TI NE5532 Audio Opamp Changes

[floobydust]

But the RC4580 exhibits phase-inversion... see Reply #172

1. I don't understand your point.  We know that the new production TI NE5532/A are TI RC4580 die.  So TI decided that was the closest substitute, phase inversion or no phase inversion.

2. Have you actually tested the TI RC4580 for phase inversion?  It may not be the same as the other RC4580 designs.

"TI decided that was the closest substitute" 
Keep on trusting them, see how that works out for you. Phase-inversion in op-amps is a big deal, in audio it's a basic nightmare.

Fellow EE did an EV charger design 6kW using TI op-amps that exhibit phase-inversion. JFET quads.
Well, the IGBT's would explode when that happened. I think for the current-sense. Who has time to test TI's bullshit?

[edavid]

"TI decided that was the closest substitute" 
Keep on trusting them, see how that works out for you. Phase-inversion in op-amps is a big deal, in audio it's a basic nightmare.

I didn't say that I trusted them, or that the RC4580 is a good substitute for the NE5532.  I just said that they shouldn't have decided to remark the RC4580s as NE5532s.

[The Chump]

Here's someone else having problems with the new part, mainly THD
https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1627549/ne5532a-ne5532a-high-thd-from-new-batches-that-are-clearly-not-ne5532a

[floobydust]

"TI decided that was the closest substitute" 
Keep on trusting them, see how that works out for you. Phase-inversion in op-amps is a big deal, in audio it's a basic nightmare.

I didn't say that I trusted them, or that the RC4580 is a good substitute for the NE5532.  I just said that they shouldn't have decided to remark the RC4580s as NE5532s.

TI has made so many transgressions here and there's no way to consider the "new chip" a suitable replacement for the "legacy chip".
They are totally different birds, with some dodgy datasheet deletions/revisions to cover that up. How gross, for the micropennies. Or to keep Chips Act funding and tax concessions flowing in.

The big problem is once this lack of ethics becomes the norm in a company, it just spreads and gets worse. Ref. Boeing's culture fell down that slippery slope.

[exe]

Here's someone else having problems with the new part, mainly THD
https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1627549/ne5532a-ne5532a-high-thd-from-new-batches-that-are-clearly-not-ne5532a

Wow, actual THD measurements. I'm impressed that old TI's part was any good. A wide-spread belief is that TI's part is inferrior to competition due to die shrinkage. But it seems it's doing fine there.

[tom66]

Here's someone else having problems with the new part, mainly THD
https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1627549/ne5532a-ne5532a-high-thd-from-new-batches-that-are-clearly-not-ne5532a

Another "let's discuss this matter privately, just you, me, and the boys" kind of response.  Hmm, why not in public? 

[floobydust]

Here's someone else having problems with the new part, mainly THD
https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1627549/ne5532a-ne5532a-high-thd-from-new-batches-that-are-clearly-not-ne5532a

Wow, actual THD measurements. I'm impressed that old TI's part was any good. A wide-spread belief is that TI's part is inferrior to competition due to die shrinkage. But it seems it's doing fine there.

No it's not doing fine there, plus the new "new chip" is shit. I would say slew-induced distortion. On-semi is king? Not sure about ST.

[pope]

Here's someone else having problems with the new part, mainly THD
https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1627549/ne5532a-ne5532a-high-thd-from-new-batches-that-are-clearly-not-ne5532a

Wow, actual THD measurements. I'm impressed that old TI's part was any good. A wide-spread belief is that TI's part is inferrior to competition due to die shrinkage. But it seems it's doing fine there.

No it's not doing fine there, plus the new "new chip" is shit. I would say slew-induced distortion. On-semi is king? Not sure about ST.

Onsemi seems to have the highest THD across the whole bandwitdh. Excluding the new TI fakes of course.

[tom66]

Onsemi seems to be around 0.025% till about 20kHz, TI original 0.02%.  Pretty good for an inexpensive device.  TI new also about 0.02%, until you get beyond 3kHz.  So from this chart, we can surmise the opamp is good if you want high quality audio in the telephonic spectrum.  A new market for TI perhaps? Audiophile phone calls?

Ok, all kidding aside I doubt anyone can really hear the difference between these parts, but it could really matter if you have several of them chained into a filter setup.

[thm_w]

-If you want to rely on the build in ESD protection, good luck.

