how would i go about getting about 6db of gain out of an opamp?

[ode2no1]

i'm not sure why i'm having trouble figuring this out, but can you guys point me in the right direction? basically i want to go from a sound card to some outboard gear, but i want a volume control with a buffer in between. instead of wiring it as a voltage follower though i'd like to get a bit of gain if possible. how do i determine what value resistors to use? any help would be much appreciated.

[brett]

Hi
6dB is about 4x.
If you are using an opamp in an inverting scheme, make the ratio of the input to feedback resistors about 3 or 4 or 5 to 1. Say 220k input, and 1M feedback.
cheers

[ode2no1]

how do you go about choosing the values though? the audio will be going thru a dual gang pot via XLR connector. do i need inverting or not inverting? sorry, clearly i'm no expert in this area, but i do appreciate any replies.

[petemoore]

  Use a pot wired as variable resistor to be the 'other' resistor, to make the gain adjustable, fiddle with values until the gain range adjustment knob covers the gain range desired. It's the proportion, or ratio of the feedback and input resistors values which control gain, changing either value changes the ratio.
   The opamp data sheet or FAQ should tell more, there's also links which explain how to set/alter gain in the chip circuit.

[WhenBoredomPeaks]

Slightly related but how would you get the maximum possible gain out of an opamp? Using like an 1M and a 1ohm resistor for a gain of 1million? (that sounds like a pretty big number so it is cool) Or they will start to oscillate and stuff after a point?

edit: to be ontopic, i used this to work out opamp gain in a mixer: http://www.radio-electronics.com/info/circuits/opamp_basics/operational-amplifier-gain.php

[nexekho]

Surely the absolute max gain you could get out of an op-amp is comparator mode?

[earthtonesaudio]

The maximum gain of an op-amp is called its "open loop" gain.  Basically you omit any feedback to avoid "closing the feedback loop" and then your gain from input to output pin of the op-amp is exactly the open-loop gain.  The input current goes up somewhat but is still relatively low, as in, don't worry about voltage division until you use an input resistor a lot bigger than 1M.

[merlinb]

Hi
6dB is about 4x.
If you are using an opamp in an inverting scheme, make the ratio of the input to feedback resistors about 3 or 4 or 5 to 1. Say 220k input, and 1M feedback.
cheers

6dB is 2x. So feedback resistor would be two times greater than input resistor for inverting configuration.

[Gurner]

personally, I'd go with non-inverting ......  it seems a bit slapdash to just go and invert the polarity when there's no need 

[ode2no1]

so if i went non-inverting and Av = 1 + R2 / R1...then i just make the two resistors the same value, right? and does it make any difference what the value is?

[PRR]

> just make the two resistors the same value, right?

Correct.

> and does it make any difference what the value is?

Take it to extremes.

Two 1-ohm resistors: the opamp has to drive 2 ohms. It will NOT be happy. Too small.

Two 100-Meg resistors. 1 nanoAmp bias current makes sinificant offset. A dozen pFd stray capacitance swamps 100Meg one way or the other. Too big.

Two 1K resistors is a reasonable low end.

Two 47K resistors is as high as I would go without some thinking.

So what you got too-many-of with red or orange 3rd-stripe? I think I used to use 6K8 because I never seemed to use 6K8 for anything else.

[ode2no1]

is there a super informative page about op amps that goes into detail as to how to choose your values? i want to know why you say 47k is about as high as youd go without some thinking....or why 1k is a reasonable low end. for now though, i super appreciate your help . thank you for giving me a ballpark to work with.

[ode2no1]

also, do op amps need to be powered by DC or can you power them with AC?

[earthtonesaudio]

also, do op amps need to be powered by DC or can you power them with AC?

DC!

[waltk]

is there a super informative page about op amps that goes into detail as to how to choose your values? i want to know why you say 47k is about as high as youd go without some thinking....or why 1k is a reasonable low end. for now though, i super appreciate your help . thank you for giving me a ballpark to work with.

Try these two app notes from TI:

http://www.ti.com/lit/an/sloa030a/sloa030a.pdf
http://www.eng.yale.edu/ee-labs/morse/compo/sloa058.pdf

[sault]

Two 100-Meg resistors. 1 nanoAmp bias current makes sinificant offset. A dozen pFd stray capacitance swamps 100Meg one way or the other. Too big.

PRR, could you please verify if I understand this comment correctly? I just happen to be in the Op Amp section of "the Art of Electronics" (study, study, study in my spare time), and while I completely get the stray capacitance (in and of itself a reason not to use 100M resistors!)...

