Replace OP amps. Completely pointless, or not?

[Sokel]

Any change (not just OAs) has to have a goal.
So,what is one try to achieve changing an OA?

Is it deeper lows?Is it lower audible distortion,IMD,etc?
Is it something mysterious else?
Except for the "something else" answer non of the above will happen and no goal will be reached.
It can change the sound dramatically though in some cases,that is if it breaks the whole circuit and all hell breaks loose.

Most importantly,what is it that bothers someone to reach to a point that must change it?

 

[cgallery]

What we are saying is that there may be audible differences if you make the test hard enough; lots of gain, large signals, multiple OAs, large loads, but not if you use OAs they way they are used in home audio systems. This is only applicable to music production, as in mic prees and other analog devices which are getting rare.

Oh I know what you guys are saying.

And I do agree that the distinction between what the op amps in the Aiyima 07, and a phono or perhaps tape head stage are doing.

OTOH, I also note that the distortion levels in that test are still very low, and I honestly don't think most here would have believed it would be so easy to detect differences until they saw it done.

 

[cgallery]

Any change (not just OAs) has to have a goal.
So,what is one try to achieve changing an OA?

Is it deeper lows?Is it lower audible distortion,IMD,etc?
Is it something mysterious else?
Except for the "something else" answer non of the above will happen and no goal will be reached.
It can change the sound dramatically though in some cases,that is if it breaks the whole circuit and all hell breaks loose.

Most importantly,what is it that bothers someone to reach to a point that must change it?

Right, in my case it was unnatural sounding cymbals. Old recordings that admittedly aren't the best (but where the performances were great) had become nearly unlistenable for me.

To the point made by @Cbdb2 above, for all I know, distortion increased, and I couldn't care less because those recordings are once again available to me.

Keep in mind, BTW, that it took months and significant trial and error before swapping the op amps. I was trying all sorts of EQ (I'm biamped and use MiniDSP for my crossover). I was trying room treatments. I was looking for alternative masterings and pressings, trying digital vs. analog, etc. Not a single change ever resulted in a welcome change until I swapped those op amps.

 

[solderdude]

OTOH, I also note that the distortion levels in that test are still very low, and I honestly don't think most here would have believed it would be so easy to detect differences until they saw it done.

When you look at the results II assume you meant the AES paper) you might have noticed that the ones that were clearly audible all had audible level (around -60dB) amounts of higher order distortion.
As this is not masked as much as say 2nd or 3rd it is audible.
Also note that these probably weren't audio op-amps and they were not used in a -1x or 1x buffer stage (or up to 2x at most).

Also .. treble quality is solely determined by the output stage 3255 with its relatively higher distortion due to the load dependent class-D design nature.
20kHz is easy for those opamps in buffer config.

 

[Sokel]

Right, in my case it was unnatural sounding cymbals. Old recordings that admittedly aren't the best (but where the performances were great) had become nearly unlistenable for me.

To the point made by @Cbdb2 above, for all I know, distortion increased, and I couldn't care less because those recordings are once again available to me.

Keep in mind, BTW, that it took months and significant trial and error before swapping the op amps. I was trying all sorts of EQ (I'm biamped and use MiniDSP for my crossover). I was trying room treatments. I was looking for alternative masterings and pressings, trying digital vs. analog, etc. Not a single change ever resulted in a welcome change until I swapped those op amps.

If my problem was cymbals I would first make sure that I got a load independent amp as the oposite can have all kinds of interactions,speakers with smooth 1-7kHz response on and off axis and verify all the above with a dirty measurement up close.

I would also check about filter interactions (since you biamp),gain structure,"corrected" response upwards of 500Hz (that's a big NO if not done anechoically with speakers that can be EQ'd,not all can),etc.

The very last place I would look would be opamps in a device working in it's linear range and is checked for problems,like DC,etc.

But that's me.

 

[cgallery]

If my problem was cymbals I would first make sure that I got a load independent amp as the oposite can have all kinds of interactions,speakers with smooth 1-7kHz response on and off axis and verify all the above with a dirty measurement up close.

I would also check about filter interactions (since you biamp),gain structure,"corrected" response upwards of 500Hz (that's a big NO if not done anechoically with speakers that can be EQ'd,not all can),etc.

The very last place I would look would be opamps in a device working in it's linear range and is checked for problems,like DC,etc.

But that's me.

Well I still may swap-out the amps, there are some tremendous deals out there for used A/B amps these days.

 

[Sokel]

Well I still may swap-out the amps, there are some tremendous deals out there for used A/B amps these days.

I forgot about an important one.

Have enough power for my desired SPL after subtracting any gain loses (EQ,negative gains,etc) ,adding enough for my music's transients (as - wait for it -cymbals! ) ,etc.
But that's a given I presume.

 

[cgallery]

When you look at the results II assume you meant the AES paper) you might have noticed that the ones that were clearly audible all had audible level (around -60dB) amounts of higher order distortion.
As this is not masked as much as say 2nd or 3rd it is audible.
Also note that these probably weren't audio op-amps and they were not used in a -1x or 1x buffer stage (or up to 2x at most).

Also .. treble quality is solely determined by the output stage 3255 with its relatively higher distortion due to the load dependent class-D design nature.
20kHz is easy for those opamps in buffer config.

Not audio op amps? You think they weren't fit for the application, or that we should only use op amps that say "audio" on them, or ???

I would like to know the op amps used.

But if I said "these weren't audio op amps" here, you guys would probably shred me to pieces.

 

[cgallery]

I forgot about an important one.

