Are low cost E12 series resistors linear enough in my case?

[bonzer]

Hello everyone! I know that E12 resistors with 10% tollerance can't give me the exact nominal value. What I'm interested in is linearity:
if it's for example instead of 10kOhm - 990Ohm, will this value remain stable for a wide range of voltages? In my case it is not that large but anyway : 0.1V - 20V DC? Unfortunately I can't access to lab.

I'm talking about DC current but anyway if you know tell me if it could still be valid for example for triangle or square wave 20kHz...?

[madires]

As long as the voltage isn't higher than rated maximum. Based on the construction resistors can have inductance, but for 20kHz it shouldn't be an issue.

[Siwastaja]

Also note the temperature coeff specification, which could cause non-linearity seen at very low frequencies: basically the resistor heating up causes a drift in value. This is easy to avoid by derating power dissipation and hence, self-heating of the resistor, and is often irrelevant at signal levels if the instrument sits at a fairly constant room temperature.

If you have some unknown cheap resistors with no specsheet, then the best course of action is to minimize temperature changes.

Note that +/-0.5%, 50ppm/degC resistors are very cheap as of today, so if you need to order parts anyway, it's often easier to just overspec the resistors a bit if you are unsure.

[Zero999]

Hello everyone! I know that E12 resistors with 10% tollerance can't give me the exact nominal value. What I'm interested in is linearity:
if it's for example instead of 10kOhm - 990Ohm, will this value remain stable for a wide range of voltages? In my case it is not that large but anyway : 0.1V - 20V DC? Unfortunately I can't access to lab.

I'm talking about DC current but anyway if you know tell me if it could still be valid for example for triangle or square wave 20kHz...?

Where did you get these resistors from? Please post a link?

10% tolerance isn't that common these days. Most E12 resistors assortments are really 5%, but it wouldn't surprise me if you see them being advertised as 10%, since any E12 value will be within 10% of tolerance of the target.
https://www.velleman.eu/products/view/?id=350890&country=us&lang=en

[TimFox]

Ohm's Law is only a good approximation.  The deviation from linearity in the current vs. voltage relationship depends on the material and construction.  Wirewound resistors are extremely linear, as are metal-film (or thin-film SMD) and metal-foil resistors.  Cermet (or thick-film) and carbon are less so.  Very few data sheets quantify this, usually as a voltage co-efficient of resistance in ppm/V.  As a practical matter, the worst linearity is found in short SMD thick-film resistors of high resistance value, where the voltage gradient along the path is large.  In critical applications, I found 50 megohm thick film 0805 devices to have substantial non-linearity at 10 V, worse than the same manufacturer's 1206 parts.  The non-linearity, however, is independent of the manufacturer's sorting tolerance for initial resistance.
Often, audio designs require resistor pairs matched in value to 1% or better.  Any DMM can do that match when sorting, even if the absolute accuracy is worse than the match requirement.

[Zero999]

Ohm's law is not an approximation, but a fundamental law of physics. Yes, it's true some materials are non-Ohmic, but Ohm's law still applies, it's just the resistance is dependant on some other property, such as voltage, current, temperature etc. Yes, I know I'm being pedantic here and applying Ohm's law to something like a diode is futile, because its resistance is extremely non-linear.

Practically speaking, for beginner purposes, in most applications where 10% tolerance is adequate, non-linearities over 20kHz, 990R to 10k and 0V to 20V, bandwidth/ranges are a non-issue. The only exception might be in some very low distortion/noise applications, but as a general rule you'll want better tolerance than 10% for those, so it's a moot point. Using as physically larger resistors helps because the voltage gradient and heating due to power dissipation are minimised and low noise/high linearity materials such as metal film help, but again, if the schematic only specifies 10% tolerance, noise/linearity are normally a non-issue.

[TimFox]

Ohm’s law as usually stated is that the current through a conductor is directly proportional to the voltage applied to it: 
I = V/R.
Physically, in materials such as metals, this is a complicated process due to electrons interacting with the atoms in the solid structure.  Otherwise, the electrons would accelerate as they proceed along the voltage gradient, rather than achieve a limited “drift velocity”.  The proportionality fails noticeably, for example, if the conducting path is short in length compared to the mean-free-path for scattering, as in some micro-electronic structures.
Maxwell’s equations are examples of fundamental laws, where the proportionality is exact in vacuum.
Similarly, the capacitor equation Q = C V is exact for an ideal vacuum capacitor, but the proportionality fails for practical capacitors with a dielectric.  The proportionality is good enough for most purposes with “good” dielectrics such as polypropylene film or C0G ceramic, but is measurably bad for “bad” ceramics that demonstrate ferroelectric behavior and dielectric absorption.

[David Hess]

As far as I know, carbon film resistors have a worse voltage coefficient of resistance than metal film and other precision resistors but I think the only way to really know is to test it.  However I suspect it is an insignificant source of error compared to their temperature coefficient.

[dmills]

Voltage coefficient should be in the datasheet, for all that some manufacturers can be a bit coy about admitting there is such a thing.

If designing something where it matters, just pull the datasheet and check, or pick your parts to fit your requirements.

[David Hess]

Voltage coefficient should be in the datasheet, for all that some manufacturers can be a bit coy about admitting there is such a thing.

If designing something where it matters, just pull the datasheet and check, or pick your parts to fit your requirements.

I have only seen it listed for some precision parts.