"Low noise" 5V linear PSU

[UpLateGeek]

Just looking for some suggestions/advice on designing a low noise linear power supply for a DAC that's powered by 5V USB. I've only ever worked with digital stuff, so I have no idea if my bypassing is good enough or overkill, or if there's any basic analogue circuit design tricks I should know.

Also, I'm guessing the shield should be grounded (the metal case of the DAC is connected to the shield of the USB power connector, coax input and analogue outputs), but should that ground connect to the 0V rail? And if so, should I just use a ground plane, or is it worth it to use a star ground?

Here's my design:



Any comments would be appreciated!

[Zero999]

The ripple rejection can be improved, by putting a fairly large capacitor across R2, say 10µF to 100µF.

Why is R1 a potentiometer? For fixed power supplies, it's better to just use the right resistor values.  R1 = 1k5 and R2 = 4k3 will get 4.988V, which is within 0.24% and better than the NJM2397's tolerance. Even if you manage to adjust it to better than this, the voltage can still change by more than this due to the line and load regulation. If you're a volt nut than you could use E96 values: R1= 1k21 and R2 = 3k48, but it's overkill.

[UpLateGeek]

The ripple rejection can be improved, by putting a fairly large capacitor across R2, say 10µF to 100µF.

Why is R1 a potentiometer? For fixed power supplies, it's better to just use the right resistor values.  R1 = 1k5 and R2 = 4k3 will get 4.988V, which is within 0.24% and better than the NJM2397's tolerance. Even if you manage to adjust it to better than this, the voltage can still change by more than this due to the line and load regulation. If you're a volt nut than you could use E96 values: R1= 1k21 and R2 = 3k48, but it's overkill.

Fine, I'll go buy some more resistors. 

But thanks, both good suggestions.