>>2948441
>the entire system is a laser lock-in amplifier.
oh i see, the circuit you posted is a laser driver? lol im retarded
thats cool, spectroscopy?
>we use Vishay foil resistors as references.
huh, didnt even know those existed :/

here's a drawing of exactly what im trying to do. left is my current setup. we have some photoconductors with resistances of around 500 kOhms. the easy/usual way of using them is in a 1:1 voltage divider. in this configuration, we get around 600 nV/rtHz of noise on the output. this setup is good enough (and cheap) but we want to test devices that have a much larger range of resistances, from 50 kOhms to 2 MOhms, and manually swapping the load resistance for every device is a hassle. i proposed using a current source for biasing instead (pic rel right), but we need to make sure the noise it creates (i.e. the voltage noise generated by the current source driving a 2MOhm ideal resistor) is well below the noise of the device itself. reading up on things, im getting confused about *current* sources generating *voltage* noise. my best guess is voltage noise depends on the load while current noise doesnt, but im not 100% sure.
my original "<300 fA/rtHz" requirement came from taking the thermal noise of a 500 kOhm resistor (128 nV/rtHz) and dividing it by 500k. technically my "real" requirement is less (or equal) noise than thermal at the output of pic rel.