1) DC offset - almost none
2) common-mode noise - common mode noise is reduced in a diff amp stage
3) difference in amplitude between + and - signals (impossible to make these identical with active circuits) - hello current source
4) isolation of ground between DAC and amps - entirely different topic
You're right in that it's not perfect. You should also realize that none of this is perfect. Especially a transformer!
Transformers are for isolation and impedance matching. They are inefficient and sloppy. They are not extremely linear.
I've heard good music coming through a xformer, who hasn't. I just would not use one unless I had to.
Let's just agree to disagree. From a design methodology standpoint I prefer to keep nonlinear devices, like transformers and capacitors, out of the signal path, and bias devices in their linear region.
Here's a little review for you:
http://www.ecircuitcenter.com/Circuits/BJT_Diffamp1/BJT_Diffamp1.htm
If you bias this with a current source, it's almost perfect. Certainly as perfect as any transformer.
Balanced in, balanced out, and you set the bias voltages so you bias your next stage, which can also be fully balanced. Nothing new.
RM
2) common-mode noise - common mode noise is reduced in a diff amp stage
3) difference in amplitude between + and - signals (impossible to make these identical with active circuits) - hello current source
4) isolation of ground between DAC and amps - entirely different topic
You're right in that it's not perfect. You should also realize that none of this is perfect. Especially a transformer!
Transformers are for isolation and impedance matching. They are inefficient and sloppy. They are not extremely linear.
I've heard good music coming through a xformer, who hasn't. I just would not use one unless I had to.
Let's just agree to disagree. From a design methodology standpoint I prefer to keep nonlinear devices, like transformers and capacitors, out of the signal path, and bias devices in their linear region.
Here's a little review for you:
http://www.ecircuitcenter.com/Circuits/BJT_Diffamp1/BJT_Diffamp1.htm
If you bias this with a current source, it's almost perfect. Certainly as perfect as any transformer.
Balanced in, balanced out, and you set the bias voltages so you bias your next stage, which can also be fully balanced. Nothing new.
RM