opamp inverting input sounds better?
- Thread starter dorkus
- Start date
I am not sure it is an issue of non-infinite common mode rejection though this could be a factor.. ?
The problem you are referring to generally occurs with FET input op-amps. What you have to consider is the actual circuit implementation.
When you use an inverting configuration for an op-amp, both the inverting and non-inverting terminals are generally kept pegged a a fixed voltage level. Hence that voltage level is kept constant w.r.t. the rails. Between the input terminals and the body of the semiconductor, there is a non-linear capacitance.
With an inverting configuration, the voltage at the input terminals changes w.r.t. the body of the semi. The effect of this is a non-linear input resistance. That non-linear input resistance causes distortion and potentially bad sound quality.
One way to fix this is to use a low impedance source to drive the non-inverting configuration. That often defeats the whole purpose of using the non-inverting configuration. The other way of reducing the distortion to to vary the voltage on the supply rail.
I have attached a schematic of a circuit I whipped together from the parts bin... no smart *** comments of the 2N2222, etc. please...
! Notice how the output is feedback into the supply rail. This reduces the distortion by a good order of magnitude. [and this circuit actually does not sound to bad. The pin used on the AD744 is before the ClassAB output stage.]
I can't claim the idea for doing this... I am sure if you did a search on the web you would find it somewhere!
The problem you are referring to generally occurs with FET input op-amps. What you have to consider is the actual circuit implementation.
When you use an inverting configuration for an op-amp, both the inverting and non-inverting terminals are generally kept pegged a a fixed voltage level. Hence that voltage level is kept constant w.r.t. the rails. Between the input terminals and the body of the semiconductor, there is a non-linear capacitance.
With an inverting configuration, the voltage at the input terminals changes w.r.t. the body of the semi. The effect of this is a non-linear input resistance. That non-linear input resistance causes distortion and potentially bad sound quality.
One way to fix this is to use a low impedance source to drive the non-inverting configuration. That often defeats the whole purpose of using the non-inverting configuration. The other way of reducing the distortion to to vary the voltage on the supply rail.
I have attached a schematic of a circuit I whipped together from the parts bin... no smart *** comments of the 2N2222, etc. please...
! Notice how the output is feedback into the supply rail. This reduces the distortion by a good order of magnitude. [and this circuit actually does not sound to bad. The pin used on the AD744 is before the ClassAB output stage.]I can't claim the idea for doing this... I am sure if you did a search on the web you would find it somewhere!
Attachments
hmm, interesting. i've known about capacitance-modulation-related distortion, but did not think about how it relates to which opamp input is used. (we're assumming IC devices here, thus the common substrate.) i guess my real question is which configuration for a given circuit yields the lowest distortion, or at least exhibits distortion that is least significant to the system. in my case, the application is a simple line-level preamp, say with 20dB gain and an input impedance of at least 20kohms. which configuration would you pick? (let's forget about absolute polarity for now...)
p.s. the feedback circuit looks very clever, but i wonder if it's fixing a small distortion while introducing new ones from the additional feedback? the need for additional circuitry probably means imposing additional sonic signature on top of the basic circuit...
p.s. the feedback circuit looks very clever, but i wonder if it's fixing a small distortion while introducing new ones from the additional feedback? the need for additional circuitry probably means imposing additional sonic signature on top of the basic circuit...
That was what I was thinking the whole time I was making the circuit. However, with a large input resistance, 20K would not be large or small, distortions can be higher than 0.1%.
Here is the link to the application note on the ADI web-site that covers this topic. Racked the brain and remember where I had seen this written up and where the circuit idea came from. I basically modified the circuit shown in the attached application note to use the Class-A output of the AD744 and added a class-A discreet buffer. The AD811 works well for the feedback, but I have an improved version, can't find it right now that uses an LM7171 instead. As opposed to a simple feeback circuit, I also treated the feedback loop as a "control system" and tuned it for minimum overshoot and ringing. This improved the sound quality yet a bit more.
I mainly did this because I was playing around with something where I wanted a high input resistance and low distortion.