Why "good luck"? Its completely reasonable to have designs that rely on the IC for ESD protection, often in low cost devices.
Comment from one of the TI links above suggests its worse than that:

"There are no input protection diodes as one would expect on an NE5532"

And who will buy this "new" creature? Nobody.

So TI has 165,000 in stock and Digikey has 40,000 in stock, but.. no one will ever buy them?

[edavid]

Why "good luck"? Its completely reasonable to have designs that rely on the IC for ESD protection, often in low cost devices.
Comment from one of the TI links above suggests its worse than that:

"There are no input protection diodes as one would expect on an NE5532"

That is not referring to ESD clamp diodes, it's referring to the back to back diodes commonly used to protect superbeta transistor inputs.  Original NE5532 had them, RC4580 doesn't (since it has PNP inputs).  Most people see them as a nuisance rather than a useful feature though.

[Alex Nikitin]

Onsemi seems to be around 0.025% till about 20kHz, TI original 0.02%.  Pretty good for an inexpensive device.  TI new also about 0.02%, until you get beyond 3kHz.  So from this chart, we can surmise the opamp is good if you want high quality audio in the telephonic spectrum.  A new market for TI perhaps? Audiophile phone calls?

Ok, all kidding aside I doubt anyone can really hear the difference between these parts, but it could really matter if you have several of them chained into a filter setup.

0.0025% and 0.002% !

And, depending on the circuit, different opamps do sound different, the RC4580 is not something I would ever consider for anything audio related (same for the LM833 and the majority of its clones, with one possible exception in the NJM2068).

Cheers

Alex

[temperance]

Why "good luck"? Its completely reasonable to have designs that rely on the IC for ESD protection, often in low cost devices.
Comment from one of the TI links above suggests its worse than that:

"There are no input protection diodes as one would expect on an NE5532"

Those ESD ratings serve a different purpose. They inform you how much ESD a device can handle on an assembly line. For components with very low ESD ratings like some RF components you have to adjust the pick and place machine tape peel off speed and other things and have ionizers installed.

What does a typical audio input contain:

Some AC coupling followed by a low pass filter, which is your first line of defence against ESD. Followed by a current limiting resistor and back to back zeners or steering diodes to the supply rails. The low pass filter and what follows are often omitted in those expensive guitar pedals developed by idiots.

 

[David Hess]

Anyway, there was no need for a new part number, since the new parts are RC4580s.  They should have just issued an EOL notice with a suggestion that people use the RC4580 instead, instead of pretending there was a new version of the NE5532.

When Raytheon released their PNP input RC4558/RC4136 operational amplifiers, instead of using the 741 type of part number, their datasheets referred to them as having 741 performance.

Part number suffix is usually the package designation.

Parts like the 741 used part of its suffix for absolute maximum supply voltage.  The original Fairchild part numbers only listed the C (commercial) version as +/-18 volts, while the A, E, and blank were +/-22 volts.  Signetics, National, and Texas Instruments at least followed this.

[floobydust]

Anyway, there was no need for a new part number, since the new parts are RC4580s.  They should have just issued an EOL notice with a suggestion that people use the RC4580 instead, instead of pretending there was a new version of the NE5532.

When Raytheon released their PNP input RC4558/RC4136 operational amplifiers, instead of using the 741 type of part number, their datasheets referred to them as having 741 performance.

Part number suffix is usually the package designation.

Parts like the 741 used part of its suffix for absolute maximum supply voltage.  The original Fairchild part numbers only listed the C (commercial) version as +/-18 volts, while the A, E, and blank were +/-22 volts.  Signetics, National, and Texas Instruments at least followed this.

I'm against keeping the root part number between quite different op-amps - PNP/NPN/CMOS changes. As a "brand" that is stretching it way too far. Maybe for old popular transistors, 2N2222 etc. the die can be a different process - but not IC's?
TI has in a way both obsoleted the NE5532, LM833 etc. and replaced them with "beancounter specials" that perform poorer, not the same.

Grocery store checkout has tiny sized 34g bags of potato chips for $1.50 and I couldn't stop laughing at that ripoff. I may have raised a fist and yelled at the chips lol.
TI version, less for more, would also be a different recipe that tastes terrible.

[Zero999]

Grocery store checkout has tiny sized 34g bags of potato chips for $1.50 and I couldn't stop laughing at that ripoff. I may have raised a fist and yelled at the chips lol.
TI version, less for more, would also be a different recipe that tastes terrible.