Have enough power for my desired SPL after subtracting any gain loses (EQ,negative gains,etc) ,adding enough for my music's transients (as - wait for it -cymbals! ) ,etc.
But that's a given I presume.

Yeah plenty of power, I'm using Mean Well 36v/5.9a power supplies and was experiencing the cymbals issue at sane listening levels.

 

[antcollinet]

To the point made by @Cbdb2 above, for all I know, distortion increased,

And you are still talking about uncontrolled listening. So for all you know (and by far most likely), nothing audible changed at all.

Just out of interest, have you done the klippel listening test to find out your own personal sensitivity to distortion? With speakers (since we are talking about a speaker amp here). I think you would find it interesting.

If you do, don't forget to post back screenshots of your result here.

Listening Test

www.klippel.de

 

[DanielT]

Then we have this, just to get some sense of levels regarding distortion detection:

 

[cgallery]

And you are still talking about uncontrolled listening. So for all you know (and by far most likely), nothing audible changed at all.

I'm working on it.

I was swapping op amps again last night. The sockets (machined) on these amps are a massive PITA when it comes to extraction. Stamped sockets would be so much easier. This is the reason I said people don't tend to blind test this chit, the swap is a PITA and each time you risk mangling your DUT.

I thought it may make sense to find an adapter with a SOIC-8 socket that would fit in the tight quarters, as it seems like it would be easier to swap op amps that way. BUT, I wouldn't be able to use the stock op amps, and I can't find an SOIC-8 adapter that would fit anyhow.

The smallest ZIF socket I can find is 16-pin (IIRC), I've thought about making an adapter for a couple of these (I CNC my own boards for my projects like my phono preamp). I'm not sure I could make it fit and even if I did, one of you would insist that the adapter was causing an issue (proximity to decoupling or whatever).

 

[cgallery]

Then we have this, just to get some sense of levels regarding distortion detection:

Listening to jazz trio performances you've heard 100 times before (no exaggeration), on gear in your own room, goes a long way in helping to identify differences, IMHO.

When I designed my own tonearm, I was listening to music six to eight hours a day for months. Familiarity goes a long way towards isolating changes.

 

[antcollinet]

This is the reason I said people don't tend to blind test this chit, the swap is a PITA and each time you risk mangling your DUT.

This is true.

Another reason - it is actually quite difficult to do the test sufficiently well controlled. Getting the level matching sufficiently accurate. Making sure whatever method you use for blind switching doesn't give you any tells etc etc. So even if you set the test up and pass it, the first thing you'll face is a bunch of people (me included) challenging you on the robustness of your test setup.

(It should be this way - when it is difficult to get the test right, but also incredibly unlikely to get a pass result, it is far more likely to be a problem with the test setup, than for there to be a reliably audible difference between the devices.)

 

[Blumlein 88]

Listening to jazz trio performances you've heard 100 times before (no exaggeration), on gear in your own room, goes a long way in helping to identify differences, IMHO.

When I designed my own tonearm, I was listening to music six to eight hours a day for months. Familiarity goes a long way towards isolating changes.

I had not seen this. I tried listening with headphones and not looking at it once I saw what was being done. I listened to it 5 times and stopped when I was pretty sure I heard distortion. 4 times it was between .8 and .9 % which was a bit of a surprise. Once it was right about 1%.

 

[antcollinet]

Familiarity goes a long way towards isolating changes.

You could use @pkane s "distort" program to apply distortion of various types and at various levels. Then you can listen as long as you like to whatever music you like to find out what distortion you can hear.

But much easier to do the klipel test first. Perhaps you won't be as sensitive in that test as you would be with your own music - but I doubt very much the difference would be more than 6dB.

 

[cgallery]

Level matching is easy with the MiniDSP and a DMM. Preset 1 is the stock 5532, preset 2 is the 4562.

Edit to add: I'll have no way of knowing what preset is selected, the MiniDSP 2x4HD has no display for this.

 

[DanielT]

Level matching is easy with the MiniDSP and a DMM. Preset 1 is the stock 5532, preset 2 is the 4562.

Edit to add: I'll have no way of knowing what preset is selected, the MiniDSP 2x4HD has no display for this.

Now you can blind test if you feel like it and haven't read the thread from which I took the two files:

A:

gfa565-a.wav | Powered by Box

app.box.com

B:

gfa565-b.wav | Powered by Box

app.box.com

Do you hear any differences between them or not?

I can link to the thread they came from after you listen to them.

 

[AnalogSteph]

Protip: Never-ever-ever upload WAV files to the ol' interwebs if you can help it. They're just pointlessly bigger than FLACs.

 

[cgallery]

Now you can blind test if you feel like it and haven't read the thread from which I took the two files:

A:

gfa565-a.wav | Powered by Box

app.box.com

B:

gfa565-b.wav | Powered by Box

app.box.com

Do you hear any differences between them or not?

I can link to the thread they came from after you listen to them.

That isn't the sort of track where I'd have expected to hear a difference, though.

As I said earlier, I'm hearing a difference very specifically on cymbals in recordings of trios. This is my original post describing the difference:

Replace OP amps. Completely pointless, or not?

You listen to music with half of its power at 6 kHz or above? Or you are just interested in measurement for measurement's sake? There is a valid argument for those of us who like measurements and that comes for the very class D amp's companies as in their evolution they state higher and higher...

www.audiosciencereview.com

When I later found the thread where the audio club members were using recorded noises like clicks to differentiate chips, I thought to myself, "yeah, clicks aren't too unlike a stick hitting a cymbal.

Maybe I heard a difference on vocals, I've listened to a handful of tracks with vocals after the first change, that was it.

But I know I hear a difference on cymbals.