op-amp what voltage to use?

[steven2583]

How do you determine the optimal voltage to use an op-amp at? I had my sound processor op-amps replaced with OPA627s and PGA2310. How do I determine the best voltage to run them at? The op-amp takes a range from +/-4.5 v to +/-18v and the PGA2310 Analog voltage goes up to 0 to +/-16. I've read the higher the voltage the better the dynamics. Is there a downside of raising the voltage? Is the best way is just try a few values and see how they sound? Does anyone know the characteristics of the sound as the voltage goes up?

[indianajo]

I was replacing much more pedestrian 4558 op amps with 33078's, but since nobody else answered- - The used mixer was running the 4558's at +-5V with resistors off the +-15 supply, which decreased the hum generated by the AC power switch right next to the op amps. Jumpering the resistors increasd the hum. I put the resistors back for a while. When I went from 4558 to 33078 the annoying -40db?? hiss went away, but the hum was still there. I reduced the hum by moving the transformer out of the steel box, isolating RCA jack grounds, and moving the root of the star ground system to the safety ground, but there was still some hum. I increased the rail voltage to +-8 to allow more headroom for an FM radio input that hissed itself it I turned the radio pot down too much. If the radio pot was up, the op amps clipped the radio @ +-3v power (33078's draw more current than 4558, the voltage dropper was a resistor). The byproduct of increasing the PS rails, was that I was able to turn down the gain on the power amp,and turn up the sliders (gain) on the mixer inputs, which decreased the system hum of course. Didn't notice a lot of difference in the sound on magnetic phono cartridge (50x gain max) between the two P.S. voltages. This was listening on SP2 speakers that by spec are 30-40 db down on second harmonic distortion, which is the floor of my listening accuracy. My hearing is down 6 db at 14khz, your results may differ. (Thank you USAR for the howitzer experience.)

[AndrewT]

Increasing the supply voltage allows an increased maximum output voltage. Both the continuous maximum sinewave voltage and the peak transient voltage capability increase with supply voltage.

This greater output capability results in a higher max signal to noise ratio. The dynamic range is effectively increased due to the higher supply voltage.

This dynamic range is quite different from dynamics in a musical reproduction.

Dynamics is a description of how well the reproduced audible sound follows the original sound from very quiet to very loud.
Changing the opamp supply voltage will not change the way the "system" will reproduce the dynamics of the performance.
It simply allows one to pass a higher signal through before clipping distortion become audible.

[jackinnj]

Run the opamps with as high a voltage as can be tolerated -- further to Andrew's point, the closer the output signal gets to one of the rails the more likely the signal will be distorted. I have a friend who mods high-end equipment for a living and he hates the stuff with low rail voltage, i.e. 3.3V and 5V rails.

Higher voltage makes the opamp run hotter -- someone more knowledgeable than me can explain it with more eloquence -- but the opamp seems to stress less (thermal distortion) with wider excursions of input/ouptut.

[qusp]

many opamps will also swing closer to the rails into a low impedance with a higher supply voltage, rather than just being less in comparison

[indianajo]

All 4 of my power amps will generate full power with 1.6 VAC input.
The lowest impedance input of any of the 4 is probably about 150k on the St120. So neither of the above reasons matters much if driving normal power amps. If driving a headphone, use something better than an op amp. That is why my disco mixer had +-15V supplies, to drive the headphone amps, not for the power amp drivers which were +-5 v rails with original 4558's. Clipping distortion is what I had when I tried to run a 5 VAC radio signal into +-3V rail op amps.

[AndrewT]

Using Indianajo's Power Amplifier sensitivity of 1.6Vac to produce maximum power.
I would suggest that the preceding stage can pass that same 1.6Vac+10dB (~5Vac) without excessive distortion.
I would prefer to recommend +20dB but that takes one beyond what normal opamps can provide. The end result is to run at +-15Vdc or +-18Vdc or +-21Vdc depending on how well the opamp/s cope with the high supply voltages.
For 2Vac sensitivity Power Amplifiers I would certainly recommend at least +-18Vdc for the pre-amp. That is also what I would use for a CD Player which normally has a maximum output in the range 2Vac to 2.3Vac.

[Conrad Hoffman]

I'm not sure there's any rule of thumb as it depends on the opamp and situation. If you have plenty of headroom with the output voltage, there's no reason to run the opamp near the absolute maximum voltage. Also, power dissipation will go up, more or less, with the square of the voltage, for the resistive part of the load. Current sources will suck whatever they need. You can cut a few volts and run the chip cooler. In theory there should be less noise. OTOH, performance probably suffers if you get near the minimums. I'm quite happy running +/-15 to +/-22 volt parts at +/-12 volts. If the difference between 12 and 15 matters, the design needs more margin!

[Mooly]

My old Micromega Stage CD player runs at -/+9 volts and the original opamps were NE5534