2

is it possible to get higher output voltage "stability" by cascading voltage references?

-- REPHRASING --

Suppose I have IC1 and IC2. Both are 0.5% accurate TL431 parts. IC1 takes 15V and outputs 12V. This is fed to IC2 and IC2 outputs 5V. Will the 5V output from IC2 be more stable, than if this output was directly taken from IC1?

Output current is well within the limit, and my main concern is line-regulation?

Thanks in advance,

Vishal

Vishal
  • 479
  • 1
  • 5
  • 16

1 Answers1

5

Study the data sheet.

The errors in a voltage reference include ...

a) Initial accuracy (as delivered, at 25C, with nominal load and supply)
b) Temperature coefficient (with variation from 25C, ambient and self heating)
c) Line regulation (mV of output change per volt of input)
d) Load regulation (mV of output change per mA of load change)
e) Long term stability (only specified for the higher quality references, often ppm per month)

Providing a regulated input will remove (c) from the buildup of output errors, all the others will remain. The effect of (c) would have to be much larger than the other terms for it to be worth providing a regulated input.

Neil_UK
  • 166,079
  • 3
  • 185
  • 408
  • 1
    With most voltage references having very good power supply rejection ratios (PSRRs), the effect of (c) can be made very low just by having a linear regulator before the reference, which is fairly common to have in the analog domain's power circuitry already. – user2943160 Aug 07 '16 at 16:39
  • Thanks for the details Neil_UK. I have rephrased the question to convey the exact requirement that concerns me for this design. – Vishal Aug 07 '16 at 18:31
  • Thanks for the details user2943160. I have rephrased the question to be specific about line-regulation. – Vishal Aug 07 '16 at 18:33