dac I/V convertion with very low distortion

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I made another experiment. Instead of one battery i tried two with a center tab. They are now parallel to the output caps. Offset is still zero. I did not listen to it but i will.
Maybe someone can make a simulation if that works better.

Where is the center tap connect? ground? It's probably fine for a moving coil head amp but maybe not for a DAC I/V converter. I could be wrong though.
 
What do you like to win with the BISS ? I asume less distortion.
In terms of speed we are alraedy at 4.5Mhz and a gain over 30 and noise is really low.
And with a much simplified circuit.
Anyway, i am open to sugestions and look forward to your full steam implementation.

sorry for the long delay.

forget the biss, at this current level maybe the difference in distortion is minimal.

what i propose to you is connect your circuit like the Pass zen and you get rid of the electrolytics in the signal path, use 10uf film capacitors, you can also use batteries.

Other porpose is to connect the collectors of helper transistores to the input like shinja suggested in one post, you will get less harmonic distortion (you can try this with your circuit).
 

Attachments

  • cen_Joachim2.jpg
    cen_Joachim2.jpg
    79.1 KB · Views: 377
I made another experiment. Instead of one battery i tried two with a center tab. They are now parallel to the output caps. Offset is still zero. I did not listen to it but i will.
Maybe someone can make a simulation if that works better.

If you increase voltage in the input transistors, you will get less distortion.
But at this voltage output levels, the difference is minimal, with voltages like 1 Vrms will make a difference
 
I measured ca. 80mV over the 5.1 Ohm emitter resistors ( i did not have 4.7 Ohm on hand ). That is 0.08 / 5.1 = ca. 16mA.
I have also thought further and your new circuit comes handy. With my floating supply there is the problem that the helper transistors are connected to ground with the bases.
That way they easy pick up hum on the ground line that should be the shield. I tidied up the layout but i could not get the hum away 100%. Other noise like hiss is extremely low though.
 
I put the new circuit in the system and after mucho experimentation the hum is gone.
What a releave ! So it was not the bases connected to ground. This has consequences on the sound, especially in the bass that is now much more robust and tight. That small hum was fogging a bit and took dynamics and resolution. The light on the feat and liquidity is still there but on a much more firm fundament. I will now try the double battery option.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.