MPP

[RCruz]

Saw that too, I think its a lot of components... Paradise is a good example of how that can be achieved with much less components, and in my opinion much more elegant. But I am not a phono stage designer at the level of you or Self, so who am I to tell....

This is what "Paradise regained" looks like right now. Schematic is captured, placement of components done, routing has started. This now has the current source on the left side, input stage in the middle, the RIAA elements and current source next to it and very close, and the output buffer. The servo is at the bottom in the middle, close to the buffer. The shunt regs are where they were before. Board dimensions are the same as before. More as it progresses.....

Very nice

Maybe this simpler design is the way to go.

 

[Joachim Gerhard]

I had a loger discussion with MiiB and he thinks the problem could be in the current mirror. I rebuild it now from a Cascode Wilson ( with feedback ) to a Cascode Widlar ( without feedback ) that the new Paradise also has.

 

[hesener]

I had a loger discussion with MiiB and he thinks the problem could be in the current mirror. I rebuild it now from a Cascode Wilson ( with feedback ) to a Cascode Widlar ( without feedback ) that the new Paradise also has.

It may be "brute force" but do you think it might be worthwhile to check if we use 2N4401/4403 in the current mirrors? They are ok-ish lower noise and also lower current gain, but I am not sure about the impact on noise and circuit performance...

 

[Bksabath]

It may be "brute force" but do you think it might be worthwhile to check if we use 2N4401/4403 in the current mirrors? They are ok-ish lower noise and also lower current gain, but I am not sure about the impact on noise and circuit performance...

I was thinking abaut 2N4403/4401 but was reluctant to make suggestion as my experience in this is very limited.

When I bought the 337 I also got a 1000 strip of 2N4403 for other project
(Sallas uses them on Reflektor as current mirrors).

I can run the batch on the Curve tracer and sorth same out for you to try

If possible I wuld like to swap same 2N4403 for 2N4401.

 

[Joachim Gerhard]

This transistors have comparable noise then the BC327, 337 but much lower Hfe.
They will work but i would like to stick with the BCs in the new Paradise, also in the mirror. The new circuit should work without problems.
Should a problem come because of the high Hfe we can use BCs from a lower Hfe group.

 

[Joachim Gerhard]

For example there are BCs with extention 16 ( BC327-16, BC337-16) that have Hfe from 100 - 250. The 2N4401 and 2N4403 have Hfe around 60 and that is too small i think. Also we found that the BCs are true complimentary and very low distortion and they sound very good. They also have very low flicker noise. I do not like to give that up after all the struggle we went through. They are "our" transistors.

 

[Bksabath]

Tanks Joachim.

Briliant explanation

 

[hesener]

Yes, thats very clear. Have you made progress with the four Rev1 boards?

 

[hesener]

Paradise regained

Here is the new layout, after checking we'll proceed with the new PCBs. Hope you like it!!

 

[Benedetto]

Looks great,
but the pattern for the earlier discussed "heat dissipators" has changed.
Shall we go further in it ?

 

[Joachim Gerhard]

Very good, Alfred, and more compact.

 

[Joachim Gerhard]

Alfred, i started to modify the current mirrors. I put that in the lab today.

 

[hesener]

its the same heatsink as before, if that is what you mean. the board size etc is the same.
Note this now has jumpers to test the shunt regulators independently, before connecting to the preamp...

 

[hesener]

Very good, Alfred, and more compact.

This time I made sure the input connectors, input structure, and output structure are as close as possible and have heavy-duty GND connections... ;-) learned from that mistake....

 

[Joachim Gerhard]

I think Benedetto is talking about cooling the input transistors. I think that is not necessary.

 

[Joachim Gerhard]

Have send you an e-mail. I see that you use 4 x 4700uF and not 2 x 4800uF plus x 6800uF as shown on the schematic. The schematic is OK otherwise as far as i can tell. MiiB should double check that.

 

[Benedetto]

I think Benedetto is talking about cooling the input transistors. I think that is not necessary.

Exact, but we discussed to keep them at more or less the same temperature.

 

[Joachim Gerhard]

They do not get very warm anyway. We are on track solving the mystery of oszillation.

 

[hesener]

just to let you guys know, the layout of the second revision of paradise is done and being checked as we speak. we should be ready to order the new PCBs pretty soon....

 

[RCruz]

I would like to beta build those also.