Zen -> Cen -> Sen, evolution of a minimalistic IV Converter

[EUVL]

In my own experience there is no need to ground the battery cases.
It is likely you have some input output coupling.
You may try gate stopper resistors of say 100R for each FET.
And you should also trim the current source to zero the output DC.
The input will then be automatically zeroed.

Without a schematics, it is impossible to figure out what exactly the problem is from here.

But I salute you for the courage to experiment.
You are the only one so far to do a bread broad test.


Patrick

[Joachim Gerhard]

I build a breadboard version of the SEN. I send it to Mathias Kurz of Solid Core Audio to test it in an old Phillips CD player.

[jameshillj]

A question & please pardon my ignorance -

Is the value of the output impedance the same as the Riv? If so, a buffer on the o/p to reduce it would need another seperate supply with it's 0 volt tied to the dac's pin 5 gnd point, same as signal 0 volt on the o/p socket?

Perhaps this been covered before, but ......

[EUVL]

> Is the value of the output impedance the same as the Riv?

Yes. If you use an Riv at 2.2k or 1.4k, I consider that low enough to not warrant an additional buffer.

> If so, a buffer on the o/p to reduce it would need another separate supply with it's 0 volt tied to the dac's pin 5 gnd point, same as signal 0 volt on the o/p socket?

Yes. A buffer would need fixed rail supplies, separate from this circuit.
And if you wish to do that, then you may also consider using the buffer as a Sallen Key filter.

See also :

Anybody using the new ESS Vout DAC (ES9022)?

> Perhaps this been covered before, but ......

Yes, in the discussion with Calvin about the Bipolar version of CEN.

As also mentioned, the bipolar version, as well as the current mirror version, has significantly more noise.
The original JFET SEN / CEN circuits are so quiet, that you hear absolutely nothing on a headphone with max volume.
The difference is not subtle.

This is all thanks to the separate current loop of bias and signal.
And the floating supply is a essential feature of that.




Patrick

.

[EUVL]

> I build a breadboard version of the SEN.
> I send it to Mathias Kurz of Solid Core Audio to test it in an old Phillips CD player.

And his feedback ?


Patrick

[jameshillj]

Thanks, Patrick.

[ryanj]

Originally Posted by EUVL

In my own experience there is no need to ground the battery cases.
It is likely you have some input output coupling.
You may try gate stopper resistors of say 100R for each FET.
And you should also trim the current source to zero the output DC.
The input will then be automatically zeroed.

Without a schematics, it is impossible to figure out what exactly the problem is from here.

But I salute you for the courage to experiment.
You are the only one so far to do a bread broad test.


Patrick

Thanks for the encouragement.

I do have some dc offset in the output of one channel. Whats the best way around this, trimmers in series with the 560k resistors?

Ryan

[ryanj]

Sorry, the resistors I refer to are in the original cen circuit, R1 and R2.

[mfrimu]

I'm still waiting for my Linear Audio issue.

I'm wondering (as I'm ordering parts right now) if there would be any benefit to upgarde the 220uF Nichicon ES to something like MUNDORF or Blackgate NX or Elna BP/NP ?

Thanks

[EUVL]

> I'm wondering (as I'm ordering parts right now) if there would be any benefit to upgarde the 220uF Nichicon ES to something like MUNDORF or Blackgate NX or Elna BP/NP ?

Can you still get Elna NP ?
Can you get a reasonably sized Mundorf (MKP?) for 100uF, or even 22uF ?

I already explained in the article why the "distortion" (i.e. non linear V-I relationship) of the caps are unimportant, as long as current in = current out (Kirchoff's law). One of two profs asked me by PM why I did not simply use ordinary Eleytrolytics (and they will work just fine).

But if you want to try Blackgate NX, please do.
And come back and tell us whether you can hear a real difference.

I would probably try a 10uF MKS-2 one day. But it has no priority.


Patrick