Yeah, but the NE5532 that you buy today is not made by Signetics because they don't exist anymore. I also guarantee you that it isn't made on the same equipment as originally used because that was decommissioned long ago. So an NE5532 bought today will be a device that's been redesigned many times to be as close to the original as possible while being manufacturable on current equipment.
You then add that everybody was copying everybody back then. So you have the original Signetics part, the TI "NE5532", the NSC "NE5532", etc. I think there's a JRC version of it as well. In Doug Self's article on audio opamps in AudioXpress 01/25 he mentions that he tested a "not TI" NE5532, but doesn't actually list which brand it is. So maybe some are better than others.
But I don't think there's any doubt that the NE5532 was miles ahead of the competition in 1979. It pushed the limits of the test equipment that existed at the time and much of the current equipment too. It's not really until within the last decade or two that we've had sources that started pushing the NE5532. Then someone, reasonably, asked, "why can't we have a better opamp that's also affordable?" and developed the LM4562.
Not thereby said that there was nothing until the LM4562. The AD797 and OPA627 spring to mind. But they're both very expensive. The OPA2134 had a good reputation, but it isn't much better (if better at all) than the NE5532 as I recall.
Tom
Just to mention some more descent opamps with PNP inputs apart from NJM2068 - NJM4560/65/80.
AFAIK (and probably from DIYaudio) LME49720 / 49860 also have PNP input BJTs inherited from LM4562.
Not really modern opamps, but not horribly old))
What's interesting about PNP-input devices: they usually have no protective diodes between inputs and allow differential input voltage in full Vcc-Vee range. Sometimes it is useful.
AFAIK (and probably from DIYaudio) LME49720 / 49860 also have PNP input BJTs inherited from LM4562.
Not really modern opamps, but not horribly old))
What's interesting about PNP-input devices: they usually have no protective diodes between inputs and allow differential input voltage in full Vcc-Vee range. Sometimes it is useful.
Should just point out there is no such thing as an ordering between opamps for being "better". The LM4562 is better than the NE5532 in some respects, and worse in others. Each application asks for different combinations of parameters from an opamp, so you can only compare opamps w.r.t. a particular application. For instance in many circuits the AD797 is much noisier than the NE5532, despite being ultra-low voltage noise. Current noise is often forgotten and some applications require low current noise, such as an MM RIAA preamp. An MC RIAA preamp on the other hand cares more about voltage noise. NE5532 is a good fit for MM, poor for MC, and vice-versa for the AD797.
The LM4562 has two achilles heels, its very sensitive to RF pickup, and its not screened during manufacture for excess noise.
And its worse when a particular application is a PCB. The amount of stray capacitance, stray inductance and the decoupling scheme may work with one opamp, but not with a different one, even if that different one is theoretically superior for the application (for instance it oscillates on the PCB and performs very poorly).
Rolling opamps won't necessarily tell you which is best, even for the application... Proper circuit design and testing is the key.
The LM4562 has two achilles heels, its very sensitive to RF pickup, and its not screened during manufacture for excess noise.
And its worse when a particular application is a PCB. The amount of stray capacitance, stray inductance and the decoupling scheme may work with one opamp, but not with a different one, even if that different one is theoretically superior for the application (for instance it oscillates on the PCB and performs very poorly).
Rolling opamps won't necessarily tell you which is best, even for the application... Proper circuit design and testing is the key.
I could humbly add few ignored competitors from old age.
For NE5532, NJM4580. It is even lower cost, lower noise ( noises I refer to MM riaa apication) and has unusualy high output current capability, can drive many headphones directly.
On NE5534, i had very good experience with 20 year younger competitors from early 1990's (still 35 y. o.) OP27 and TLE2027, both are little bit more expensive, like 4 euro, also low SR (they have uncompensated brothers, min gain of 3 like 5534, OP37 and TLE2037 that are bit faster) , but with better noise performance and perfect DC precision, usually no need for coupling caps there.
All 3 mentioned are still today avaliable at Mouser, and I found them as common as 5532 in pro audio gear.
For NE5532, NJM4580. It is even lower cost, lower noise ( noises I refer to MM riaa apication) and has unusualy high output current capability, can drive many headphones directly.
