New driver stage for dynaco mk3

[gabdx]

I want to re-design the driver stage of the Dynaco MK3.

My idea is the following:

- no pre-amp, (I have already too much gain from the phono)

- volume pot 100k log

- phase splitter is LL1545A

- 6sn7 as drivers ( this is to be determined)

- a good dose of global feedback (this is to be determined how it is applied)

- kt88 in ultralinear with second grid 1k resistor with 40% taps.

 

[indianajo]

That is a lot of effort for monaural sound. I buy mono LP's to hear Toscanini or CR Fine at work, but stereo is one of those improvements that I can actually hear. You sure you don't want to buy a new chassis and a pair of matching output transformers? KT88's are pretty big, and hot.

 

[artosalo]

I want to re-design the driver stage of the Dynaco MK3....

Do you mean to replace the 6AN8 with 6SN7 and the transformer ?
What sensitivity do you plan to get ?
And with what amount of GNFB ?

 

[gabdx]

I have 2 and they are actually working. I would use something like 15 db of feedback.

I thought about this:

100k volume - 6sn7 cathode follower with active load 6sn7

this is going to a LL1545A transformer for creating a balanced output,

the balanced output is feed to the long tail pair 6sn7 , feedback is applied to this part.

 

[jgf]

I was mulling over similar ideas, maybe without the input transformer. I'm thinking the 6SN7 won't have enough gain to apply global feedback, maybe consider 6922 or 6gk5?

 

[gabdx]

I was mulling over similar ideas, maybe without the input transformer. I'm thinking the 6SN7 won't have enough gain to apply global feedback, maybe consider 6922 or 6gk5?

Hi! I have too much gain as it is. I can cut easy 20 db.

The idea behind the input phase splitter transformer was my discovery that the long tail pair is far from perfect. I will draft a schematic...

 

[jgf]

Sounds cool! I'd like to see it.

 

[gabdx]

basically the idea is this one : Audio Virtual Systems very similar... I even use the same output trans...

 

[gabdx]

Hi, I did more research and found that I could connect the volume pot to a doubled 6sn7 in triode mode connected directly to a LL1660S connected directly to KT88...

The feedback would be feed to the cathode resistors 6sn7 cath.

There would be 2 capacitors at the balanced ends of the LL1660S because I want to bias the tubes individually with grid bias.

Shematic similar to the SUN audio amplifier for EL34 PP...

LL1660S specifies 220 V R M S max

 

[gabdx]

I just don't know for the feedback and values of resistors.

I am figuring out if it needs capacitors because the KT88 are biased individually per grid and it could upset the transformer ability to split the music if there is modulation and different fields.

*** maybe I will abandon this idea and adopt this instead: input -- long tail pair -- capacitors -- KT88

 

[gabdx]

I will try this out before going to transformer splitter...

I don't know which way to take 1, 2, 3 and the need for the capacitor on the grid 2. I learned that I could use feedback from the output transformer and send it to Grid 1 or Grid 2, but I ignore the pro or cons.

 

[artosalo]

No CCS will work without some bias voltage.
You should give positive bias to tubes or connect the lower end of the CCS to negative voltage instead of GND.

For example LM317 regulator works fine with some 20 V bias.

 

[gabdx]

Hi, I actually use figure 1 from lm334 with the positive voltage of the cathodes. I know it can be improved with thermal stability (figure 3) and it can oscillate if the stage overloads because not enough cathode potential will exist.

 

[gabdx]

I will try this out before going to transformer splitter...

I don't know which way to take 1, 2, 3 and the need for the capacitor on the grid 2. I learned that I could use feedback from the output transformer and send it to Grid 1 or Grid 2, but I ignore the pro or cons.

I will use Grid 2 similar to LEAK amplifiers. Now about the values of resistors I have no clue to give me a start...

 

[gabdx]

I just don't know for the feedback and values of resistors.

I am figuring out if it needs capacitors because the KT88 are biased individually per grid and it could upset the transformer ability to split the music if there is modulation and different fields.

*** maybe I will abandon this idea and adopt this instead: input -- long tail pair -- capacitors -- KT88

I will just do that without capacitors, 0 caps amplifier (haha 2 transformers) I saw that the transformers allow for indifidual biasing current to circulate in both half without any issues.