How best to power precision opamps

Thread Starter

spitsnsparkin

Joined Jun 19, 2011
29
Hi All,

I have a question about powering opamps.
Suppose an opamp specs the supplies in the range of +/-2V to +/-20V (OP297). Likewise, you are only passing small signals through: 0-1V, 0-5V, 0-10V, etc. Is there a benefit to using the largest supply swing you can provide, for example +/-15V, instead of +/-12V or +/-5V (if the signal can be offset and scaled to fit within the input/output swing of the opamp)?
As the input offsets, noise, etc. are spec’d at absolute values (75uV, etc.), is the signal to noise ratio better at larger supply voltages? Do opamps perform better with larger supplies vs. smaller?
Thanks in advance for the insight.
Spitsnsparkin
 

crutschow

Joined Mar 14, 2008
31,123
Generally the power supply DC voltage has little effect on the S/N of the output signal.
More important is the absolute value of noise on the supply, as you generally want to keep that as small as practical.
A larger supply voltage is mainly of use to increase the maximum swing of the output voltage.
 

Thread Starter

spitsnsparkin

Joined Jun 19, 2011
29
I could guess that using the largest supply voltage may make the inputs more tolerant to an expected, out of bounds inputs signal where the closed-loop feedback cannot compensate to keep the input voltages at the plus and minus inputs within the input voltage range, which is usually linked to the supply.
 

Papabravo

Joined Feb 24, 2006
19,578
I could guess that using the largest supply voltage may make the inputs more tolerant to an expected, out of bounds inputs signal where the closed-loop feedback cannot compensate to keep the input voltages at the plus and minus inputs within the input voltage range, which is usually linked to the supply.
You should consult the Absolute Maximum Ratings section of your favorite datasheet and NOTE WELL any restrictions on common mode input voltage ranges. You might be surprised at the results of that survey.
 

MrSalts

Joined Apr 2, 2020
2,617
Your chip has 120db power supply rejection ratio. What kind of ripple do you have in your power supply that concerns you?
That's taking 1volt of ripple in a terrible power supply and seeing 1uV of ripple in the output.
 

Thread Starter

spitsnsparkin

Joined Jun 19, 2011
29
Hi crutschow,
Thanks for the reply. Does the PSRR become better with larger supplies? 100mV ripple on +/-5V would seem ratiometrically smaller on +/-10, 12 or 15. Right?
 

Thread Starter

spitsnsparkin

Joined Jun 19, 2011
29
Thanks to all,
My original post concerned the selection of the power supply for an opamp.
If you can process the signal using +/-5V supplies, would there be a benefit to using rails near the maximum (+/-12V or 15V) provided that the magnitude of the input and output signals were the same. Does some aspect of opamp circuts improve with larger supplies, including crutshow’s comment of larger output swing capability?
 

crutschow

Joined Mar 14, 2008
31,123
Would larger than necessary supplies simply cause more self-heating and be a net negative in opamp health and or performance?
It would cause more op amp power dissipation, but most op amps have such low dissipation that it is unlikely to have a significant effect on either their health, life, or performance.
 

DickCappels

Joined Aug 21, 2008
9,512
Probably the best practice is to just use enough voltage to assure that everything works correctly. I have memories of watching a TL431 drift as it warmed up. 50 ppm/°C is a lot if there is significant heating. Haven't seen that on precision op amps... yet.
 
Top