On NE5534, i had very good experience with 20 year younger competitors from early 1990's (still 35 y. o.) OP27 and TLE2027, both are little bit more expensive, like 4 euro, also low SR (they have uncompensated brothers, min gain of 3 like 5534, OP37 and TLE2037 that are bit faster) , but with better noise performance and perfect DC precision, usually no need for coupling caps there.
All 3 mentioned are still today avaliable at Mouser, and I found them as common as 5532 in pro audio gear.
I like the OPA2134 and OPA2132. JFet inputs on these so they are better for other applications.
The NJM parts were good, I have some NJM4580 and others.
Older parts are not necessarily better. In fact as technology and materials move forward, plus processes, I can't think of any reasons an older op amp would be better. As Tom said, current manufacture of old part numbers are no longer the original design. Typically better as well. For sure, noise should be lower from better manufacturing.
The NJM parts were good, I have some NJM4580 and others.
Older parts are not necessarily better. In fact as technology and materials move forward, plus processes, I can't think of any reasons an older op amp would be better. As Tom said, current manufacture of old part numbers are no longer the original design. Typically better as well. For sure, noise should be lower from better manufacturing.
No? I find it perfectly possible to compare two opamps and find one objectively better than the other. Without looking at the data sheet I'd be willing to wager that the LM4562 is better than the µA741 on just about every parameter except price. That doesn't mean that I would use the LM4562 if I was designing, say, a doorbell. Common sense does apply.
The DIP version is sensitive to RF pickup. I've never had issues with the SOIC version. And, frankly, if you want the manufacturer to screen the parts for 1/f noise, be prepared to pay through the nose as that's a very expensive test (it takes a long time). The NE5532 isn't screened for 1/f noise either...
And the NE5534 would be even better. Because I can compare the current noise of the NE5532 and NE5534 and conclude that the NE5534 has lower current noise, thus, is better in the MM preamp.
Tom
Well yes, yours is a necessary clarification.
But if we had an original one i think we could say that it is a very current component.
In addition to this i wonder if the other op-amps are also subject to various copies.
If so, can there be substantial differences between one factory and another?
However i once tried the direct change in an amplifier between NE5532 and OP275, and when listening to it there was no difference.
Looking at the datasheet values, the NE5532 noise current at 1 kHz appears to be largely 1/f. If it keeps dropping until it reaches √(2qIB), its typical value should actually be better than the NE5534(A) at high audio frequencies, where it matters most for MM.
Hi Gioba,
Human beings are terrible test instruments for one. There must be a large difference between two things before most can reliably determine which is better.
The original NE5532 has been redesigned to match processes in use at the time. The lithos (artwork) are secret to each factory, so no two are identical. What they are is "work a-likes" with similar but not identical characteristics. So there can be large differences between factories. Even production runs can be a bit different. Detectible by test equipment, rarely the human ear.
Using the proper device for an application is most important. Device characteristics depend on circuit configuration as well.
Human beings are terrible test instruments for one. There must be a large difference between two things before most can reliably determine which is better.
The original NE5532 has been redesigned to match processes in use at the time. The lithos (artwork) are secret to each factory, so no two are identical. What they are is "work a-likes" with similar but not identical characteristics. So there can be large differences between factories. Even production runs can be a bit different. Detectible by test equipment, rarely the human ear.
Using the proper device for an application is most important. Device characteristics depend on circuit configuration as well.
Ok but in the end if it is not detectable by the human ear perhaps it takes on little importance on a practical level.
Then it is possible that my ear is not sufficiently reliable but, it is the only one i trust.
I do not have the possibility of making laboratory measurements.
Yes i certainly agree.
Thanks for your info anatech.
As mentioned, low power, micro power , single supply, many rail to rail in those applications will all use PNP inputs.
Very common for comparator circuits
Not as common for audio use.
The processing well extended for medical and military use. Many not even public
Lower input bias current , better matching, lower noise and higher speed in lower power, single rail applications.
Probably hundreds of them. Wont waste much time scrummaging through data sheets or memory to " prove it"
Common mode will extend to the negative rail or very slightly above.
