Why is it better to use a differential pair to guide the current to the output node, rather then using just one simple switch?
Anyone an idea?
I made a current steering dac and somebody asked me that question but I just took the diff. pair because everbody uses it, i don't know why it is better.

 

If you really use complementary switching signals for V1 and V2, you can use lower peak-to peak input signal swings, the voltage across the current source M3-M4 doesn't change (yielding temperature stability and faster output settling), you will get outputs that are "complementary", and there will be minimal (zero, ideally) change in IDD, which keeps supply noise to a minimum.
In the Electro-Tech forum, I missed your statement that this was in a current-steering DAC. Sorry about that.