I will have to buy a precise linear laboratory supply and use a potentiometer to reduce the voltage
I'd just build a fixed voltage divider, say 10:1 or 100:1, and adjust the voltage from the power supply/function generator. Be aware of the trade-off between output impedance / load regulation and dissipation in the resistors. If you use a pot, make sure you don't exceed their maximum dissipation.
If I was doing something like what you're doing, I would look and see if the specs on something like the SPD3303X-E were sufficient for your needs. If not, then you need to look at voltage sources specifically, and that can get expensive for high accuracy.
Is the SPD3303 really any better at the 1 mV level?
In addition to the Elektor project the PVDS2mini comes to my mind as reasonably affordable DC voltage source. Though not sure if Ian is currently able to produce them.
Sorry a $ $ Fluke universal calibrator, 5200 is wrong then.
Maybe I'm dense, but what good does a multifunction calibrator, which can generate a variety of DC and AC voltages (typically up to 1 MHz), currents and resistances do for adjusting a source like a function generator or power supply?