Far as pnp lateral, that is where they belong, high speed, low noise small signal. Not power devices.
Fairchild, National Semi was rather proud of their early PNP lateral / BJT developments and extend further now with TI.
Including my father and his name included on multiple patents.
Yes National included the engineers names on the patents, and numerous awards and plagues.
Which I watched him receive and walked past daily in my childhood
Very common for comparator circuits
Not as common for audio use.
The processing well extended for medical and military use. Many not even public
Lower input bias current , better matching, lower noise and higher speed in lower power, single rail applications.
Probably hundreds of them. Wont waste much time scrummaging through data sheets or memory to " prove it"
Common mode will extend to the negative rail or very slightly above.
Far as pnp lateral, that is where they belong, high speed, low noise small signal. Not power devices.
Fairchild, National Semi was rather proud of their early PNP lateral / BJT developments and extend further now with TI.
Including my father and his name included on multiple patents.
Yes National included the engineers names on the patents, and numerous awards and plagues.
Which I watched him receive and walked past daily in my childhood
Last edited:
It'd be interesting to pinpoint the exact differences between the Ti NE5532 and the NJM5532DD. I've read (here?) that they are different parts, but the datasheets for each pretty much mirror each other give or take a few %. In practice, I have found the NJM part to be trickier to use than the Ti 5532 (I had an intermittent oscillation at the low-ish freqency of 280KHz in a filter section, and the output appeared to be completely flipping phase at the same time). Considering both parts look almost identical on paper and both have pedestrian slew rates, I am intrigued as to what might make the NJM part less flexible than the Ti. I have some Signetics 5532s and tested them against the Ti parts many years back. I was surprised to see the Signetics part outperform the Ti part in terms of THD+N, but only by a fraction of a dB. If I had time I'd compare all 3 versions.
Well, those "NE5532" all come from different artwork and processes. I'm not surprised at all. It's like tubes from different factories. Same number, not exactly the same part but close enough.
Hi Gioba,
Whether you can actually hear something or not depends on your entire system, ambient noise levels and so forth. Then there is your own mood and alertness level at the time. That is a huge variable. Throw in different people and you have a massive problem issue with correlation. On top of that, people tend to agree with others they feel are "gifted" or respect. Kind of a "me too" setup.
If you can't hear it, it doesn't matter to you and that's okay. But an opinion on sound quality is only valid for you in your current system - and that can change. The real performance of the equipment doesn't change beyond wear, tubes being more susceptible to this.
Measured performance using industry correct loads and equipment, and showing a spectrum are highly accurate for determining sound quality that does correlate between people. If you can't hear noise or distortion below a certain level, you're right. It won't matter. However a designer can see problems and work to solve them. More often than not fixing this stuff results in a better piece of equipment. I have two totally different amplifiers that measure pretty much the same in distortion residuals. They sound the same on speakers. I mean identical. Now this statement assumes the equipment performs well at all levels in it's operating range (power and load for an amplifier). If, say a power amplifier, current limits or has other odd behavior it will not sound good. Again, easily measured.
Back in the 1980's we had to work to "see" issues below, say -80dB below 1 watt. Today's equipment can "see" 40 dB below that level or better. That is well below the limits of a human body to perceive anything in a silent space. Most people complaining about measured response are referring to a meter pointer in 1970's equipment, or even 1980's test equipment and methods. We have moved well beyond that time. I'll be honest, designers and companies who dispute the validity of proper measured response have defective product. So, they attack the truth. You'll also see comments like "synergy" between equipment. Naw, incorrect. Make each piece perform properly and you can mix and match and get the best sound given they are designed properly and perform equivalently. That's assuming your power amplifier has enough output power for your speakers and SPL levels. However that puts many "golden ear" people and companies out of business. Ahh - the romance and sales tactics.
Hi Gioba,
Whether you can actually hear something or not depends on your entire system, ambient noise levels and so forth. Then there is your own mood and alertness level at the time. That is a huge variable. Throw in different people and you have a massive problem issue with correlation. On top of that, people tend to agree with others they feel are "gifted" or respect. Kind of a "me too" setup.
