer CDP sounded like reverb.

Couple of thoughts come to my mind:
1. connect the buffer after LPF, i.e. second opamp in the path, then why use tubes?
2. connect the buffer to DAC outputs and add LPF, as in Musical Fidelity PAN,
3. there is something wrong with my understanding of a simple CF,
4. abandon this "mod" and replace opamps, clock, caps, mute transistors
5

Where did I go wrong?
I will post schematic later.

[yagoolar]

See attached schematic.

[yagoolar]

See attached.

[leadbelly]

2 things I can see: your CF is terrible, #3 on that one, and also you haven't coupled to the output of the DAC properly.

Your CF is a mishmash, too small cathode load and you have somehow merged the grid leak and bootstrap resistors into 1 undersized entity. Poor DAC chip.

Note on the schematic that the input of the 1st opamp stage says +2.4V? That means the signal has a DC component (i.e. when the music is quiet you get constant 2.4V out of the DAC chip.

[yagoolar]

As for DC offset - true - output DC from LC7881 is blocked by 120nF not shown on the schematic; sim results were strange after I had added it.

I know LC7881 is not perceived as the best DAC, yet I would like to start here, and than use AD1865K or TDA1541. I simply learn from practising.

I do not quite understand what you mean by "terrible" and "mishmash" (BTW. we use this phrase in Polish (pl. miszmasz) to express something is totally mixed up, chaotic).
Me-not-understanding-CF is what I experience in real world - I read books, articles etc. with no use and finally I learn I should use complex BCF, WCF SyCF etc. Sims are wrong either? No offence.

[edit]
What is the answer to the question: the reverb sound? Comes from "terrible" CF design or ...?

Thanks anyway.

[SY]

Let's back up a bit. What are you trying to accomplish with this stage? Adding distortion?

[yagoolar]

SY,

I admire your achievments, still this kind of questioning (me) does not help (me) understand why/where I go wrong. Last question is rhetorical. If I go ignorant, show me the right way. Did you mean go #4?

PS.
My RDL84 based on your design works great after 2 years from assembly.

[SY]

No, I'm being absolutely serious. You have to start with specific design goals before anyone can judge if your design is appropriate for achieving them. Very often, people want to add distortion to an otherwise clean circuit, and that's what I'm asking. If the answer is "no," and there's no other reason I'm missing, then the buffer gets you nothing except some heat and expense. If the answer is "yes," then we can go to the next step.

[yagoolar]

Good point.
Goals:
- learn if it can sound better (totally personal perception of music)
Boundary condition:
- if it can, can I do it with 1 tube.

So far the result is: it cannot. Q: why? #1-3? Or else?
#4 is to be considered. If abandoned, no lesson learnt, isn't it?

[SY]

Fair enough. If a cathode follower is designed to have low distortion, it will make no sonic difference. If you want so-called "tube sound," you want to design to get several percent harmonic distortion, weighted toward second and third. That means low impedance load.

In your proposed circuit, the load impedance is indeed low enough to give you lots of 2nd and 3rd. You do want to reference the input to ground, though. That will cause compression of the voltage swing, making the sound very thick and fat.

If you want a transparent CF circuit, take a look at some of the non-servo versions of the Heretical.