Converting 5v to 3.3V for GPS purpose while keeping noise at the lowest

I'm trying to create a very simple module that consist in powering a GPS ublox - NEO-M8U - with a micro USB & using an active antenna.

Later on, I will add more features but I just want to make sure I'm doing things the right way by keeping the noise at the lowest.

Here's what I did using information I found mainly on stackexchange.

I decided to go only with a 10uf capacitor for the Vin. Some people are in using 22uf, 100nf & 10uf in parallel. Would that make sense to use more capacitors on this side on the circuit ?

For the GPS, I went with this :

I'm just not sure about the inductor. The problem I have is that I want my card to be directly assembled by the PCB supplier and they did not have the 27nh recommanded by the GPS manufacturer. Also I really don't know how to choose such inductor. Do people have good practices for that ?

My last question is for the board design. Should I have the biggest capacitor the closest to Vin ?

• Your connection to the antenna looks odd. Do you mean to have D connected to $V_{CC}$? – awjlogan Dec 4 '17 at 18:02

On the USB side, the specification is for 10 $\mu$F maximum. In my experience, higher values are usually fine but that's anecdote, so stick with the specification. It's followed by a linear regulator anyway.
You almost certainly don't need that amount of capacitance on the 3V3 line either, and just adding capacitors does not lower the noise (and you should think about what you need as well). Stick to the datasheet or reference implementation, unless you need something radically different. For modules like this, a good starting point is a 47 $\mu$F electrolytic cap with a parallel 10-100 nF ceramic cap.