Alvaius
http://www.analog.com/library/applicationNotes/amplifiersLinear/precision/AN232.pdf
Here is the link to the application note on the ADI web-site that covers this topic. Racked the brain and remember where I had seen this written up and where the circuit idea came from. I basically modified the circuit shown in the attached application note to use the Class-A output of the AD744 and added a class-A discreet buffer. The AD811 works well for the feedback, but I have an improved version, can't find it right now that uses an LM7171 instead. As opposed to a simple feeback circuit, I also treated the feedback loop as a "control system" and tuned it for minimum overshoot and ringing. This improved the sound quality yet a bit more.
I mainly did this because I was playing around with something where I wanted a high input resistance and low distortion.
Alvaius
http://www.analog.com/library/applicationNotes/amplifiersLinear/precision/AN232.pdf
dorkus, almost every opamp has 10 times higher distortion in non-inverting mode. An example is here but it goes for everyone (more or less).
http://www.linear.com/prod/datasheet.html?datasheet=202
Check the diagrams of distortion.
http://www.linear.com/prod/datasheet.html?datasheet=202
Check the diagrams of distortion.
Hi All
The facts mentioned above ore the ones I already read about.
I could imagine another one as well: As soon as you use the inverted circuit configuration there is no component , apart from two resistors, that is not within the forward part of the feedback loop (which isn't the case for non inverting circuit configuration).
And the two resistors, even when of the worst type available, are still closer to the ideal from the linearity point of view than any other component ever will be.
I made some simulations on a fully symmetrical amp which is inverting and that doesn't use a differential stage at the input. The results looked promising so I will build a prototype a soon as I find time to.
Regards
Charles
P.S.:
My subtractive Manger crossover uses inverted stages only. The only exception is the differential line receiver (INA 134) on the input.
The facts mentioned above ore the ones I already read about.
I could imagine another one as well: As soon as you use the inverted circuit configuration there is no component , apart from two resistors, that is not within the forward part of the feedback loop (which isn't the case for non inverting circuit configuration).
And the two resistors, even when of the worst type available, are still closer to the ideal from the linearity point of view than any other component ever will be.
I made some simulations on a fully symmetrical amp which is inverting and that doesn't use a differential stage at the input. The results looked promising so I will build a prototype a soon as I find time to.
Regards
Charles
P.S.:
My subtractive Manger crossover uses inverted stages only. The only exception is the differential line receiver (INA 134) on the input.
Another thing you can expect to work better in inverting opamp's is that if you use a feedback cap (Cf across Rf) to stabilize and to raise the phaseshift, it will able to give you a gain of less than 1 at higher frequency which is needed to stabilize an opamp when driving dificult loads.
This will be very hard to optain in a noninverting design, as this cap will turn your opamp into a follower (a little bit less than 1 times gain) at higher frequency.
Am i wrong PerAnders?
Sonny
This will be very hard to optain in a noninverting design, as this cap will turn your opamp into a follower (a little bit less than 1 times gain) at higher frequency.
Am i wrong PerAnders?
Sonny
peranders said:
i dunno. i think it's a little more difficult in some applications (due to less flexibility with the input impedance). also, a lot of people can't keep track of absolute polarity and wouldn't know if they need to reverse their speaker cables or not. boo hoo.
anyway, i'm thinking building a little IC-based amplifier, a 47 Laboratories Gaincard clone if you will. i currently have the Blue Circle Music Pump mono amps (basically Gaincard clones in a woman's pump shoe) which i'm reviewing for a magazine, and i'm shocked at how good they sound. no not the ultimate sound, but really quite good and very musical. i'm pretty sure they use the LM1875 chip, which is so simple to implement, so i would like to build a really compact amp to use as a mini-system or for computer speakers. i am just wondering if i should try something new, and maybe run the IC in inverting configuration... wonder if that would work with what is essentially a power op-amp? any possible side effects to this in a higher-current application like this one?
cheers,
dorkus
Noise gain
One disadvantage of the inverting config is that the noise gain is always higher than the signal gain. Sounds cryptic?
Consider that for a gain of -1 (both resistors equal) the input signal really is attenuated by the input resistor and the feedback resistor (halved in this way), so even if it looks like a gain of 1, the opamp really internally has to amplify the signal by 2. That means that all noise and distortion generated in the opamp is also amplified by 2 rather than one (hence the name 'noise gain').
