[time-nuts] Low noise power supplies?

Garry Thorp GThorp at pascall.co.uk
Thu Jan 31 08:41:28 EST 2013


If the main concern is about low-frequency noise the LT1763 might be worth considering for a simple solution, as it is capable of lower noise than the data sheet shows. With a bypass capacitor (Cbyp) fitted, the noise floor is ~30nV/rtHz. Although the data sheet only recommends up to 10nF for Cbyp, higher values continue to reduce the low-frequency noise. Correspondence with Linear Tech confirmed that it is OK to increase Cbyp as long as its leakage current is low enough and the output shunt capacitance is large enough.

With Cbyp=22uF and metal film resistors used, the noise at 1Hz (with +10.6V output) came down to not much above 30nV/rtHz. The downside is that the output voltage rise time is a few minutes! (And if you short the output, it takes just as long for it to come back again.) Maybe not a problem for lab use.....

The LT1763 produces a noise peak around 150-200kHz if Cbyp is fitted (any value), but increasing the output capacitor to 100uF brings the peak down to ~30nV/rtHz.

Garry Thorp


More information about the time-nuts mailing list