RES11A: Meaning of each drift spec

Hi Ryuichi-san,

The drift specifications are best understood by relating them back to the room temperature spec definitions. See below I have collected each room temperature error term with its relevant temperature drift term.

1.) Absolute tolerance (per resistor) and Absolute temperature coefficient of resistance (per resistor).

These specifications refer to the exact resistance value of each resistor in the divider pair. For example, the nominal resistance of RIN is 1kΩ but the exact resistance could be 950Ω.

As the RES11A is designed to be used as divider pairs rather than individual discrete resistors, the absolute resistance spec is not as important. The ratio tolerance and matching tolerance (specifications 2 and 3 below) are much more significant.

2.) Ratio tolerance of divider and Divider temperature coefficient of resistance.

These specifications refer to how closely each divider ratio matches the nominal ratio, where the divider ratio is defined as RG / RIN. For example, the nominal ratio for RES11A40 is 4 Ω/Ω and the exact ratio of a divider could be 4.001 Ω/Ω.

This divider ratio term determines the voltage accuracy when used as a voltage divider and the gain accuracy when used as an amplifier's gain network.

3.) Matching tolerance of dividers and Matching temperature coefficient of resistance.

These specifications define how close the ratios of each divider are to each other. For example, if divider 1 has a ratio of 4.001 Ω/Ω then divider 2 may have a ratio of 4.0015 Ω/Ω.

This matching tolerance is important for determining common-mode rejection and gain accuracy of fully-differential amplifiers.

For more information on the matching tolerance specification and common-mode rejection, please refer to the following application note: Optimizing CMRR in Differential Amplifier Circuits With Precision Matched Resistor Divider Pairs

Ryuuichi machida said:

If one of "RIN1" of RES11A tend to observe higher drift spec (Ex +0.05ppm/℃), "RG1" which is included same RES11A also tend to observe same polarity (In this example, temp drift of RG1 also observe "+"(positive) drift.)

You are correct that the absolute drift terms of RIN1 and RG1 are highly correlated, and this feature preserves the ratio tolerance over temperature (Divider temperature coefficient of resistance). See the following excerpt from RES11A datasheet.

Best Regards,

Zach

Hi Ryuchi-san,

The nominal resistance of RIN for all RES11A variants is 1kΩ. The exact value for RG depends on the ratio selected and the absolute tolerance specification. For example, RES11A40 has a ratio of 4:1 which means the nominal RG is 4kΩ, RG/RIN = 4kΩ/1kΩ.

The exact value of each resistor depends on the tolerance as defined by the absolute tolerance specification. However, when used as a ratiometric pair, the exact value of each resistor is much less important than the tolerance of the resistor ratio.

Taking RES11A40 for example, exact resistor values could be 3925 Ω for RG and 981 Ω for RIN. However, the ratio of these resistors is 3925/981 = 4.001 which results in very low gain error.

Regards,

Zach

The number I gave is just an example showing one possible value for RG and RIN. Every individual resistor will have a different exact value based on a statistical distribution. If we take the typical absolute tolerance of +/-2%, we can rearrange the formula to show that Rx = Rnom * (1 ± 2%)

For Rin, Rnom = 1kΩ and the typical Rin is therefore 1kΩ * (1 ± 0.02) which means about 67% of the devices will have RIN between 980Ω and 1020Ω

The same analysis can be done for RG, you simply have to replace Rnom with the correct value depending on your selected ratio. For RES11A40 the nominal RG is 4kΩ and the typical RG is 4kΩ * (1 ± 0.02) which means about 67% of the devices will have RG between 3920 Ω and 4080 Ω.

I hope this helps.

Zach