D1 I/U converter...

[Skorpio]

I am building the output stage from the D1 (I/U-konverter and buffer).

I will use it together with a Denon DCD-2560 with diferential AD1862 DACs, so I will use the diferential mode on input.

Looking at the circiut and also the SOZ gives some questions:

Why are there no connection between the Sources on the two halves of the circuit (like the resistor between Sources in SOZ)?

Instead of two 3K3 resistors to the negative supply, could I use a CCS? And must I use two or could I use on and let it go to junction between to small resistors, each going further to the Sources? As in a normal long-tail pair?

Would a resistor between Sources (100-500R) give problems for the DACS?

[Skorpio]

The circuit is running now and performing quite well...next step will be to try it with bipolar devices with a temperature compensated bias scheme...

[Skorpio]

The circuits worst technically limitation is the poor PSSR, but it performs much better that the op-amp circuit it has replaced....perhaps PSRR is overrated?

[relder]

Originally posted by Skorpio
...perhaps PSRR is overrated?

If you're running it balanced then there's perhaps some truth to that.

[relder]

Originally posted by Skorpio
Looking at the circiut and also the SOZ gives some questions:

Why are there no connection between the Sources on the two halves of the circuit (like the resistor between Sources in SOZ)?

Instead of two 3K3 resistors to the negative supply, could I use a CCS? And must I use two or could I use on and let it go to junction between to small resistors, each going further to the Sources? As in a normal long-tail pair?

Would a resistor between Sources (100-500R) give problems for the DACS?

If you mean a resistor between the sources of Q3 and Q6, that wouldn't really do anything in a static analysis. The sources of Q3 and Q6 are supposed to be at a constant voltage (0V or whatever the DAC likes) so there's nothing for the resistor to do. In real life where Q3 and Q6 are making quick changes to adjust to what the DAC is outputting I'd think a resistor would only mess things up, but it's easy enough to play with if you want. This is an I/V converter, not a voltage amp.

You could use a CSS, just make sure you think about what happens when you fire the circuit up so that your DAC doesn't get a short blast of +- 30V. Maybe that's fine, maybe not, I don't know. You'd have to use two CCS per channel or it would seem performance would suffer, you want the current on each side to be constant, tying the two sides together through won't do that as well.

[lineup]

Originally posted by Skorpio
I am building the output stage from the D1 (I/U-konverter and buffer).

I will use it together with a Denon DCD-2560 with diferential AD1862 DACs, so I will use the diferential mode on input.

Looking at the circiut and also the SOZ gives some questions:

Why are there no connection between the Sources on the two halves of the circuit (like the resistor between Sources in SOZ)?

Instead of two 3K3 resistors to the negative supply, could I use a CCS? And must I use two or could I use on and let it go to junction between to small resistors, each going further to the Sources? As in a normal long-tail pair?

Would a resistor between Sources (100-500R) give problems for the DACS?

There are a few good threads here in Pass Labs forum
dealing with Pass D1 DAC.
For example this: modifying the D1 I/V stage

[Skorpio]

Originally posted by relder



If you mean a resistor between the sources of Q3 and Q6, that wouldn't really do anything in a static analysis. The sources of Q3 and Q6 are supposed to be at a constant voltage (0V or whatever the DAC likes) so there's nothing for the resistor to do. In real life where Q3 and Q6 are making quick changes to adjust to what the DAC is outputting I'd think a resistor would only mess things up, but it's easy enough to play with if you want. This is an I/V converter, not a voltage amp.

You could use a CSS, just make sure you think about what happens when you fire the circuit up so that your DAC doesn't get a short blast of +- 30V. Maybe that's fine, maybe not, I don't know. You'd have to use two CCS per channel or it would seem performance would suffer, you want the current on each side to be constant, tying the two sides together through won't do that as well.

The reason for the resistor between the soruces is that I want some interaction between the DACs in order to use the extra 2x oversampling feature from the player (se thread: Need help with this I/V stage replace...).

But perhaps I better make it simple, as no balanced output is needed.....