If you get more fancy with feedback topologies the signal gain and the noise gain can be quite different and it will look like the feedback factor is less effective than expected.
Jan Didden
One disadvantage of the inverting config is that the noise gain is always higher than the signal gain. Sounds cryptic?
Consider that for a gain of -1 (both resistors equal) the input signal really is attenuated by the input resistor and the feedback resistor (halved in this way), so even if it looks like a gain of 1, the opamp really internally has to amplify the signal by 2. That means that all noise and distortion generated in the opamp is also amplified by 2 rather than one (hence the name 'noise gain').
If you get more fancy with feedback topologies the signal gain and the noise gain can be quite different and it will look like the feedback factor is less effective than expected.
Jan Didden
Hi folks, I won't get into the deep technical critique, cos I would have no idea what I was talking about , but here is a forum you might find interesting-
http://pub4.ezboard.com/ffakeidsfrm1
I am currently building inverting and non inverting versions using the LM3875, using the info from this forum, and hey, I think I might be learning something!
, but here is a forum you might find interesting-http://pub4.ezboard.com/ffakeidsfrm1
I am currently building inverting and non inverting versions using the LM3875, using the info from this forum, and hey, I think I might be learning something!
This was one of the later schematics
I have to update this circuit, but this was one of the later configurations with some better transistors on the output stage and a better tuned feedback loop. The comment above of loading the 3906 is quite valid depending on what op-amp you are using. I actually loaded the AB output of the op-amp into the negative rail which improved the quality a bit.
I have to update this circuit, but this was one of the later configurations with some better transistors on the output stage and a better tuned feedback loop. The comment above of loading the 3906 is quite valid depending on what op-amp you are using. I actually loaded the AB output of the op-amp into the negative rail which improved the quality a bit.
Attachments
inverting input and input impedance
hmm ok, so even if i need a relatively high input impedance, is inverting-input configuration still recommended? e.g. i want to have 50k input impedance with 26dB gain, this would require a 1Mohm feedback resistor - very high, would noise and distortion performance suffer? would i be better off using non-inverting to get this high an input impedance?
hmm ok, so even if i need a relatively high input impedance, is inverting-input configuration still recommended? e.g. i want to have 50k input impedance with 26dB gain, this would require a 1Mohm feedback resistor - very high, would noise and distortion performance suffer? would i be better off using non-inverting to get this high an input impedance?
Every time you mention inverting phase or phase
splitting to the semi technical audiophile, they go into
shock.
I have another hypothesis which occurs when the stage
is not being used for much gain:
It is my experience that at low and unity gains, a lot of
op amps are on the edge of oscillation. Over the years
I have made a number of bad sounding circuits (no, not
ones you've seen) and many of them were bad sounding
not because they were oscillating, but because they were
near the edge. If I opened up the feedback gain, they
got better as they moved farther from the edge.
As a result, whenever I use op amps in an audio circuit, which
is not often, I make sure that I either use up or throw away
about 20 dB of open loop gain. If the circuit has a gain of
10 or greater, this is automatically achieved.
If the circuit has less than 20 dB gain, I toss some away.
With an inverting amp, a resistor from the - input to ground
will do it. If not, then resistance from both + and - to
ground is used.
Works like glue....
splitting to the semi technical audiophile, they go into
shock.
I have another hypothesis which occurs when the stage
is not being used for much gain:
It is my experience that at low and unity gains, a lot of
op amps are on the edge of oscillation. Over the years
I have made a number of bad sounding circuits (no, not
ones you've seen) and many of them were bad sounding
not because they were oscillating, but because they were
near the edge. If I opened up the feedback gain, they
got better as they moved farther from the edge.
As a result, whenever I use op amps in an audio circuit, which
is not often, I make sure that I either use up or throw away
about 20 dB of open loop gain. If the circuit has a gain of
10 or greater, this is automatically achieved.
If the circuit has less than 20 dB gain, I toss some away.
With an inverting amp, a resistor from the - input to ground
will do it. If not, then resistance from both + and - to
ground is used.
Works like glue....