This is from the Vitrohm ZCM data sheet. The larger resistor has lower current noise as it should for 20 k and higher. Below, noise no longer decreases as resistance decreases. The smaller resistor has the same behavior below 2 k. This results in lower current noise in the smaller resistor from about 7.5 k.
I don't get why this should be the case. The larger resistor body needs a wider trace to achieve the same resistance as the shorter trace on the smaller body.
PS: Vishay Draloric SMM0207 has the exact same graph for current noise
The non-linearity graph of the Vitrohm is also exactly the same. Was there a standardization or is Vitrohm trying to be a second source, much as many second sources for classic Japanese transistors simply copy the data sheet?
I don't get why this should be the case. The larger resistor body needs a wider trace to achieve the same resistance as the shorter trace on the smaller body.
PS: Vishay Draloric SMM0207 has the exact same graph for current noise
The non-linearity graph of the Vitrohm is also exactly the same. Was there a standardization or is Vitrohm trying to be a second source, much as many second sources for classic Japanese transistors simply copy the data sheet?
Last edited:
This paper has the basics - there is no reason noise and nonlinearity should change at those values - unless that is where the construction is different:
https://pub.dega-akustik.de/ICA2019/data/articles/001261.pdf
https://pub.dega-akustik.de/ICA2019/data/articles/001261.pdf
My guess is that the original data is extremely noisy and sparse (only a few values measured) and thus the plots are rough trend lines only (the sharp kink is a tell-tale), and the horizontal section is the measurement noise floor.
Quite possible, especially for the noise measurement. But why would the measurement of the bigger resistor have a higher noise floor?
My initial guess was that the kink was related to a change in mechanical construction or the type of alloy used. But then I saw that the kink in the nonlinearity graph occurs at the same resistance, not different resistances as the noise graphs.
My initial guess was that the kink was related to a change in mechanical construction or the type of alloy used. But then I saw that the kink in the nonlinearity graph occurs at the same resistance, not different resistances as the noise graphs.
Firstly you can't really speak of noise-floor for these measurements - they are not just current noise, they include flicker noise (measured in µV/V is the clue, current noise is in pA/√Hz normally, i.e. as a noise spectral density - noise floors are typically a spectral density (where white noise is a horizontal plot))
My best guess about this is:
Flicker noise is likely due to charge traps gaining and losing charges at random, causing the actual resistance to change over time. Charge traps on a metal film resistor are likely to be embedded in the surface oxide (think about it!). So geometry of the construction will have an effect on the population of charge traps - and we don't know the film geometry for either size of resistor. Flicker noise is pink noise.
At lower frequencies flicker noise dominates, and at higher frequencies the Johnson current noise (white noise) predominates. If these plots were done as spectral densities they would be flat at high frequency, but rise at lower frequencies where flicker noise picks up.
My best guess about this is:
Flicker noise is likely due to charge traps gaining and losing charges at random, causing the actual resistance to change over time. Charge traps on a metal film resistor are likely to be embedded in the surface oxide (think about it!). So geometry of the construction will have an effect on the population of charge traps - and we don't know the film geometry for either size of resistor. Flicker noise is pink noise.
At lower frequencies flicker noise dominates, and at higher frequencies the Johnson current noise (white noise) predominates. If these plots were done as spectral densities they would be flat at high frequency, but rise at lower frequencies where flicker noise picks up.
The term current noise has more than one meaning. I wish no one would call the flicker noise or 1/f noise of resistors current noise, but unfortunately, it is fairly common to call it that.
The rationale is that you only get a random voltage variation due to random resistance fluctuations when there is a current flowing through the resistor. With a similar reasoning, they might as well call it voltage noise, as you only get a random variation of the current due to random resistance fluctuations when there is a voltage applied across the resistor.
The rationale is that you only get a random voltage variation due to random resistance fluctuations when there is a current flowing through the resistor. With a similar reasoning, they might as well call it voltage noise, as you only get a random variation of the current due to random resistance fluctuations when there is a voltage applied across the resistor.