TI's new RES60A makes high-voltage attenuation easy!

[laichh]

An alternative to Caddock, Vishay, & VPG divider.


https://www.ti.com/product/RES60A-Q1

[mendip_discovery]

Tempting solution for an issue I have of measuring output of some Hipot testers, it states it is happy up to 4k for 60 seconds. I wonder if it could be pushed to 5kV without long term issues.

[HalFoster]

Very nice!  And they have a related low voltage part with dual matched dividers as well: https://www.ti.com/product/RES11A

Hal

[razvan784]

Tracking TC 1 ppm/K typ., 5 ppm/K max.
Voltage Coeff. 2 ppm/V typ.
Very nice.
I see one slight drawback, the 15% absolute tolerance, which does allow you to claim >10M input resistance, but not 10M precisely.
Long-term drift and humidity sensitivity might be some other issues.

[laichh]

Tracking TC 1 ppm/K typ., 5 ppm/K max.
Voltage Coeff. 2 ppm/V typ.
Very nice.
I see one slight drawback, the 15% absolute tolerance, which does allow you to claim >10M input resistance, but not 10M precisely.
Long-term drift and humidity sensitivity might be some other issues.

Voltage coefficient of 2 ppm/V could be an issue! Imagine up to ±2 V of measurement error when measuring a 1000 V source, that is between 998 V ~ 1002 V, or 3½ digit resolution limited just by the voltage coefficient alone.

In contrast, the similar 10 MΩ voltage divider from Vishay & Caddock has a voltage coefficient of less than 0.02 ppm/V.

https://www.vishay.com/docs/60043/cns471.pdf

https://www.caddock.com/online_catalog/mrktg_lit/TypeHVD.pdf

https://www.caddock.com/online_catalog/mrktg_lit/TypeUSVD.pdf

[Kleinstein]

The voltage coefficient would be a problem. However the main nonlinearity is usually more a thermal thing and thus more a square law, not a linear part as implied by giving ppm/V.  The thermal effect is a principle limitation for such a small case - for improved accuracy some better DMMs use additional heating to compensate. For the thermal effect it also make a difference which voltage is assumed (e.g. 500 V , 1000 V or maybe 1700 V). The RES11 data-sheet shows using a rather high voltage used for the test of the voltage coeffcient.

For the Vishay and Caddock resistors the voltage coeffient given in the data-sheets may be just the the linear part, so excluding the direct thermal effect on the TCR. A linear effect is possible from uneven heating and thermal EMF. So the numbers may not be comparable.

A problem with the high voltage dividers from Ti is there ESD rating - when used for a DMM input, they would need extra ESD protection !  Once damages by ESD the resistor may no longer be able to withstand a high voltage. So the dividers are really made to be used in a more fixed circuit, like a high voltage power supply or checking on the high voltage battery.

After all the divider is a rel. small and cheap part ($2 range), not really made for highest performance. The plastic case also makes it possibly sensitive to humidity effects. For high performance applications the divider ratio is also a bit large - one would like more 1:100 and I would even like 1:50.

It is still nice to see a new source for precision resistor arrays. The lower resistance versions like RES11 are mor something for high precision, though these are missing doubled pins (4 wire connections) at the ends.