opamp supply voltage question

[j.y]

hi guys..

i was reading up and lookin some schematics and i notice some opamp circuits use +ve and -ve supplies.. and some uses +9v and gnd, with a 4.5v ref like the tube screamers..

im curious about the difference in powering the circuit using these 2 ways..

can somebody explain it to me??

thanks,
jy

[Mark Hammer]

Alternating AC signals go positive and negative relative to some midpoint.  "Bipolar" (+/-) supplies have a true ground or midpoint that represents zero volts.  Circuits that use +9v and a +4.5 midpoint simulate having a plus, minus and ground.  The 4.5v point substitutes for true ground in the bipolar supply, and zero volts fulfills the same function as the  negative supply.  Keep in mind that if you had an audio signal that ranged as wide as +100mv/-100mv peak to peak, the only way it can go negative is if there is a negative direction to go in.  The zone between 4.5v and 0v serves as the "negative zone".  The reason it is 4.5v and not some other value is because the exact middle of the battery's range permits the most positive AND negative swing, simultaneously, that you can get with 9v.  You will sometimes see circuits where the biasing is NOT at 4.5v.  Such "asymmetrical biasing" is intended to provide more clean headroom for one side/half-cycle of the signal than for the other.  As you can expect, when you are aiming for "clean", 4.5v biasing is needed.  If some sort of soft-clipping is the intention, then biasing which limits voltage swing on one side of the Vref is used.

While the power supply may be more complex in bipolar contexts, the audio circuit will be LESS complex and often quieter because of the numerous components that can safely be left off.  If a synthesised Vref of 4.5v is used then it has to be added to the input of each amplification stage.  Since it is possible that some drift or error occurs in that gain stage, best practice is to "reset" the signal by using a DC-blocking capacitor on the output and then rebiasing all over again.  All those extra caps and resistors here and there add noise to the equation.

[gez]

Just to add to Mark's comments.  The outputs of many op-amps don't swing rail-to-rail and often can be a few volts short (see data sheets).  Doesn't give you much headroom with a single 9V supply.  Often that isn't a problem with 'guitar only' circuits, but if an effect is intended to be used with other instruments too (keyboards and bass can kickout signals of much greater amplitude than the average guitar signal), or if a stage needs significant gain but you don't want the signal to be clipped, then split supplies are a good way to go as more headroom is available.

[j.y]

ic...

thats its more of a headroom issue? if so.. its still possible to fix split supply for the opamp?? like lets say the tubescreamer?

now that we are at opamps.. theres another question i would like to clarify.. is there a difference in audio applications that the opamp be confgured in inverting or non inverting? i always notice in schematics that if its a 2 opamp circuit.. both opamp will be in inverting config so that the output will be same polarity as input.. if so.. why dont designers use non inverting config for all audio applications??

thanks.

[gez]

ic...thats its more of a headroom issue? if so.. its still possible to fix split supply for the opamp?? like lets say the tubescreamer?

Not sure what you mean by 'fix' the supply.  You can re-jig a circuit for split supply, but whether that's worth doing is another matter.  In the case of the tubescreamer I'd say no as the diodes limit the amplitude of the signal, so headroom isn't really an issue. 

Although headroom is primarily the reason many designers opt for split supply, it's not the only reason - some building blocks are easier to do/function better with that arrangement.

is there a difference in audio applications that the opamp be configured in inverting or non inverting? i always notice in schematics that if its a 2 opamp circuit.. both opamp will be in inverting config so that the output will be same polarity as input.. if so.. why dont designers use non inverting config for all audio applications??

Different horses for different courses.  There are some things the inverting configuration is better at doing - mixing, integration, less than unity gain, some filters etc - but one of the drawbacks is that input impedance tends to be low if you want high gain (first transistor tends to be low value).  The main virtue of the non-inverting mode is that input impedance can be a lot higher - especially if you bootstrap - so you often see this configuration preceding an inverting stage.