Post Go back to editing

LT5400 Resistor Network Matching

Category: Hardware
Product Number: LT5400

Hey all,

I've got a question concerning the matching properties of the LT5400. I want to use the LT5400-3 and need a very precise divider of 6 for one of our ATE units (dividing a up to 60V analog voltage to 10V for a multiplexer). Therefore I connected both 100k resistors in parallel and one of the 10k resistors in series to GND. The datasheet states a matching error of 0.01% (grade A). Does this mean 0.01% between each 100k and 100k / 10k and 10k resistor or is this also meant to be applicable for the ratio between 100k and 10k (10:1)? Because another datasheet parameter is the absolute resistor tolerance which states up to 7.5%. This confuses me a bit and I'm uncertain if this is the correct resistor network for my application of a precise 6:1 divider.

Thanks in advance!

Greetings, BuckiED

  • Hi BuckiED:  The exact values of the resistors part to part can be as much as 7.5% away from the nominal value, but within a specific LT5400-3 the (same value) resistors are matched to each other within 0.01%, and also the ratio.  The exact equation applied in production test is given in Note 6 of the data sheet.

  • Hi Glen,

    thanks for your quick answer! So if I understand your reply correctly I don't have to care for the absolute resistor values, because the ratio is still the same every time (in the error margin of 0.01%). Even if the absolute value of R2 and R3 are 10k + 7.5% this means the absolute value of R1 and R4 is 100k with also + 7.5% (+-0.01% in respect to the ratio)? All in all this would always lead to my desired 6:1 voltage divider like I configured above, correct?

    Greetings, BuckiED