1
\$\begingroup\$

Let's say I had a low power op amp (pre-amp stage) going to an ADC (LTC2412). enter image description here

My question is: If I wanted a super precise, low noise, low TC drift, positive supply line could I just power the +A5V line from a voltage reference IC like the LTC6655LN-5? My circuit only needs about 3mA ( 6mA projected worse?)to run.

I have added the schematic to show that the +2.5V line is produced from the +A5V line. So therefore changes in supply to the ADC will also affect the +2.5V common mode voltage ( going to the CH- channel on the ADC), allowing for more linearity. So I am looking for a very low noise +5V source, with preferably low TC drift. I am unsure of the tradeoffs of using a voltage reference as the supply or using an LDO or using a reference+LDO combo (like the LTC6655LN-5 and LT3045) Like this: enter image description here

Thanks

\$\endgroup\$
12
  • 6
    \$\begingroup\$ This isn't really answerable without some numbers, so datasheets will be useful. It looks like you're trying to use the load regulation and line regulation of the voltage reference to work around a poor power supply rejection ratio of the op amp, but the actual numbers involved aren't clear. An op amp with a better PSRR, running off the same (or better) LDO, is probably the way to go. \$\endgroup\$
    – nanofarad
    Jul 20, 2021 at 23:26
  • 3
    \$\begingroup\$ Plus, a precise supply voltage doesn't matter to the opamp (as long as it's sufficiently high) - a low noise one might matter, but exactly as nanofarad says: numbers are important, hand-waving isn't. You will need to tell us which problem you're solving, to which degree (SNR) you need to solve it and how far you are from solving it so far. Then we can talk about potential improvements, everything else is just conversion of air to hot air. \$\endgroup\$
    – mmmm
    Jul 20, 2021 at 23:50
  • 3
    \$\begingroup\$ Please edit your question with the additional information about your op-amp and ADC. Stackexchange likes the answer itself to be standalone, so you can just read question & answer and not have to plow through the comments. \$\endgroup\$
    – TimWescott
    Jul 21, 2021 at 1:10
  • 1
    \$\begingroup\$ Also please give us a schematic of your proposed circuit -- "ratiometric" and "op-amp" do not go together -- you are either misunderstanding what an op-amp does, or you're doing something strange with it. \$\endgroup\$
    – TimWescott
    Jul 21, 2021 at 1:11
  • 2
    \$\begingroup\$ I think people were a bit harsh on you. If you already have the reference for precision, you can also use it to power an op-amp that is part of that precision circuit. However, beware of capacitive loading on the reference. The op-amp may require a certain amount of capacitance on VCC. The reference may not like a lot of capacitance. So that could be an issue. And of course, you have to watch out for and respect maximum output currents and such. \$\endgroup\$
    – user57037
    Jul 21, 2021 at 2:39

2 Answers 2

6
\$\begingroup\$

This Vref IC up to $21/pc is overkill for most Op Amp applications. It is limited to +/-5mA so a 6mA load exceeds spec.

Unless a specific need is defined as designed as a voltage reference and not a power source, this is a poor match to power Op Amps.

The LT3045 is a suitable buffer. But choosing an ultra-stable supply does not make a design great alone. In fact, it draws doubt on the design without specs.

\$\endgroup\$
2
  • \$\begingroup\$ I thought the IC can output a Short-Circuit Current up to 20mA to ground ? \$\endgroup\$ Jul 21, 2021 at 2:38
  • 3
    \$\begingroup\$ what good is a 0V short cct as a precision power source?? \$\endgroup\$ Jul 21, 2021 at 3:51
2
\$\begingroup\$

The whole idea of a precision reference is that it is a reference. You plop it into your circuit, and then you handle its output with kid gloves. If you look at the datasheet for the LTC6655, you'll see that it does have a non-zero sensitivity to load current.

Moreover, if you look at the graph labeled "2.5V Output Impedance vs Frequency", you'll see that at around 10kHz the output impedance goes up to \$1\Omega\$. Couple that with whatever varying current demands your op-amp and ADC put on the thing, and it may not be that precise of a precision regulator.

I suggest that you don't try to save a dime on a run-of-the-mill LDO, but rather you make a 5V line with a real regulator, then make a reference voltage for the op-amp from a real reference. Deriving your 2.5V from that reference is probably a good idea -- but I wouldn't power any devices from it.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.