Replace OP amps. Completely pointless, or not?

[sq225917]

As I wrote earlier, any unity gain stable opamp will suffice in such a circuit. But that's only 1 circuit.

Take the nuerochrome amps, loads of feedback, bucket loads in fact, but can you use another opamp in that circuit? Nope, the compensation is so tightly defined almost any other part would be a shit show.

My point stands, saying opamp rolling doesn't make an audible difference is a million miles from saying rolling similar opamps in a specific circuit type doesn't make an audible difference.

I'm not being pedantic, I'm being rigorous, because the statement of conditions ensures that we 'objectivists' don't get misunderstood and written off as wrong for making obviously Incorrect generalisations.

 

[Killingbeans]

Really, shouldn't you guys be helping one and other?
"I heard this change..."
"Okay lets find out what caused it..."

Believe me, we've done that.

Most of the time it goes:
"I heard this change..."
"Okay lets find out what caused it...
The measurements show nothing. There's no difference whatsoever. Most likely placebo. Do a blind test if you really want to know."
"Don't tell me what I'm hearing!!! It MUST be something science hasn't discovered yet! You need to be open minded!"
And around we go...

 

[MakeMineVinyl]

The sweeping generalisations obscure the truth. You have to consider the circuit. Two opamps of similar specs "should" sound the same in the same circuit. But only if you've accounted for all the parameters that contribute to audibility. Which is no easy task.

Please list those, and please do some blind testing on those parameters, one at a time, to determine if what you're hearing is real or an illusion.

 

[mhardy6647]

... "Put down in layers like 3d printing" - what you mean, constructed using the same xray lithography techniques as the chips in Tesla cars, iphones and super computers...

you forgot "... and military weapons systems that cost billions of dollars..."

 

[sq225917]

Please list those, and please do some blind testing on those parameters, one at a time, to determine if what you're hearing is real or an illusion.

Please, read the rest of my posts in the thread.

 

[MakeMineVinyl]

Please, read the rest of my posts in the thread.

Sorry, I'm not about to read through over 200 posts in order to research what you have to say. Please shortly list the 'parameters' and some hard evidence with blind testing. If you have a valid argument for your views, please take this opportunity to make it.

 

[mansr]

An obvious example is output voltage range. If the replacement is more limited, it might clip, something that can easily be audible, even if the same two opamps are indistinguishable when operated within the narrower range.

 

[MakeMineVinyl]

An obvious example is output voltage range. If the replacement is more limited, it might clip, something that can easily be audible, even if the same two opamps are indistinguishable when operated within the narrower range.

How so? Just about any op amp can swing within a couple volts of the +/- rails.

 

[mansr]

How so? Just about any op amp can swing within a couple volts of the +/- rails.

Some can get a lot closer to the rails than others. If the design has several volts of margin, it'll probably be fine. That isn't always the case, however.

 

[MakeMineVinyl]

Some can get a lot closer to the rails than others. If the design has several volts of margin, it'll probably be fine. That isn't always the case, however.

Even the ancient 741 and 301 can get within a couple volts of the rails; I have yet to see any op amp which cannot do this, especially current designs. If a designer chooses to use lower voltage rails, well that's a design issue and choice of op amp is not going to cure a compromised design from the start.

Changing op amps is a solution in search of a problem. Personally, I've had better sonic results by changing the on/off power switch on my gear.

 

[Deleted member 46664]

Believe me, we've done that.

Most of the time it goes:
"I heard this change..."
"Okay lets find out what caused it...

Then, the part that should not happen...

The measurements show nothing. There's no difference whatsoever. Most likely placebo. Do a blind test if you really want to know."
"Don't tell me what I'm hearing!!! It MUST be something science hasn't discovered yet! You need to be open minded!"
And around we go...

There are two acceptable outcomes ...
"Okay ... here's the difference...."
--OR--
"I can't find any difference..."

The rest is up to the other person to figure out on their own.

 

[mansr]

Even the ancient 741 and 301 can get within a couple volts of the rails; I have yet to see any op amp which cannot do this, especially current designs. If a designer chooses to use lower voltage rails, well that's a design issue and choice of op amp is not going to cure a compromised design from the start.

The NE5532 is only guaranteed to reach 3 V from the rails (24 Vpp at ±15 V) whereas the OPA1612 manages 0.6 V. That's quite a difference, and a circuit designed for the latter could easily hit the limits of the former. It's not "compromised" to utilise the full performance of the chosen parts.

 

[Deleted member 46664]

The NE5532 is only guaranteed to reach 3 V from the rails (24 Vpp at ±15 V) whereas the OPA1612 manages 0.6 V. That's quite a difference, and a circuit designed for the latter could easily hit the limits of the former. It's not "compromised" to utilise the full performance of the chosen parts.

And what are the odds, in an audio circuit, of either of those chips having to produce 24volt peak to peak signals?

Virtually all line level devices operate at about 1 volt p/p ....

 

[mansr]

And what are the odds, in an audio circuit, of either of those chips having to produce 24volt peak to peak signals?

Virtually all line level devices operate at about 1 volt p/p ....

Nobody says you have to use ±15 V rails. Moreover, opamps are frequently used in many places besides line level output drivers.

 

[sq225917]

I've already stated them, with measurements, if you can't be arsed to read them that's your call. I'm not Google.

 

[MakeMineVinyl]

And what are the odds, in an audio circuit, of either of those chips having to produce 24volt peak to peak signals?

Virtually all line level devices operate at about 1 volt p/p ....

Yes, that's the point - in normal usage and with a competent design, no op amp is likely to be stressed for output voltage. Certainly op amp swapping is not likely to make any practical difference; if the design is so close to the edge so that a 1 volt difference in ability to reach the rails makes a difference, that design is broken!

 

[antcollinet]

Even the ancient 741 and 301 can get within a couple volts of the rails; I have yet to see any op amp which cannot do this, especially current designs. If a designer chooses to use lower voltage rails, well that's a design issue and choice of op amp is not going to cure a compromised design from the start.

Changing op amps is a solution in search of a problem. Personally, I've had better sonic results by changing the on/off power switch on my gear.

Some op amps can go rail to rail. If you replace one of these with one that cannot, there is risk that you break the design.

 

[MakeMineVinyl]

Some op amps can go rail to rail. If you replace one of these with one that cannot, there is risk that you break the design.

So why would any competent designer make a design so critical that a change breaks the desvice? That is a poor design, full stop. I've been designing with op amps since the early 1970s; I would never in a million years consider designing in such a narrow margin of proper operation.

 

[antcollinet]

Why do op-amps with rail to rail capability exist if that capability is not going to be used.

Every design decision is a tradeoff. You shouldn't have to design expecting idiots who don't know what they are doing to change devices for something else less capable.

On the other hand - if you put sockets into the design, and advertise to people that they can roll their own - you'd better damn well make sure your design is tolerant to typical variations between devices.

 

[egellings]

Most audio applications employing op amps never ask for more than maybe 2Vrms from an op amp, so that requirement is easy to meet with an op amp powered on +/-15V rails.