If you can't hear it, it doesn't matter to you and that's okay. But an opinion on sound quality is only valid for you in your current system - and that can change. The real performance of the equipment doesn't change beyond wear, tubes being more susceptible to this.
Measured performance using industry correct loads and equipment, and showing a spectrum are highly accurate for determining sound quality that does correlate between people. If you can't hear noise or distortion below a certain level, you're right. It won't matter. However a designer can see problems and work to solve them. More often than not fixing this stuff results in a better piece of equipment. I have two totally different amplifiers that measure pretty much the same in distortion residuals. They sound the same on speakers. I mean identical. Now this statement assumes the equipment performs well at all levels in it's operating range (power and load for an amplifier). If, say a power amplifier, current limits or has other odd behavior it will not sound good. Again, easily measured.
Back in the 1980's we had to work to "see" issues below, say -80dB below 1 watt. Today's equipment can "see" 40 dB below that level or better. That is well below the limits of a human body to perceive anything in a silent space. Most people complaining about measured response are referring to a meter pointer in 1970's equipment, or even 1980's test equipment and methods. We have moved well beyond that time. I'll be honest, designers and companies who dispute the validity of proper measured response have defective product. So, they attack the truth. You'll also see comments like "synergy" between equipment. Naw, incorrect. Make each piece perform properly and you can mix and match and get the best sound given they are designed properly and perform equivalently. That's assuming your power amplifier has enough output power for your speakers and SPL levels. However that puts many "golden ear" people and companies out of business. Ahh - the romance and sales tactics.
The NE5534 started life in the 1970-ies as the TA1034 designed by Philips.
I still have a few of those. Maybe I should test them. Or sell them ;-)
Jan
I still have a few of those. Maybe I should test them. Or sell them ;-)
Jan
Hey Jan,
I didn't know that. Yes, test them! I always thought Signetics designed them, I have an early data book.
You never know, there may be someone out there that would buy them.
I didn't know that. Yes, test them! I always thought Signetics designed them, I have an early data book.
You never know, there may be someone out there that would buy them.
Hi anatech, sorry but unfortunately it's the same old story.
The objectivity of laboratory measurements is....objective in the laboratory.
You yourself state that the variable is enormous and that for many different people there is a huge correlation problem.
On this basis, the component that works correctly is measurable by an instrument but completely questionable to the ears of many people.
I can't believe that every person has the same perception, or taste, or competence, or experience.
It is precisely from this reasoning that i said that i only trust my perception, which does not mean that mine are "golden ears", but they are mine, and i don't change them with anyone else, because i make the listening experience with my own.
It is a perfectly logical equation.
In my opinion, this is valid for the 70s, for the previous ones, for today, and for the future.
Maybe i'm wrong but according to my listening experiences, the synergy between equipment is fundamental, first of all because no component is perfect, then because the resulting combination must be the one that reflects your personal taste, and since the devices are not all the same and each device has its own unique characteristics, the combinations in my opinion are very important.
Even cables, regardless of their market price.
The speech is probably different in the factory, production needs standard elements to make components that can be defined with X characteristics, for which a buyer can choose to purchase a stock.
Semiconductors, IC... here the measured tests have their objective to give objectivity to the buyer.
But for hi-fi devices the measurements are only indicative, it's the listening test that gives the final judgment, and it is subjective.
If i only like vintage it isn't romanticism and i don't sell vintage, but it's a different thing constructively, first of all.
As i said, I don't like SMD because for me it's a technology born for IT and introduced for audio for the only purpose of containing production costs.
Furthermore, for the maintenance of the devices, the through hole is much more convenient and uses classic and more readily available components.
This is according to MY experience, which may be wrong but is based on a probably common life, i think.
Vintage because all the components, once were produced all over the world.
To give just a few examples, made in USA, made in Germany, made in Japan, had a meaning.
A precise, excellent meaning.
The best designers, we had them, USA, Europe, Japan..
Today almost (?) everything is produced in China, it produces for the national market and it produces for others.
Let me ask you a question: in the 70s, 80s and first half of the 90s, when the market reached its maximum peak, where was China?
Why did it emerge only in the new millennium?
