Is there anyone who has the idea what resolution (xxx ppm to the central freq) a si-based oscillator can at most achieve over temp or process respectively.
in common case, without using calibration at the system level, a thermal stability 100...500 ppm/'C can be achived. In general case 1/RC frequency oscillator can be designed,- independent from process/voltage variation of active components. R can be external or built in. In later case it can be composed from two types with total ~zero TC. Capacitors can have 20...100 ppm/'C, depending from it's type and process.
All above is IMHO.
Is there any recommended structure? i.e. for RC oscillator
mikersia said:
in common case, without using calibration at the system level, a thermal stability 100...500 ppm/'C can be achived. In general case 1/RC frequency oscillator can be designed,- independent from process/voltage variation of active components. R can be external or built in. In later case it can be composed from two types with total ~zero TC. Capacitors can have 20...100 ppm/'C, depending from it's type and process.
All above is IMHO.