UNIO a minimalist I/V converter

[smms73]

Dirk you have to use the same values i use for the resistors and a 25mA current source. R3 , R4 is 3k5 , R1,R2 is 18 ohms

 

[Joachim Gerhard]

As a said i am a dumb with I/V converters for DACs. I just wanted to know if the offset screws up the PCM1794A. In another thread we talked about canceling the offset from this DAC and then i had the idea if your UNIO can be set up to cancel that offset because it has an offset already. Your circuit is balanced and AC coupled and it plays already fine so there seems to be no problem.

 

[smms73]

Hi Rosendorfer.

It´s funny that you ask that. I have see it an hour ago in the post #15 of the thread

http://www.diyaudio.com/forums/analog-line-level/216557-xcen-balanced-single-ended-converter-2.html

It was you that design it ?

 

[dirkwright]

OK, THD is about 0.001% @ 1kHz, 7.8mA p-p input. Q1 and Q2 are running at a few hundred micro amps, which is not very much. Output is about 1.86Vrms. I don't have spice models for the BC327-40, so used the -25 version of each instead.

 

[Joachim Gerhard]

The XCEN was designed by EUVL and me. I have posted the circuit in post 20 there because it also converts Bal-Unbal. I think that circuit is from a french website.

 

[Joachim Gerhard]

No, i was wrong. It is a website from the Netherlands.Ray's Audio Page

 

[smms73]

1.86 Vrms is strange. you should have 1 Vrms, with 1 Vrms the distortion should be lower.

 

[jean-paul]

Hi I have such a PCB almost completely populated with A grade parts for sale (never used). I used 2SK170 on this one. If someone wants to try it please send me a PM. Very nice output stage !

 

[smms73]

Hi joachim i have started today to read that thread.
i like jfet, but the problem is the increase difficulty to obtain a p channel jfet.

 

[Joachim Gerhard]

Yes, the P-Channel problem. It is worse then solving hunger in Africa.

 

[smms73]

if someone prefer to use jfet. The q1 and q2 (bc337) can be replaced by a jfet like the 2sk170 without changes in the rest of the circuit.

 

[Joachim Gerhard]

That is good opportunity to listen if BJt or Fet sounds better. The circuit is simple enough to do that experiment fast and easy. How about the distortion with the Fets ? I think Bob Cordell has models of N-Channel Fets on his website CordellAudio.

 

[smms73]

UNIO FET

with jfet at input

 

[Joachim Gerhard]

Cool, that could also be a balanced line stage when C4, C5 are taken away and R1, R2 raised to 100kOhm.

 

[Joachim Gerhard]

Maybe the gain is too high then. Something around 10 - 16dB would be good.

 

[smms73]

I have simulate with lsk389 from Cordell model and the distortion is the double of the bc337, but that is normal as the gm of the jfet is lower than the bc337.

 

[dirkwright]

1.86 Vrms is strange. you should have 1 Vrms, with 1 Vrms the distortion should be lower.

I'm getting 51dB of gain. Flat from 10Hz to 100kHz. I didn't put your filter capacitors in yet.

 

[smms73]

The gain is 30db in the simulation and in real circuit. without capacitors the frequence response is wider than that ( 5 Mhz at -3db ). try other transistors, maybe there is something wrong with the bc337 models that you use.

 

[Joachim Gerhard]

To lower gain to say 16dB what is best to do ? I would like to try the Fet version as preamp. Is that 30dB single ended gain of each side or bal. gain ?

 

[dirkwright]

To lower gain to say 16dB what is best to do ? I would like to try the Fet version as preamp. Is that 30dB single ended gain of each side or bal. gain ?

Without loop feedback, gain is controlled by the local feedback degeneration resistors.