The TL072 datasheet says 65 pA of offset current, so multiply that by the impedance the inverting input sees ( 100M || 100M = 50M ) multiplied by the gain ( 2 ) means that the offset voltage could be something like 7ish mV of offset voltage, add the 3 mV of offset voltage already listed, and you're looking at something like 10 mV. The effect only gets worse with higher gain (which is pretty much a given for us, right?), the offset voltage could easily approach a volt, and that means less headroom within the op-amp itself.... right? A 100x gain would put us at about a third of a volt, and that's awful close to a full diode drop...

I know that you can minimize this offset by putting a resistor on the non-inverting input, but that only works with the inverting amplifier...


Saul t

[PRR]

> The TL072....

That's an exceptionally low input current chip. Many popular chips have higher inptu current. '741/4558 is significantly higher than my 1nA example.

> ....datasheet says...

Datasheets have strange truthyness.

TL07x is 65nA "Typical". Is the chip in your hand "typical"? Surely not. Datasheet says 200pA max room temperature. 20nA when very-hot (boiling spit).

500pA/0.5nA would be a safe bet for a worst-case TL072 on an outdoor sunny stage.

> worse with higher gain (which is pretty much a given for us, right?)

Well, with higher DC gain. We don't need DC gain for guitar, we need AC gain. We set the gains different for AC and DC with a cap. The simplest case gives unity DC gain.

It would be reasonable to have 0.5nA*50Meg= 0.025V input offset, and with unity DC gain, the same at the output.

Under 9V supply we can't have +/-4.5V swings, probably more like +/-3V swings. If offset error changes that to +3.025V/-2.975V, we have lost 0.07db headroom. Utterly insignificant.

> a resistor on the non-inverting input, but that only works with the inverting amplifier...

No, it can be done with the non-inverting. Select the NFB network on the -in to be equal to the resistor on the +in pin. In guitar work this usually leads to excessively high NFB impedance. But it is common on power amps. The NI input is cap-coupled and biased with 10K. The NFB network is 10K series and 500+50uFd to ground. At DC, both input currents pass through 10K. If input bias current is 100uA, both inputs sit 1V off of ground, but if both input currents are equal these 1V offsets cancel. The input currents are never eXactly equal, but offset current is typically much smaller than bias current.

> "the Art of Electronics"

A fine book for a rushed semester; there's better material around.

OpAmps For Everyone, Mancini, http://www.ti.com/lit/an/slod006b/slod006b.pdf

Op Amp Circuit Collection, NatSemi, http://www.national.com/an/AN/AN-31.pdf

Op Amp Applications Handbook, Jung, http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html

[sault]

Thank you!

I've been interested enough in audio electronics that I've been slowly teaching myself the fundamentals. Hey, gotta keep yourself busy, right? I haven't gotten familiar enough with AC theory yet, I'm still trying to make sure I know the DC basics first. At least I got the basic concept, though, that reassures me a little.

I figured that AoE was a good all-around reference, covering enough of the material in enough detail to get the basics down. I'm trying very hard to pry myself away from the "all hail the Jfet, BJT's are evil, regard op-amps with vague suspicion" mentality I somehow managed to acquire!

I mean, I find myself looking at a Sallen-Key bandpass and thinking to myself "I bet I could do that with just JFETs". You know that saying... when you're holding a hammer, everything looks like a nail?   


Saul t

[PRR]

> Sallen-Key bandpass and thinking to myself "I bet I could do that with just JFETs".

Yes, but....

1) the JFET has significant output impedance; the S-K equations usually assume an amp with zero output impedance.

2) the more-general forms of filters need a precise gain, or a very-high gain. The basic low/high-pass in near-Butter alignment starts with a gain of +1.7(??). JFETs can have gain of 0.95 with Zout near 1K, or -30 very variable with Zout above 10K.

Analysis of the Sallen–Key Architecture  http://www.ti.com/lit/an/sloa024b/sloa024b.pdf

[sault]

Exactly! I want to educate myself enough to draw away from that JFET-preferred mentality, precisely for those reasons. I mean sure, I could probably use a mu-amp configuration to get some higher gains...  Anyways, the component spread with JFETs means that I'd either have to match by hand or use trim-less biasing... I haven't tried figuring out how to do a trimless mu-amp (yet), come to think of it...

So much to learn... how to find a balance between component count and creating a circuit that is tolerant to the variation in device characteristics, for instance. Op amps do seem to be a way to achieve this, however, even though they seem to be a little taboo among stompboxers around these parts...