Everyone give their own answer.
Please excuse the length of my post, but i would like to emphasize that it is not "a war" between you and me, or between analog and digital, or cables yes cables no, or old against new, simply everyone has their own vision of things but behind it there is a very precise reasoning, disillusioned above all, for me.
My 2 cents.
Well...maybe they are 4..
Hi Gioba,
Manufacturing moved to other countries for purely economic reasons. Production moved to Japan and they eventually priced themselves out of the market. Some of the best production was Japanese. They went to Taiwan, then China and maybe starting in India now. It's 100% about how much it costs to build something and then transport it to market. We are now in a age of contract manufacturing (very bad for consumers) where companies buy a design and it is shipped out. The factory now drops everything and has zero responsibility for support of manufacturing errors, or component defects (as in - they used fakes).
Other Asian countries are on that hit parade. The USA moved other production to Mexico in the 1970s purely for labour cost considerations.
Now, you know what I love about electronics (physics)? Everything on this planet follows the rules of physics. Nothing is suspended just because it is audio. Just like gravity is always the same no matter how much you don't want to hit the ground. Everything is not only predictable, it's been proved empirically with real people for a long time now. If you have a problem with measured results, it's probably because of the measurements taken, how and what equipment and method. Done properly, measured results are accurate. Especially over the long term for the vast majority of people.
If you have equipment that doesn't perform well, people split into groups of what they can accept or not. But almost everyone prefers an accurate system. People tend to prefer what they are used to hearing, but over time they do gravitate to better systems. Now, we also have some folks who have a strong preference for a "sound". That's fine for them, but it does not reflect the majority, they are a "flyer" in the data set. Many have measured, listened and observed people for decades. As instrumentation got better we could figure out what amplifiers and speakers (everything else) was doing. That's why equipment has been getting better. There isn't any equipment "designed by ear" that actually does sound good compared to equipment designed with the aid of testing. Note that he testing is in addition to listening tests. These days the listening tests are more of a sanity check if people are doing things right.
So as nice as it sounds that we can hear what cannot be measured ... well that became complete false a long time ago. Not everyone can measure what is needed, but many can and do daily.
No fight, this is merely reality today.
Yup. Some folks stuck in the dark ages. lol! I'm not going to debate this.
Manufacturing moved to other countries for purely economic reasons. Production moved to Japan and they eventually priced themselves out of the market. Some of the best production was Japanese. They went to Taiwan, then China and maybe starting in India now. It's 100% about how much it costs to build something and then transport it to market. We are now in a age of contract manufacturing (very bad for consumers) where companies buy a design and it is shipped out. The factory now drops everything and has zero responsibility for support of manufacturing errors, or component defects (as in - they used fakes).
Other Asian countries are on that hit parade. The USA moved other production to Mexico in the 1970s purely for labour cost considerations.
Now, you know what I love about electronics (physics)? Everything on this planet follows the rules of physics. Nothing is suspended just because it is audio. Just like gravity is always the same no matter how much you don't want to hit the ground. Everything is not only predictable, it's been proved empirically with real people for a long time now. If you have a problem with measured results, it's probably because of the measurements taken, how and what equipment and method. Done properly, measured results are accurate. Especially over the long term for the vast majority of people.
If you have equipment that doesn't perform well, people split into groups of what they can accept or not. But almost everyone prefers an accurate system. People tend to prefer what they are used to hearing, but over time they do gravitate to better systems. Now, we also have some folks who have a strong preference for a "sound". That's fine for them, but it does not reflect the majority, they are a "flyer" in the data set. Many have measured, listened and observed people for decades. As instrumentation got better we could figure out what amplifiers and speakers (everything else) was doing. That's why equipment has been getting better. There isn't any equipment "designed by ear" that actually does sound good compared to equipment designed with the aid of testing. Note that he testing is in addition to listening tests. These days the listening tests are more of a sanity check if people are doing things right.
So as nice as it sounds that we can hear what cannot be measured ... well that became complete false a long time ago. Not everyone can measure what is needed, but many can and do daily.
No fight, this is merely reality today.
I must say I enjoy reading what anatech and gioba have shared. 🙏 Thanks.