0xDEADBEEF
- 815
- 1
A few questions about noise. I am doing a noise budget and I am not sure if everything is correct...
1/f flicker noise
Resistors show this noise. Horowitz specifies microvolt per decade frequency per applied volt.
This would give me a false reading on a four point measurement, at least that is how I understand it, because the applied voltage won't really flicker will it. Is this noise real noise in the resistivity? Does it mean the resistance is not very stable over a year? So when I measure the resistance with an AC Voltage of 100KHz will I see a resistance drift in time caused by flicker noise?
Let's say I measure some voltage, and the measurement is such, that it produces the average value over one second. Can I calculate the expected error due to flicker noise? Can I only specify the expected drift till the next measurement?
If I take 1 measurement every 90 seconds for 5 hours. Am I correct in specifying the bandwidth to be the frequencies between 1/(3 minutes) (sampling theorem) upper and 1/(300 minutes) lower frequency, so two decades for the purpose of flicker noise?
Thermal drift noise
How much and how fast does the temperature of a smd resistor or op amp typically change in a metal box due to air movement in a lab? (external power supply, low power chips) What is the frequency spectrum?
Capacitor noise
Is the noise on a high pass the same as on a corresponding low pass?
1/f flicker noise
Resistors show this noise. Horowitz specifies microvolt per decade frequency per applied volt.
This would give me a false reading on a four point measurement, at least that is how I understand it, because the applied voltage won't really flicker will it. Is this noise real noise in the resistivity? Does it mean the resistance is not very stable over a year? So when I measure the resistance with an AC Voltage of 100KHz will I see a resistance drift in time caused by flicker noise?
Let's say I measure some voltage, and the measurement is such, that it produces the average value over one second. Can I calculate the expected error due to flicker noise? Can I only specify the expected drift till the next measurement?
If I take 1 measurement every 90 seconds for 5 hours. Am I correct in specifying the bandwidth to be the frequencies between 1/(3 minutes) (sampling theorem) upper and 1/(300 minutes) lower frequency, so two decades for the purpose of flicker noise?
Thermal drift noise
How much and how fast does the temperature of a smd resistor or op amp typically change in a metal box due to air movement in a lab? (external power supply, low power chips) What is the frequency spectrum?
Capacitor noise
Is the noise on a high pass the same as on a corresponding low pass?