At least food retailers and manufactures have to print the weight of the contents on the package. TI can put anything inside an op-amp package and sell it for what they like.

[Kleinstein]

The is still slight difference between the transistor numbers like 2N2222 and the OP-amps like NE5532.
The OP-amp names are given by the companies that first made the amplifiers (e.g. Phillips - signetics for the NE5532). Other manufacturers than were 2nd sources via a licence or made a similar competing part.
The transistors are given out by a kind of standard oranization. So the specs there are more fixed and a transistor is simpler with usually less specs to care about.

[floobydust]

I was trying to say a jellybean transistor number is a different die/process over the decades, between the manufacturers etc.- but with the ethical rule "meets or exceeds" datasheet specifications.
Not TI's sleazy version which would give less P_D, I_C, lower V_CE and f_T for the same p/n lol.

I'm not sure if transistor numbers are from JEDEC or not.

[Zero999]

I was trying to say a jellybean transistor number is a different die/process over the decades, between the manufacturers etc.- but with the ethical rule "meets or exceeds" datasheet specifications.
Not TI's sleazy version which would give less P_D, I_C, lower V_CE and f_T for the same p/n lol.

I'm not sure if transistor numbers are from JEDEC or not.

Yes. To use your food analogy, in many countries, trading laws mean that in order for a product to be sold as chocolate it must contain a certain percentage of cocoa butter, sausages must contain a certain percentage of meat etc. Or course this means food companies still sell colocate flavour bars and bangers with lower cocoa butter and meat content respectively.

[squadchannel]

https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1652528/ne5532-ne5532-lm833-rc4580-and-mc33078-all-now-the-same-die

NE5532: NE5532, LM833, RC4580 and MC33078 all now the same die?
Andy Grove
Part Number: NE5532

Hi,

Following the recent datasheet changes, including the representative schematic on the new NE5532 datasheet, as well as its degraded performance specifications, it appears that these parts have all converged to the same die and are now effectively the same thing.

I am one of many who would like confirmation or refutation of this, because the revised specifications effectively make the new NE5532 not an NE5532, regardless of its architecture or process.

It's only fair that TI are open about this, otherwise we are designing in or attempting repairs with parts which no longer correspond to historical and accepted NE5532 performance.

Michael Hartshorne 10 hours ago

Hi Andy,

To answer your top-line question, yes, these products are on the same die. We recognize the critical importance of consistency for your designs and appreciate your feedback. The NE5532 data sheet updates were made to help clarify these differences. Additionally, the PCN procedure has been updated to help give more detail to PCN recipients. We are here to support you if you have additional questions.

Regards,
Mike

 

[tom66]

Interesting that the MC33078 datasheet states the absolute maximum for the part is ±18V but also provides a recommended operating condition up to ±18V.

https://www.ti.com/lit/ds/symlink/mc33078.pdf

Have they been making the die for MC33078 for some time, and just discontinued NE5532 and decided to package these chips as NE5532?

[David Aurora]

Just received my latest Digikey order and sure enough, the TI 5532's in it are the new version (I had previously asked Digikey what would actually ship and they were unable to tell me). I've just requested a return given they can't be used as a drop in replacement for the exact same part number from the exact same brand. Let's see how this goes.

[David Aurora]

Accepted basically immediately with no argument or further questions, they've told me to throw them in the bin rather than return them and they've already started a refund.

I'd be interested to know how many return requests they'll get over this. I thankfully only bought a handful to see which version I'd get given the lack of clarity when I enquired (in the off chance I got proper ones I would have quickly ordered a bunch of them, but no luck there).

[sparkydonkey]

well David Aurora..I think this is very much a case of 'action speaks louder than word' and thank you for being yet one more of these, I only can hazily guess that if TI had to keep sorting out a lot of refunds left&right especially to big ordered-many-reels-many-times commercial customers they would perhaps finally wise up?

[David Aurora]

well David Aurora..I think this is very much a case of 'action speaks louder than word' and thank you for being yet one more of these, I only can hazily guess that if TI had to keep sorting out a lot of refunds left&right especially to big ordered-many-reels-many-times commercial customers they would perhaps finally wise up?

Hopefully so! I mean nobody is going to notice or care about a few op amp refunds here and there, but hopefully some larger manufacturers do the same with parts they can't use